Table of Contents

--------------------------
Chapter 1: Introduction
--------------------------
• 1.1 Preliminary Remarks
• 1.2 Symbolic Computation
• 1.2.1 Global Rules
• 1.2.2 Local Rules
• 1.2.3 Pattern Matching
• 1.2.4 Using Patterns in Rules
• 1.2.5 Restricting Pattern Matching with Predicates
• 1.2.6 Pure Functions
• 1.2.7 Map and Apply
• 1.2.8 Nest and Fold

-------------------------------------
Chapter 2: The Duffing Equation
-------------------------------------
• 2.1 The Duffing Equation
• 2.2 Straightforward Expansion
• 2.3 The Lindstedt-Poincaré Technique
• 2.4 The Method of Multiple Scales
• 2.4.1 Second-Order Real-Valued System
• 2.4.2 First-Order Real-Valued System
• 2.4.3 First-Order Complex-Valued System
• 2.5 Variation of Parameters
• 2.6 The Method of Averaging

-------------------------------------------------------------------
Chapter 3: Systems with Quadratic and Cubic Nonlinearities
-------------------------------------------------------------------
• 3.1 Nondimensional Equation of Motion
• 3.2 Straightforward Expansion
• 3.3 The Lindstedt-Poincaré Technique
• 3.4 The Method of Multiple Scales
• 3.4.1 Second-Order Real-Valued System
• 3.4.2 First-Order Real-Valued System
• 3.4.3 First-Order Complex-Valued System
• 3.5 The Method of Averaging
• 3.6 The Generalized Method of Averaging
• 3.7 The Krylov-Bogoliubov-Mitropolsky Technique
 • 3.8 The Method of Normal Forms

-------------------------------------------------------------
Chapter 4: Forced Oscillations of the Duffing Equation
-------------------------------------------------------------
• 4.1 Straightforward Expansion
• 4.2 The Method of Multiple Scales
• 4.2.1 Preliminaries
• 4.2.2 Primary Resonance
• 4.2.3 Secondary Resonances Due to Cubic Nonlinearities
• 4.2.4 Secondary Resonances Due to Quadratic Nonlinearities
• 4.2.5 First-Order Real-Valued System
• 4.2.6 First-Order Complex-Valued System
• 4.2.7 The Function MMS1
• 4.3 The Generalized Method of Averaging
• 4.4 The Method of Normal Forms

-----------------------------------------------------------------------------------------
Chapter 5: Higher-Order Approximations for Systems with Internal Resonances
-----------------------------------------------------------------------------------------
• 5.1 Euler-Lagrange Equations
• 5.2 Method of Multiple Scales
• 5.2.1 Second-Order Real-Valued System
• 5.2.2 First-Order Real-Valued System
• 5.2.3 First-Order Complex-Valued System
• 5.3 Method of Normal Forms
• 5.4 Generalized Method of Averaging

----------------------------------------------------------------------------------
Chapter 6: Forced Oscillators of Systems with Finite Degrees of Freedom
----------------------------------------------------------------------------------
• 6.1 Externally Excited Linearly Uncoupled Systems
• 6.2 Parametrically Excited Linearly Coupled Systems

----------------------------------------------------------------
Chapter 7: Continuous Systems with Cubic Nonlinearities
----------------------------------------------------------------
• 7.1 Solvability Conditions and the Concept of Adjoint
• 7.2 Hinged-Clamped Beam
• 7.2.1 EOM and BC's
• 7.2.2 Direct Attack of the Continuous Problem
• 7.2.3 Discretization of the Continuous Problem
• 7.2.4 Method of Time-Averaged Lagrangian
 • 7.3 Cantilever Beam
• 7.3.1 EOM and BC's
• 7.3.2 Direct Attack of the Continuous Problem
• 7.3.3 Discretization of the Continuous Problem
• 7.3.4 Method of Time-Averaged Lagrangian

--------------------------------------------------------------------------------
Chapter 8: Continuous Systems with Quadratic and Cubic Nonlinearities
--------------------------------------------------------------------------------
• 8.1 Buckled Beams
• 8.1.1 Postbuckling Deflection
• 8.1.2 Perturbation Analysis
• 8.1.3 The Function MMSDirect11
• 8.1.4 Three-to-One Internal Resonances Between the First Two Modes
• 8.1.5 One-to-One and Three-to-One Internal Resonances Between the First and
Third Modes
• 8.1.6 One-to-One and Three-to-One Internal Resonances Between the First and
Fourth Modes
• 8.2 Circular Cylindrical Shells
• 8.2.1 First-Order Solution
• 8.2.2 Second-Order Solution
• 8.2.3 Solvability Conditions
• 8.3 Near-Square Plates
• 8.3.1 First-Order Solution
• 8.3.2 Second-Order Solution
• 8.3.3 Solvability Conditions
• 8.3.4 Mixed Approach

---------------------------------------------------------------------------------------
Chapter 9: Higher Approximations of Continuous Systems Having Two -to-One
Internal Resonances
---------------------------------------------------------------------------------------
• 9.1 Two-Mode Interactions in Buckled Beams
• 9.1.1 First-Order Solution
• 9.1.2 Second-Order Solution
• 9.1.3 Solvability Conditions
• 9.2 Four-Mode Interactions in Suspended Cables
• 9.2.1 First-Order Solution
• 9.2.2 Second-Order Solution
• 9.2.3 Solvability Conditions
Chapter 1
Introduction
For convenience we turn off the spell checker by typing

Off@General::spell1D

à 1.1 Preliminary Remarks

à 1.2 Symbolic Computation

The Mathematica programming language allows not only for interactive manipulations of a wide range of objects, but it also
encourages the automation of complicated computations. Mathematica actually works by using systems of rewrite rules, and
the user can freely create and use his or her own systems of rules. This distinguishes it from traditional programming lan-
guages, which normally have no such features. We have applied rule-based and functional programming techniques in
Mathematica throughout this book. Hence, we briefly cover the basic ingredients of rule-based programming in Sections
1.2.1--1.2.5 and some functional programming features in Sections 1.2.6--1.2.8.

For further information on Mathematica, consult the standard Mathematica reference: The Mathematica Book , Stephen
Wolfram, Third Edition, Wolfram Media and Cambridge University Press, 1996. Full descriptions of the functions in the
standard Mathematica packages are available in Mathematica 3.0 Standard Add-On Packages manual, Wolfram Media
and Cambridge University Press, 1996. Alternatively, simply click on the Help menu to view the Help Browser.

In the Text cells of this book, we use bold-magenta to indicate user-defined functions or names and bold-black to indicate
Mathematica's built-in functions.

Ÿ 1.2.1 Global Rules

Global rules are applied whenever the appropriate left-hand side is encountered or matched. The advantage of using global
rules is that one does not need to apply explicitly a rule to get things done. However, it is almost impossible to prevent such
global rules from being applied to a certain expression. There are two types of user-defined global rewrite rules: those using
= and those using :=. The distinction between them lies in when the right-hand side is evaluated. Moreover, each of these
rules has two forms, depending on where the rule is stored, resulting in four rules indicated by = , ^= , := , and ^:=.
2 Chapter1.nb

Ÿ 1.2.1.1 = rules

We can think of =, an assignment statement in analogy with traditional procedural programming languages, as a (global)
rewrite rule. The characteristic property of using the rule = is that the right-hand side is evaluated immediately when the
assignment is made, and all subsequent matched left-hand sides are replaced by the evaluated right-hand side. For example,
let us consider the rule

a = 2;

The output is suppressed by ending the input with a semicolon. From now on, whenever a is encountered in any expression,
Mathematica will replace it by its value 2.

expr1 = a2 + 3 - a + Tan@aD

5 + Tan@2D

Let us try to evaluate expr1 for another value,

a = 3;

expr1

5 + Tan@2D

The new value of a is not recognized in expr1.

Let us try to clear the definition of a by using Clear or

a =.

Then, let us evaluate expr1,

expr1

5 + Tan@2D

Again, expr1 is not changed accordingly.

To be able to evaluate expr1 for different values of a, we need to either define expr1 before the assignment of a or use ":=",
instead of "=". The precise meaning of := is taken up in the next section. Here, we illustrate the former choice.

Clear@aD; expr1 = a2 + 3 - a + Tan@aD

3 - a + a2 + Tan@aD

a = 2; expr1

5 + Tan@2D
Chapter1.nb 3

a = 3; expr1

9 + Tan@3D

In traditional programming languages, the left-hand side of an assignment statement is required to be a simple identifier (i.e.,
a symbol). In computer algebra, the left-hand side can be arbitrarily complicated. For example,

look@here + 9D = Expand@H1 + xL^ 2 + yD

1 + 2 x + x2 + y

Note that the output of an = expression is the evaluated form of the right-hand side. Mathematica regards the left-hand side
as a pattern . Whenever Mathematica finds something that matches this pattern, it replaces the something by the evaluated
right-hand side.

look@there + look@here + 5 + 2 ^2DD

look@1 + there + 2 x + x2 + yD

In this evaluation, the pattern look @here + 5 + 2 2 D simplifies to look @here + 9D , which is replaced by 1 + 2 x + x2 + y. The
resulting expression does not match any pattern involving look and so it is left in unevaluated form. This rule is stored with
look.

? look

Global`look

look@9 + hereD = 1 + 2 * x + x ^ 2 + y

Again, we see the evaluated form on the right-hand side.

There are some problems associated with left-hand sides that are not symbols. For example, suppose we try to make the
following rule:

x+y= z

Set::write : Tag Plus in x + y is Protected.

We get an error message stating that Plus is protected, which means new rules cannot be added for Plus. Every time Mathe-
matica encounters Plus, it searches through the rules for Plus to see if anything applies. If we add a new rule for Plus, then
that rule would have to be examined at every subsequent addition. When a rule of the form x + y = z is given, Mathematica
interprets it as a rule of the form Plus@x, yD = z. Rules have to be stored somewhere and the default rule is that for the head of
the left-hand side. One way of adding a new rule for Plus is to unprotect Plus, make the rule, and then reprotect it.

Unprotect@PlusD;
x + y = z;
Protect@PlusD;

Now whenever Mathematica sees x + y, it rewrites it as z.

4 Chapter1.nb

x + s +y + t

s + t +z

Alternatively, we can use UpValues to associate the rule with the unprotected argument of the left-hand side. (Definitions
that attach a value to the head of the left-hand side are called DownValues of the head.) For example,

s + t ^= u

Note the caret ^ before the = sign. This rule is associated with the symbol s or t.

?s

Global`s

s •: s + t = u

t + s +t

t+u

A given symbol can have both up and down values. Let's give s a down value in addition to the up value it already has.

s@x_D := Exp@I a xD;

Then looking at s shows both kinds of values:

?s

Global`s

s •: s + t = u

s@x_D := Exp@I * a * xD

Finally, we can access the up and down values individually.

8UpValues@sD, DownValues@sD<

88HoldPattern@s + tD ¦ u<, 8HoldPattern@s@x_DD ¦ Exp@I a xD<<

Now let us try naively define the sine function using an = rule according to

m@xD = Sin@xD;

It works properly for the symbol x but not for anything else.

8m@xD, m@yD, m@1D<

8Sin@xD, m@yD, m@1D<

Chapter1.nb 5

This is where the special symbol _ comes in. The form x_ means a pattern named x. We can show the internal representation
of any expression, such as x_ , by either using

FullForm@x_D

Pattern@x, Blank@DD

or using

x_ •• FullForm

Pattern@x, Blank@DD

An underscore _ in a pattern matches anything, so it is a dummy variable or a "wild card". If it appears on the left-hand side
of "=" rule with a name, like x, then the left-hand side is rewritten as the right-hand side with x replaced by the desired
variable. Let us use this to redefine m; that is,

m@x_D = Sin@xD;

Now m works for any argument:

8m@xD, m@yD, m@1D, m@s - tD<

8Sin@xD, Sin@yD, Sin@1D, Sin@s - tD<

Thus, one can use = rules to define functions.

Ÿ 1.2.1.2 := rules

Rules using := do not evaluate the right-hand side immediately but instead leave it unevaluated until the function is actually
used. They can be used with simple left-hand sides or with left-hand sides containing patterns. As an example, let us define
and compare two rules using = or :=.

Clear@xD;
lhs1 = Expand@Hx + 1L ^2D;
lhs2 := Expand@Hx + 1L ^2D;

If these are evaluated, they give the same result.

8lhs1, lhs2<

81 + 2 x + x2 , 1 + 2 x + x2<

If we now give a value to x, then lhs1 and lhs2 will use this value in different ways.

x = p + q;

8lhs1, lhs2<

81 + 2 Hp + qL + Hp + qL2 , 1 + 2 p + p2 + 2 q + 2 p q + q2<
6 Chapter1.nb

If the left-hand side of a := rule contains a pattern, then on a subsequent occurrence of the left-hand side with actual argu-
ments, the formal (dummy) arguments (or names of patterns) on the right-hand side are replaced by the actual arguments and
then the right-hand side is evaluated. Thus, each time the left-hand side of such a rule matches something, it is replaced by a
new evaluation of the right-hand side. To see the difference, we again set up the two rules

f@k_D = Expand@k^ 2D

k2

g@k_D := Expand@k^ 2D

We note that the right-hand side of the rule for f is evaluated immediately when it is defined, whereas the right-hand side of g
is kept unevaluated. Now, try out these two definitions on the same value.

8f@p + qD, g@p + qD<

8Hp + qL2, p2 + 2 p q + q2<

Because the right-hand side of the rule for f was evaluated immediately when it was defined, there is nothing to expand, and k

when defined and it retains the whole expression Expand[k ^2]. When the two functions are subsequently used, f @ p + qD is
is replaced with p+q without expanding the result. On the other hand, the right-hand side of the rule for g was not expanded

just replaced by H p + qL2 , while g@ p + qD is replaced by Expand@H p + qL^2D , which evaluates to p 2 + 2 p q + q 2. The internal
representation of such a definition has the following form:

FullForm@g@k_D := pD

Null

To see the actual head of an expression before it is evaluated, we wrap the expression with Hold and then request its
FullForm

FullForm@Hold@g@k_D := pDD

Hold@SetDelayed@g@Pattern@k, Blank@DDD, pDD

Thus, the symbol := is the infix form of SetDelayed. This makes dramatically clear the distinction between the evaluation
when the rule is given and the evaluation when the rule is used.

When the head of the left-hand side is protected, we can use UpValues to associate the rule with the unprotected argument
of the left-hand side. For example, when we try to differentiate a user-defined definite integral with constant bounds, we
obtain

Clear@int, f, xD

D@int@f@xD, 8x, 0, 1<D, xD

intH0,81,0,0<L @f@xD, 8x, 0, 1<D + f¢@xD intH1,80,0,0<L @f@xD, 8x, 0, 1<D

which is not what we intended. We can either unprotect D and add a rule or use ^:= to assign the definition with int; that is,
Chapter1.nb 7

D@int@x_, a_D, y_D ^:= int@D@x, yD, aD

Now we check to see if Mathematica gives us the answer we want.

D@int@f@xD, 8x, 0, 1<D, xD

int@f¢ @xD, 8x, 0, 1<D

As a second example, we define Cos@n pD = H-1Ln and Sin@n pD = 0 for integer n:

Unprotect@Cos, SinD;
Cos@n_ pD := H-1L n •; IntegerQ@nD
Sin@n_ pD := 0 •; IntegerQ@nD
Protect@Cos, SinD;

If we declare m to be an integer (here UpValues is used),

IntegerQ@mD ^= True;

then

8Cos@m pD, Sin@m pD<

8H-1L m, 0<

Ÿ 1.2.2 Local Rules

These rules are applied to an expression using the operation •. (means ReplaceAll) or ••. (means ReplaceRepeated in
Local rewrite rules (lhs->rhs or lhs:>rhs) are useful for making substitutions without making the definitions permanent.

Mathematica).

Ÿ 1.2.2.1 -> rules

We can use Solve to obtain the solutions of some algebraic equations. For example,

eq1 = u2 - 3 u + 2 == 0;
sol1 = Solve@eq1, uD

88u ® 1<, 8u ® 2<<

We can check the result by substituting sol1 into eq1 and obtain

eq1 •. sol1

8True, True<

The output of Solve is a list of lists of local rules. Thus, u -> 1 is a local rule, which is the analog of the global rule u = 1. The
rule is applied to an expression by using /. , so the expression eq1/.sol1 means "use the rewrite rules u -> 1 and u -> 2 in
8 Chapter1.nb

eq1". The result of this is 80 == 0, 0 == 0<, which is then evaluated as {True, True}. The usual form of the right-hand side of
/. is a list of local rules for some of the symbols that appear on the left-hand side.

u v w •. 8u -> 3, w -> 5<

15 v

If the right-hand side is a list of lists of local rules, then the result is a list of modified expressions, one for each substitution
in the list.

u v w •. 88u -> 3, w -> 5<, 8v -> 2, w -> 3<<

815 v, 6 u<

Local replacement rules also work on expressions which have a Hold wrapped around them. For example, without a Hold,
¶x Cos@xD immediately evaluates to -Sin@xD.

¶x Cos@xD

-Sin@xD

However, the rewrite rule goes under the Hold to make the replacement.

Hold@¶x Cos@xDD •. Cos -> Sin

Hold@¶x Sin@xDD

Local rules with "->" share with "=" rules the property that their right-hand sides are evaluated at the time they are defined.

Ÿ 1.2.2.2 :> rules

The local analog of a ":=" rule is a ":>" rule; i.e., a local rule that evaluates its right-hand side only when it is used. For
example,

expr1 = Exp@I Hw2 - w1 L T0 D;

Clear@aD

expr1 •. Exp@a_D -> Exp@Expand@aDD

EI T0 H-w1 +w2L

8expr1 •. Exp@a_D :> Exp@Expand@aDD, expr1 •• ExpandAll<

8E-I T0 w1 +I T0 w2 , E-IT0 w1+I T0 w2 <

Chapter1.nb 9

where the immediate rule (lhs->rhs) fails to do the expansion. The left-hand sides of the local rules involve patterns rather

equals Exp@aD. Hence, when this is used with a equal to IHw2 - w1L T0 , we get the result EI T 0 H-w1 +w2 L . On the other hand, the
than just symbols. The difference is that the immediate rule for Exp@a_D replaces it by the evaluation of Exp@Expand@aDD which

IHw2 - w1L T0 , gives E-I T 0 w1 +I T 0 w2 . Alternatively, for this simple example, we can use ExpandAll to obtain the same result.
delayed rule (lhs:>rhs) for Exp@a_D replaces it by the unevaluated Exp@Expand@aDD which, when used with a equal to

We can check how these expressions are represented internally.

FullForm@Hold@p •. q -> rDD

Hold@ReplaceAll@p, Rule@q, rDDD

FullForm@Hold@p •. q :> rDD

Hold@ReplaceAll@p, RuleDelayed@q, rDDD

Thus, /. is the infix form of ReplaceAll, the arrow -> is the infix form of Rule, and the arrow :> is the infix form of RuleDe-
layed, corresponding to Set and SetDelayed for = and :=.

Ÿ 1.2.2.3 •. and ••.

There is another form of •. given by ••. which applies a local rule repeatedly until the expression no longer changes. Internally,
••. is represented by

p ••. q -> r •• FullForm •• HoldForm

ReplaceRepeated@p, Rule@q, rDD

An example of the difference between •. and ••. follows. This example uses a list of rules rather than just a single rule. When a

following example, the right-hand side of the •. expression consists of a list of two rules for the same symbol, fac. This list is
list of rules is applied to a single expression, then the rule for each symbol is tried from the left until a match is found. In the

searched from the left until a pattern is found that matches the left-hand side of the •. expression. In the first case using •., as
soon as a match is found, the evaluation is finished. In the second case using ••., the rules are tried repeatedly from the left
on the output of the previous evaluation until no matches are found.

fac@5D •. 8fac@1D -> 1, fac@i_D -> i fac@i - 1D<

5 fac@4D

fac@5D ••. 8fac@1D -> 1, fac@i_D -> i fac@i - 1D<

120

In the first case, the left-hand side of the rule fac@1D -> 1 does not match anything in fac@5D, but fac@i_D -> i fac@i - 1D does
with i_ equal to 5, so the output is 5 fac@4D. In the second case, the left-hand side of the rule fac@i_D -> i fac@i - 1D continues to
match a part of the existing expression until 120 fac@1D is obtained. Then the left-hand side of the rule fac@1D -> 1 matches,
leading to 120 * 1, which simplifies to 120 where neither rule matches, so the output is 120.
10 Chapter1.nb

specific rule, fac@1D = 1, first. However, in a list of local rules, applied with ••., we are completely responsible for the ordering.
If such rules are given globally, then the order in which they are given does not matter since Mathematica will put the more

Thus, the following gives the wrong answer:

fac@5D ••. 8fac@i_D -> i fac@i - 1D, fac@1D -> 1<

If several local rules are given for different symbols, then these rules are applied simultaneously. For example,

8u, v, w< •. 8u -> v, v -> w, w -> s<

8v, w, s<

If the substitutions are carried out sequentially, then the results are quite different.

8u, v, w< •. 8u -> v< •. 8v -> w< •. 8w -> s<

8s, s, s<

In particular, this means that variables can be interchanged without introducing an intermediate temporary variable.

8u, v< •. 8u -> v, v -> u<

8v, u<

Ÿ 1.2.3 Pattern Matching

Mathematica is a language based on pattern matching. Patterns refer to the structure of expressions. The symbol _ by itself,
without any symbol on the left, can be used to describe a pattern. (Recall that the FullForm of _ is Blank@D ). For instance,
the expression _ ^ _ matches anything of the form x^ y, where x and y are any (dummy) expressions.

Clear@a, xD

g1@_, _ ^ _D := u;
8g1@a, a ^aD, g1@a, b^ cD<

8u, u<

A pattern of the form x_ is matched by any expression, which is then used for evaluating the right-hand side.

g2@x_^ y_, z_D := u@x, y, v@zDD;

8g2@a^ a, aD, g2@a^ b, cD<

8u@a, a, v@aDD, u@a, b, v@cDD<

Chapter1.nb 11

g3@x_^ y_, x_D := u@y, v@xDD;

8g3@a^ a, aD, g3@a^ b, cD, g3@a^ b, aD<

8u@a, v@aDD, g3@ab , cD, u@b, v@aDD<

A pattern of the form x_h is matched by any expression whose head is h.

g4@x_^ y_Integer, z_D := u@x, y, v@zDD;

8g4@a^ a, aD, g4@a^ 3, arg2D<

8g4@aa , aD, u@a, 3, v@arg2DD<

If we give a rule for an expression involving two separate arguments with underscores, then we are constructing a function
of two variables. Such a function only works if it is given exactly two arguments.

g5@x_, y_D := x + y;
8g5@aD, g5@a, bD, g5@a, b, cD<

8g5@aD, a + b, g5@a, b, cD<

Besides rule-schemes using a single underscore_, there are rule-schemes using double and triple underscores. A double
underscore, __, is matched by any number of arguments, excluding zero arguments, separated by commas, while a triple
underscore, ___, is matched by any number of arguments, including zero arguments. The form x__ means a sequence of one
or more arguments, named x, and x__Head means a sequence of one or more arguments, named x, with the head of each
argument is Head. Similarly, the form x___ means zero or more arguments, named x, and x___Head means zero or more
arguments, named x, with the head of each argument is Head.

g6@x__D := Length@8x<D;
8g6@D, g6@aD, g6@a, bD, g6@a, b, cD<

8g6@D, 1, 2, 3<

g7@x___D := Length@8x<D;
8g7@D, g7@aD, g7@a, bD, g7@a, b, cD<

80, 1, 2, 3<

Default values and double and triple underscores are important techniques, which give optional number of arguments to
functions; that is, a variable number of arguments can be given to such functions. Alternatively, there is another rule in
which a specific argument can be optional.

g8@x_^ y_, z_D := u@x, y, zD;

8g8@a^ b, cD, g8@a, cD<

8u@a, b, cD, g8@a, cD<

12 Chapter1.nb

Because a = a ^1, one might think that g8@a, cD should match the pattern g8@a ^1, cD, which would mean that it should be
rewritten as u@a, 1, cD, but of course Mathematica cannot guess that this is what we intend. However, there is a provision to
take care of such default values. To assign a default value v to a pattern, one can write _ : v. So, the effect we wanted to
achieve is given by the form

g9@x_^ Hy_: 1L, z_D := u@x, y, zD;

8g9@a^ b, cD, g9@a, cD<

8u@a, b, cD, u@a, 1, cD<

In this case, the default value 1 for the exponent is the natural and obvious choice, and Mathematica knows this. It has
standard built-in default values for a number of such positions. The notation _. tells Mathematica to use the built-in default
value. Note the almost invisible period after the underscore. Thus, the effect we wanted at the beginning is given by a tiny
modification of the original form.

g10@x_^ y_., z_D := u@x, y, zD;

8g10@a^ b, cD, g10@a, cD<

8u@a, b, cD, u@a, 1, cD<

There is still another way that optional arguments occur in Mathematica. Some functions, such as Plot, can take named
optional arguments such as AspectRatio -> 1. By incorporating such functions into our own definitions, we can use these
named optional arguments too. Consider an example of a plotting function.

plotwithSin@func_, var_, opts___D := Plot@8Sin@varD, func@varD<,

8var, 0, 2 Pi<, opts, PlotStyle -> 8RGBColor@1, 0, 0D, RGBColor@0, 0, 1D<D;

plotwithSin@Cos@2 #D &, xD;

0.5

1 2 3 4 5 6
-0.5

-1
Chapter1.nb 13

plotwithSin@Cos@2 #D &, x, AspectRatio -> 1D;

0.5

1 2 3 4 5 6

-0.5

-1

In an expression of the form g[x_^n_Integer, z_], the patterns are named with the symbols x, n, and z, but there is no name for
the whole compound pattern x_^n_Integer. There is a way to give names to such compound patterns so that they can be
referred to directly on the right-hand side. The syntax consists of a name followed by a colon followed by the compound
pattern.

g11@expr : x_^y_Real, z_D := z expr;

8g11@Ha + b - 2L^ 0.5, 2D, g11@Ha + b - 2L^ 3, 2D<

82 H-2 + a + bL0.5, g11@H-2 + a + bL3 , 2D<

Ÿ 1.2.4 Using Patterns in Rules

Patterns play an important role in both global and local rules.

Ÿ 1.2.4.1 Patterns in Global Rules

For example, it is very easy to give rewrite rules for differentiating polynomials of one variable.

diff@x_^ n_., x_D := n x^Hn - 1L;

diff@a_ + b_, x_D := diff@a, xD + diff@b, xD;

Notice the default value for n in the first rule. Try it out on some typical functions.

8diff@y, yD, diff@z ^2.1, zD, diff@u^ 2 + v ^3, uD<

81, 2.1 z1.1 , 2 u + diff@v3 , uD<

But notice that diff does not know what to do with a constant times x, or just a constant for that matter, and we have no
obvious way as yet to teach it what to do.
14 Chapter1.nb

8diff@3, xD, diff@3 x, xD<

8diff@3, xD, diff@3 x, xD<

We could try the following: first, give a rule for products.

diff@a_ b_, xD := a diff@b, xD + b diff@a, xD

Using this for a x gives

diff@a x, xD

a + x diff@a, xD

The program does not know that a is supposed to be a constant, so we have to tell it that explicitly, with a last rule.

diff@a, xD = 0;

Then it gives the "correct" answer.

diff@a x, xD

However, this is not very satisfactory. We would like some general way to say that a is not a function of x. Section 1.2.5
below will continue this discussion.

Ÿ 1.2.4.2 Patterns in Local Rules

We will encounter this type of application quite often in the following chapters. In the following example, we show how to
change the appearance of a matrix. Using Array to make a matrix with indexed entries, we then use a local rule to display the
indices as subscripts.

Array@x, 83, 5<D •. x@i__D -> xi •• MatrixForm

i
j x1,1 x1,2 x1,3 x1,4 x1,5 y
z
j
j
j x2,1 x2,2 x2,3 x2,4 x2,5 z
z
z
j
j z
z
k x3,1 x3,2 x3,3 x3,4 x3,5 {

where the two consecutive underscores next to i on the left-hand side of the rule represent one or more arguments.

We can also use pattern matching to replace the coefficients of various terms any way we want. For example,

Clear@fD

eq1 = u² @tD + u@tD f¢ @x0D + €€€€ u@tD2 f² @x0D + €€€€ u@tD3 fH3L @x0D == 0;
1 1
2 6

eq2 = eq1 •. _. u@tD^ b_. -> kb u@tD^ b

k1 u@tD + k2 u@tD2 + k3 u@tD3 + u² @tD == 0

Chapter1.nb 15

sponding control parameters cpSearch and cpFix, such as 88x1 , x2, x3 , x4, cpSearch<, cpFix<. If we want to switch these two
Sometimes we may obtain, in a bifurcation analysis, a data set, which includes a set of bifurcation points xi and the corre-

parameters and then use the same data format as our initial input to a program, we have at least two choices. First, we take
apart the data list and then assemble them:

data1 = 88x1 , x2, x3 , x4 , cp1<, cp2<;

8Append@Drop@data1@@1DD, -1D, data1@@2DDD, data1@@1, -1DD<

88x1 , x2, x3, x4 , cp2<, cp1<

Second, we simply apply a local rule using patterns to the data list as a whole:

data1 •. 88a__, b_<, c_< -> 88a, c<, b<

88x1 , x2, x3, x4 , cp2<, cp1<

Ÿ 1.2.5 Restricting Pattern Matching with Predicates

So far, all of the rules we have considered have been "context free" rewrite rules. Whenever the pattern is matched, the
rewriting is carried out. There can be a restriction on the head of the matching expression included in the pattern. However,
there are also conditional rewrite rules, which are only applied when some condition is satisfied. First, we have to discuss
predicates (functions that return the value True or False) in Mathematica because all of the conditions will always be
expressed in terms of them. All built-in predicates that are defined for all expressions end with Q. It is easy to display all of
them.

? *Q

ArgumentCountQ LinkConnectedQ OrderedQ

AtomQ LinkReadyQ PartitionsQ
DigitQ ListQ PolynomialQ
EllipticNomeQ LowerCaseQ PrimeQ
EvenQ MachineNumberQ SameQ
ExactNumberQ MatchLocalNameQ StringMatchQ
FreeQ MatchQ StringQ
HypergeometricPFQ MatrixQ SyntaxQ
InexactNumberQ MemberQ TrueQ
IntegerQ NameQ UnsameQ
IntervalMemberQ NumberQ UpperCaseQ
InverseEllipticNomeQ NumericQ ValueQ
LegendreQ OddQ VectorQ
LetterQ OptionQ

Specifically requesting information on our function gives the usage message.

? OddQ

OddQ@exprD gives True if expr is an odd integer, and False otherwise.

Predicates are used to control pattern matching. In general, they are applied using •; , which is the infix form of Condition.
16 Chapter1.nb

FullForm@Hold@x •; yDD

Hold@Condition@x, yDD

However, the position of the predicate in an expression allows it to be used in different ways. If the predicate is placed at the
end of a global rule definition, it will restrict the application of the rule. For instance, we can use predicates to extend our
definition of differentiation in Section 1.2.4.1 to deal with arbitrary polynomials in a very natural way by adding a single
conditional rule.

Clear@diffD

diff@x_^ n_., x_D := n x^Hn - 1L;

diff@a_ + b_, x_D := diff@a, xD + diff@b, xD;
diff@a_ b_, x_D := a diff@b, xD + b diff@a, xD;
diff@a_, x_D := 0 •; FreeQ@a, xD

Now constants and products are handled properly

diff@a@xD + b x ^2 + c, xD

2 b x + diff@a@xD, xD

Using rule•;Predicate restricts the rule to those situations in which the predicate evaluates to True; that is, to those expres-
sions belonging to the type given by the predicate. An unrestricted rule is the same as a conditional rule where the predicate
always equals True.

If a predicate is immediately applied after a pattern, it will restrict the pattern matching rather than the rule application. For
instance,

fac1@1D = 1; fac1@n_D := n fac1@n - 1D

This will work perfectly well if positive integers are given as arguments.

fac1@5D

120

However, if another argument is given, then it will fail badly.

8fac1@-3D, fac1@endlessD<;

$RecursionLimit::reclim : Recursion depth of 256 exceeded.

$RecursionLimit::reclim : Recursion depth of 256 exceeded.

$RecursionLimit::reclim : Recursion depth of 256 exceeded.

General::stop : Further output of $RecursionLimit::reclim will be suppressed during this calculation.

Chapter1.nb 17

A very large output is suppressed. What happens, of course, in these cases is that the value 1 is never encountered as an
argument, so the function keeps calling itself recursively until the built-in recursion limit is reached. This bad behavior can be
corrected by using a conditional rule.

fac2@1D = 1; fac2@n_D := n fac2@n - 1D •; n > 1;

8fac2@5D, fac2@-2D, fac2@somethingD<

8120, fac2@-2D, fac2@somethingD<

One can also use the form _ ?Predicate, which restricts the pattern to something for which the predicate evaluates to True.

fac3@1D = 1; fac3@n_ ?PositiveD := n fac3@n - 1D;

fac4@1D = 1; fac4@n_ •; Positive@nDD := n fac4@n - 1D;

8fac3@5D, fac3@-2D, fac4@5D, fac4@-2D<

8120, fac3@-2D, 120, fac4@-2D<

In n_ ?Positive , Positive is a pure function, while in the form using •; , the condition is the value of the predicate for the name
of the pattern. In either case, Positive or Positive[n] is a positive test in the sense that the pattern is matched and the rule
applied only if the test succeeds. Now try fac4 on a real number and see what happens.

[email protected]

8.9856 [email protected]

The rule is applied until the negative value -0.4 is reached, where the condition fails so fac4[-0.4] is returned in unevaluated
form. Of course, the real problem is that we only intend the function to apply to integers. But this additional restriction can
easily be added.

fac5@1D = 1; fac5@n_Integer ?PositiveD := n fac5@n - 1D;

8fac5@5D, [email protected], fac5@-2D<

8120, [email protected], fac5@-2D<

The predicate that appears after •; or ? can also be a user defined expression.

test@x_Integer? HH# > 2L &LD := x + 1

8test@1D, test@2D, test@3D, test@4D, test@5D<

8test@1D, test@2D, 4, 5, 6<

The form ? Predicate can only be used after single slots, but the form •; Predicate can be used after any pattern, simple or
compound. For instance,

f1@a_, b_D •; EvenQ@a + bD := a ^ b;

f1@a_, b_D •; OddQ@a + bD := a ^H- bL;
18 Chapter1.nb

8f1@1, 3D, f1@2, 3D, f1@3, 3D<

91, €€€€ , 27=

1
8

Predicates also play an important role in manipulating lists.

Select@Range@-5, 5D, PositiveD

81, 2, 3, 4, 5<

Cases@Range@-5, 5D, _? PositiveD

81, 2, 3, 4, 5<

Cases@Range@-5, 5D, a_ •; a > 0D

81, 2, 3, 4, 5<

There is also another form of Cases in which an operation is applied to the entries that are selected.

Cases@Range@-5, 5D, Ha_ •; a > 0L :> Sqrt@aDD

• ! ! • !!! • !!!
91, 2 , 3 , 2, 5 =

The following example was the 1992 Mathematica programming competition question. The problem is to write a function
called maxima that starts with a list of numbers and constructs the sublist of the numbers bigger than all previous ones from
the given list. For instance, maxima[{3, 6, 4, 2, 8, 7, 9}] should return {3, 6, 8, 9}. The winning entry used a pattern with a
condition in a local rule.

maxima@list_ListD := list ••. 8a___, x_, y_, b___< •; y <= x -> 8a, x, b<

maxima@83, 6, 4, 2, 8, 7, 9<D

83, 6, 8, 9<

Ÿ 1.2.6 Pure Functions

we need to refer to this function by name, we can either define a rule for f in the form f @x_D := expr or use the object
Given an expression expr involving a variable x, we can think of expr as describing a function with x being the argument. If

Function[x, expr]. Either f or Function[x, expr] is a name for this function, and we can use the two interchangeably. To apply
them to an argument, we write f[arg] in the familiar way or in the so-called pure function form Function[x, expr][arg].

f@x_D := 1 + x^2
g = Function@x, 1 + x^2D;

8f@2D, g@2D<

85, 5<
Chapter1.nb 19

The name of the formal argument in a pure function does not matter. Function[x, x^2] is the same function as Function[y,
y^2]. This fact is easy to see if you apply such a function to an argument, the result is the same in both cases.

8Function@x, x^ 2D@aD, Function@y, y^2D@aD, Sin@aD<

8a2 , a2, Sin@aD<

Note that Sin and all built-in functions, with no arguments, are pure functions. Because the names of the variables in a pure
function do not matter, Mathematica denotes these variables with symbols. The symbols #1, #2, ... are used for the first,
second, ... variable. The internal form of #i is Slot[i]. If we use these forms, then we do not give the arguments of Function[ ]
in order to declare the names of the variables. Thus, instead of Function[x, x^2] we simply write Function[#1^2]. A conve-
nient abbreviation for #1 is #, so we can simplify our example a bit more and write Function[#^2]. Finally, there is a postfix
operator & for Function. That is, body& is the same as Function[body]. Using this operator, we arrive at the shortest form
#^2 & that our example can take. We have used this form quite frequently in this book. Let us see the internal form of this
short form of a pure function.

FullForm@#^ 2 &D

Function@Power@Slot@1D, 2DD

The operator & has a very low priority, just above assignment. Therefore, x = body & is understood as x = Function[body],
but x -> body & is interpreted as Function[x -> body]. If we want the right-hand side of the rule to be the pure function, we
use x -> (body &). Beware of x -> (body) &. Another case where the low priority of pure functions requires the use of
parentheses is in predicates for patterns: x_?body& is not the same as x_?(body&). The latter is usually correct.

1 + # + # ^2 & @3D

13

The whole expression to the left of & is part of the body of the pure function. No parentheses are necessary. In other
situations, parentheses are necessary.

h@aD •. h -> H#^ 2 &L

a2

Parentheses around the pure function are needed here; otherwise, the whole rule would be considered a part of the body of
the function.

Pure functions written in the form with # and & are very concise and useful, especially, in functional operations, such as
Map, Apply, Thread, and iteration functions, such as FoldList. For example,

eq1 = ¶t,t u@x, tD • a + u@x, tD == F@xD • a

uH0,2L @x, tD F@xD

u@x, tD + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ == €€€€€€€€€€€€
a a

Clear@lhs2D
20 Chapter1.nb

eq2 = lhs2 == rhs2;

eq3 = lhs3 == rhs3;

We can multiply both sides of eq1 by a and use Thread and obtain

eq1 •• Thread@a #, EqualD & •• ExpandAll

a u@x, tD + uH0,2L @x, tD == F@xD

Moreover, we can add eq2 to eq3 by

eq2 + eq3 •• Thread@#, EqualD &

lhs2 + lhs3 == rhs2 + rhs3

Ÿ 1.2.7 Map and Apply

In Mathematica, two important commands that take functions as arguments are Map[ ] and Apply[ ]. The various versions of
Map act on the arguments of an expression, while Apply acts only on its head.

Ÿ 1.2.7.1 Mapping Functions onto Expressions

The operation Map[f, list] or in infix form f •ž list applies the function f to each element of the list list. The second argument

expression h@ f @e1D, f @e2 D, ..., f @en DD.

of Map[ ] need not be a list, however. Any expression of the form h[e1 , e2 , ..., en ] will do. The result of the mapping is the

Clear@fD

8Map@f, a + b + cD, # ^2 & •ž 81, 2, 3<<

8f@aD + f@bD + f@cD, 81, 4, 9<<

Map[f, expr, levelspec] has an optional third argument that specifies the levels at which to map. The default level is {1}; that
is, at the elements of expr. In a matrix, the entries are at level 2 (there are two levels of lists). If we want to map a function at
these entries, we can use Map[f, matrix, {2}].

expr1 = 88a1,1, a1,2 <, 8a2,1 , a2,2 <<;

Map@entry, expr1, 82<D •• MatrixForm

i entry@a1,1 D entry@a1,2 D z
j y
k entry@a2,1 D entry@a2,2 D {

This is a very powerful facility and is one of our main tools in manipulating expressions. The purpose of the operations
based on Map is to make it possible to treat lists as wholes. For instance, a really poor way to square the entries in a list is as
follows:
Chapter1.nb 21

list1 = 8a, b, c<;

Table@list1@@iDD^ 2, 8i, Length@list1D<D

8a2 , b2, c2<

The term list1[[i]] tears apart the original list by extracting its parts one at a time, ^ 2 squares each part, and then Table
reassembles the parts into a new list. Map, however, allows us to treat mathematical structures as wholes so that we do not
tear them apart and rebuild them again.

Map[ ] always maps the function at all elements of the given levels. The command MapAt[f, expr, poslist ] allows you to apply
a function at selective positions.

MapAt@G, a b + c d + e f, 82, 1<D

a b + e f + d G@cD

expr2 = a + b • a + c Exp@a + 1D

b
a + €€€€ + c E1+a
a

If we want to map a function f at all occurrences of a, we first find all the positions of a in our expression.

Position@expr2, aD

881<, 82, 1, 1<, 83, 2, 2, 2<<

This list of positions is in the right form for MapAt and can be used directly.

MapAt@f, expr2, %D

b
c E1+f@aD + €€€€€€€€€€€€ + f@aD
f@aD

This particular example could have been done more easily with a replacement rule.

expr2 •. a -> f@aD

b
c E1+f@aD + €€€€€€€€€€€€ + f@aD
f@aD

In some applications, the function to be performed on an element ei may depend on its position i. The operation Map-
Indexed[f, {e1 , e2, ..., en }] behaves essentially like Map[ ], but it passes the position of each element as a second argument to
f. The resulting expression is

MapIndexed@1 + # ^2 &, 8a, b, c, d<D

81 + a2, 1 + b2 , 1 + c2 , 1 + d2 <

MapIndexed@1 + # ^2 &, 88a, b<, 8c, d<<, 82<D

881 + a2, 1 + b2<, 81 + c2 , 1 + d2 <<

22 Chapter1.nb

Therefore, the function that is mapped must be a function of two arguments. The second argument is a position (note the list
braces) that can be used to modify the operation performed on the first argument.

MapIndexed@#1^ First@#2D &, 8a, b, c, d, e<D

8a, b2 , c3 , d4, e5 <

Ÿ 1.2.7.2 Apply

The rule Apply[h, expr] or h žž expr replaces the head of expr by h.

a + b + c •• FullForm

Plus@a, b, cD

Replacing the head Plus by Times gives the product of the three terms in the sum above.

8Apply@Times, a + b + cD, Times žž Ha + b + cL<

8a b c, a b c<

With a third argument, Apply[h, expr, levelspec] replaces heads in the parts of expr described by levelspec by h.

Apply@b, 88a@1, 1D, a@1, 2D<, 8a@2, 1D, a@2, 2D<<, 82<D

88b@1, 1D, b@1, 2D<, 8b@2, 1D, b@2, 2D<<

This simple definition finds the average of a list of numbers

average@list1_ListD := Apply@Plus, list1D • Length@list1D

average@8a, b, c<D

€€€€ Ha + b + cL
1
3

Note that computing the average of a list in this way requires no do-loops or knowledge of the length of the list. Apply is
frequently used if one first wants to prepare a number of ingredients and then apply some operation to them. The ingredients
can be held in a list until they are ready and then the head of the list is changed to the appropriate operation by using Apply.

To generate more concise notations for the outputs created in the following chapters, we can introduce displayRule to meet
our needs. For example,

H2,0,0L H1,1,0L
output = u2 @T0 , T1, T2D + w2 u2 @T0, T1 , T2 D == 2 u1 @T0, T1 , T2 D - d u1 @T0 , T1 , T2D2 ;

displayRule = 9Derivative@a__D@ui_ D@__D :>

#2@@1DD-1 &, 8a<E, ui E, ui_@__D -> ui =;

SequenceFormATimes žž MapIndexedAD #1

output •. displayRule

D20 u2 + w2 u2 == 2 HD0 D1 u1L - d u21

Chapter1.nb 23

Ÿ 1.2.8 Nest and Fold

Nest and its related operations NestList and FixedPoint apply a function to its argument many times. Nest[function, x, n]
applies function to x and repeats the application n times

NestAH1 + #L2 &, x, 3E

J1 + H1 + H1 + xL2L N
2 2

NestListAH1 + #L2 &, x, 3E

9x, H1 + xL2, H1 + H1 + xL2 L , J1 + H1 + H1 + xL2 L N =

2 2 2

Here is an example producing a simple continued fraction.

Nest@1 • H1 + #L &, a, 3D

1
€€€€€€€€€€€€€€€€€€€€€€
1
1 + €€€€€€€€ 1€€€€
1+ €€€€€€€€
1+a€

An operation that is closely related to Nest is FixedPoint, which nests its operation until there is no change. For instance,
everyone is familiar with what happens if the Cos key on a pocket calculator is pushed repeatedly. In principle, FixedPoint is
what happens if it is pushed forever.

8Nest@Cos, 0.3, 5D, Nest@Cos, 0.3, 10D, FixedPoint@Cos, 0.3D<

80.784436, 0.732698, 0.739085<

Actually, FixedPoint stops after machine accuracy is achieved. We can use the option SameTest to get some control of
FixedPoint.

FixedPoint@Cos, 0.3, SameTest -> HAbs@#1 - #2D < 10^ -5 &LD

0.739089

The second pair of functions, Fold and FoldList, do something similar to Nest and NestList, but for functions of two
variables.

Fold@f, u, 8x, y, z<D

f@f@f@u, xD, yD, zD

FoldList@f, u, 8x, y, z<D

8u, f@u, xD, f@f@u, xD, yD, f@f@f@u, xD, yD, zD<
Chapter 2
The Duffing Equation
Off@General::spell1D

To symbolize some notations, we need to load the package

Needs@"Utilities`Notation`"D

à 2.1 The Duffing Equation

The free oscillations of many conservative systems having a single degree of freedom are governed by an equation of
the form

eq21a = x² @tD + f@x@tDD == 0;

H* An equation is set and then assigned to eq21a *L

where f is a nonlinear function of x@tD. Here, x² @tD is the acceleration of the system, whereas f @x@tDD is the restoring force.
Let x = x0 be an equilibrium position of the system. Then x0 is a constant and hence f @x0 D = 0. In order for Mathematica
to interpret x0 as a constant, we need to symbolize it. Otherwise, if we replace x with an expression g@tD, Mathematica will
replace x0 with g@tD sub 0 and not treat it as a constant. For example, replacing x with Sin@tD in the expression x2 + x0 yields

x2 + x0 •. x -> Sin@tD

Sin@tD2 + HSin@tDL0

Clearly, x0 is treated as a variable x with subscript 0, which is wrong. Therefore, we need to symbolize x0 as (enter and
select x0 followed by clicking Symbolize@ŸD from NotationPalette)

Symbolize@x0 D;
H* Note that symbol form is also required in certain built-in functions *L

Then, Mathematica treats the symbolized expression x0 as a constant. Consequently, replacing x with Sin@tD in the
expression x2 + x0 yields

x2 + x0 •. x -> Sin@tD

x0 + Sin@tD2

which is correct.

Next, we assume that f is an analytic function of x at x = x0 and expand it in a Taylor series around x0 as
26 Chapter2.nb

fexp = Series@f@x@tDD, 8x@tD, x0, 3<D

f@x0 D + f¢ @x0D Hx@tD - x0L + €€€€ f²@x0 D Hx@tD - x0 L2 + €€€€ fH3L @x0 D Hx@tD - x0 L3 + O@x@tD - x0 D4
1 1
2 6

where terms up to cubic are retained. Using Normal to truncate the higher-order terms from fexp, we have

fpoly = fexp •• Normal

f@x0 D + H-x0 + x@tDL f¢@x0 D + €€€€ H-x0 + x@tDL2 f²@x0 D + €€€€ H-x0 + x@tDL3 fH3L@x0 D
1 1
2 6

Substituting fpoly for f @x@tDD and using the equilibrium condition f @x0D = 0, we obtain

temp = eq21a •. 8f@x@tDD -> fpoly< •. f@x0D -> 0

H-x0 + x@tDL f¢@x0 D + €€€€ H-x0 + x@tDL2 f²@x0 D + x² @tD + €€€€ H-x0 + x@tDL3 fH3L@x0 D == 0
1 1
2 6

Now, we introduce the transformation x@tD ® x0 + u@tD in temp to obtain the final form of the equation that will be used in the

x ® Hx0 + u@#D &L; otherwise, Mathematica will not transform the derivatives of x@tD in terms of the derivatives of u@tD unless
subsequent analysis. To accomplish this, we need to express this transformation in a pure function form as

we explicitly define the rules for those derivatives; that is, it will not replace x¢ @tD with u ¢ @tD and x²@tD with u ² @tD by simply
substituting x@tD ® x0 + u@tD. Moreover, it will not replace x@pD with x0 + u@ pD as desired. To see the difference between these
two transformations, we compare the following two statements:

8x@tD, x@pD, x¢ @tD, x² @tD< •. x@tD -> x0 + u@tD

8x0 + u@tD, x@pD, x¢ @tD, x² @tD<

8x@tD, x@pD, x@a, bD, y@tD, x¢ @tD, x² @tD< •. x -> Hx0 + u@#D &L

8x0 + u@tD, x0 + u@pD, x0 + u@aD, y@tD, u¢@tD, u² @tD<

We note that the pure function form can be used for any expression with the head x, especially, for the derivatives. Conse-
quently, we introduce this transformation in pure function form into temp and obtain

eq21b = temp •. x -> Hx0 + u@#D &L

u@tD f¢ @x0D + €€€€ u@tD2 f² @x0D + u²@tD + €€€€ u@tD3 fH3L @x0D == 0
1 1
2 6

To simplify the notation in eq21b, we denote the coefficient of u@tDb by k b. The result is

eq21c = eq21b •. _. u@tD^ b_. -> k b u@tD^ b

k1 u@tD + k2 u@tD2 + k3 u@tD3 + u² @tD == 0

We relate the k b to the derivatives of the function f @xD by using the built-in function Cases as
Chapter2.nb 27

CasesAeq21b, a_. u@tD b_. -> 8k b -> a<, 82<E •• Flatten

9k1 ® f¢@x0 D, k2 ® €€€€ f²@x0 D, k3 ® €€€€ fH3L @x0D=

1 1
2 6

Most of this chapter is devoted to the special case of cubic nonlinearities. Thus, we set the coefficient of the quadratic term
in eq21c equal to zero and obtain

eq21d = eq21c •. k2 -> 0

k1 u@tD + k3 u@tD3 + u²@tD == 0

which is called the Duffing equation. We note that k 1 and k 3 may be positive or negative.

It is a good practice to nondimensionalize the governing equations before treating them with perturbation methods. To this
end, we nondimensionalize u and t using a characteristic length U and a characteristic time T of the motion and denote the
nondimensional variables by an asterisk. In nondimensional form, eq21d becomes

eq21e = eq21d •. u -> HU u*@# • TD &L •. t -> T t*

U Hu* L² @t* D
U k1 u*@t* D + U3 k3 u* @t*D3 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ == 0
T2

Next, we multiply the left-hand side of eq21e (using eq21e[[1]]) by T 2 •U so that the coefficient of u *² @t*D is unity. The result
is

eq21f = Ieq21e@@1DD T2 • U •• ExpandM == 0

T2 k1 u* @t*D + T2 U2 k3 u* @t*D3 + Hu* L² @t*D == 0

We choose T so that the linear natural frequency of the system is unity; that is, we let k 1 T 2 = 1, for ease of notation, we let
e = k 3 T 2 U 2, drop the asterisk, and rewrite eq21f as

DuffingEq = eq21f •. 9k1 T2 -> 1, k3 T2 U2 -> e= •. anything_* -> anything

u@tD + e u@tD3 + u²@tD == 0

We note that e is a nondimensional quantity, which is a measure of the strength of the nonlinearity.

à 2.2 Straightforward Expansion

To solve the DuffingEq, we augment it with the initial conditions

ic = 8u@0D == u0, u¢ @0D == v0<;

The solution u of the initial-value problem (IVP), consisting of the DuffingEq and associated initial conditions ic, is a func-
tion of the independent variable t and the parameter e. Next, we determine an approximation of the IVP for weak nonlinearity;
that is, for small but finite e.
28 Chapter2.nb

First, we set e equal to zero in the DuffingEq, augment the result with the initial conditions by using the Mathematica
command Join, and obtain the following linear IVP:

linearIVP = Join@8DuffingEq •. e -> 0<, icD

8u@tD + u²@tD == 0, u@0D == u0, u¢ @0D == v0<

The solution of this linear IVP can be obtained by using the Mathematica command DSolve. The result is

linearSol = DSolve@linearIVP, u@tD, tD@@1DD

8u@tD ® u0 Cos@tD + v0 Sin@tD<

When e is small but different from zero, the solution of the nonlinear initial-value problem is no longer given by linearSol, but
deviates from it. We try a correction in the form of a power series in e; that is, we expand u@tD in a power series in e as

uExpRule@maxOrder_D := u -> ISumAei ui @#D, 8i, 0, maxOrder<E &M

where u 0 @tD is the solution of the linear problem. For example, to first order, we have

u@tD •. uExpRule@1D

u0 @tD + e u1 @tD

and to third order, we have

u@tD •. uExpRule@3D

u0 @tD + e u1 @tD + e2 u2@tD + e3 u3 @tD

and to fifth order, we have

u@tD •. uExpRule@5D

u0 @tD + e u1 @tD + e2 u2@tD + e3 u3 @tD + e4 u4@tD + e5 u5 @tD

Next, we restrict our discussion to first order; that is, we keep two terms in the power series. Thus, we apply the rule uExp-
Rule[1] to the DuffingEq, expand the result, and obtain

eq22atemp = DuffingEq •. uExpRule@1D •• ExpandAll

u0 @tD + e u0 @tD3 + e u1@tD + 3 e2 u0 @tD2 u1 @tD + 3 e3 u0@tD u1 @tD2 + e4 u1 @tD3 + u²0@tD + e u²1@tD == 0

Since we kept only terms up to OHeL in the expansion of u@tD, we need to neglect terms of order higher than e in eq22atemp for
consistency. To accomplish this, we use the rule en•;n>k ® 0 to discard terms with order higher than ek and obtain

eq22a = eq22atemp •. e n_•;n>1 -> 0

u0 @tD + e u0 @tD3 + e u1@tD + u²0@tD + e u²1@tD == 0

Setting the coefficients of like powers of e in eq22a equal to zero, we obtain the following hierarchy of linear equations:
Chapter2.nb 29

eqEps = CoefficientList@eq22a@@1DD, eD == 0 •• Thread

8u0 @tD + u²0 @tD == 0, u0@tD3 + u1 @tD + u²1 @tD == 0<

which needs to be solved in succession. That is, we first solve the first equation in eqEps to obtain u 0 @tD and then substitute
the result into the second equation in eqEps to obtain a linear nonhomogeneous equation, which can be solved for u 1 @tD.

The general solution of the zeroth-order problem (eqEps[[1]]) can be expressed in Cartesian form as

DSolve@eqEps@@1DD, u0@tD, tD@@1DD

8u0 @tD ® C@2D Cos@tD - C@1D Sin@tD<

Alternatively, we can express the solution of the zeroth-order problem in the polar form u 0 @tD ® a 0 Cos@t + b 0D by adding two
appropriate initial conditions to eqEps[[1]] as

sol0 =
DSolve@8eqEps@@1DD, u0@0D == a0 Cos@b0D, u0 ¢ @0D == -a0 Sin@b0D<, u0 @tD, tD@@1DD •• Simplify

8u0 @tD ® Cos@t + b0 D a0<

where a 0 and b 0 are arbitrary constants. Because eqEps[[1]] represents a linear undamped oscillator, one can write down its
general solution in either Cartesian or polar form without using the Mathematica command DSolve.

Substituting for u 0 @tD in the first-order equation (eqEps[[2]]) and determining the general solution of the resulting equation for
u 1 @tD, we obtain

sol1temp = DSolve@eqEps@@2DD •. sol0, u1@tD, tD@@1DD •• Simplify

9u1 @tD ®

€€€€€€€ H32 HC@2D Cos@tD - C@1D Sin@tDL + H-6 Cos@t + b0D + Cos@3 Ht + b0 LD - 12 t Sin@t + b0 DL a30L=
1
32

We note that the homogeneous part of sol1temp consists of the terms involving the C@iD (the usual case) and Cos@t + b 0 D
(obtained from DSolve in this case). Hence, we can replace them by a 1 Cos@t + b 1D and rewrite sol1temp as

sol1 = u1 @tD ® Hsol1temp@@1, 2DD •. 8C@_D -> 0, Cos@t + _D -> 0<L + a1 Cos@t + b1 D

u1 @tD ® €€€€€€ HCos@3 Ht + b0 LD - 12 t Sin@t + b0 DL a30 + Cos@t + b1 D a1

1
32

Combining the solutions of the zeroth- and first-order problems, we obtain the general solution of the DuffingEq, to first
order, as

sol = u@tD == Hu@tD •. uExpRule@1D •. sol0 •. sol1L

i 1 y
u@tD == Cos@t + b0D a0 + e j €€€€€€€ HCos@3 Ht + b0 LD - 12 t Sin@t + b0DL a30 + Cos@t + b1D a1 z
k 32 {

where a 0 , a 1 , b 0, and b 1 are arbitrary constants.

30 Chapter2.nb

We started with a second-order equation that can support two initial conditions but it appears that we ended up with four
arbitrary constants. It turns out that only two of the constants a 0, a 1, b 0, and b 1 are arbitrary and that the two initial condi-
tions in ic are sufficient to determine all of them. To see this, we first change the Equal form in ic to the Rule form as

icRule = ic •. Equal -> Rule

8u@0D ® u0, u¢ @0D ® v0<

Then, we impose the initial conditions icRule by evaluating sol and its derivative at t = 0 according to

eq22b = 8sol, D@sol, tD< •. t -> 0 •. icRule

9u0 == Cos@b0 D a0 + e i €€ Cos@3 b0 D a30 + Cos@b1D a1 y

j €€€€ z,
k 32 {
1

v0 == -Sin@b0D a0 + e j €€€€€€€ H-12 Sin@b0D - 3 Sin@3 b0DL a30 - Sin@b1 D a1y

i z=
k 32 {
1

We equate the coefficient of each power of e in eq22b and obtain four algebraic equations for the constants a 0, a 1 , b 0 , and
b 1 . The result is

eq22c = Table@Map@Coefficient@#, e, iD &, eq22b, 82<D, 8i, 0, 1<D

98u0 == Cos@b0 D a0 , v0 == -Sin@b0D a0 <,

90 == €€€€€€€ Cos@3 b0 D a30 + Cos@b1D a1 , 0 == €€€€€€ H-12 Sin@b0 D - 3 Sin@3 b0 DL a30 - Sin@b1D a1 ==
1 1
32 32

Thus, once a 0 and b 0 are known from eq22c[[1]], a 1 and b 1 can be calculated from eq22c[[2]].

Alternatively, instead of including the homogeneous solution at each order and consider the arbitrary constants to be
independent of e as done above, one may disregard the homogeneous solution at all orders except the first and consider the
arbitrary constants to depend on e in imposing the initial conditions. With the latter approach, the solution to first order can
be rewritten as

sol •. 8a0 -> a, b0 -> b, a1 -> 0< •• ExpandAll

1 3
u@tD == a Cos@t + bD + €€€€€€€ a3 e Cos@3 t + 3 bD - €€€€ a3 t e Sin@t + bD
32 8

1
€€€ a 3 eCos@3 t + 3 bD - €€€€38€ a 3 t e Sin@t + bD is small compared to the principal term
We note that the correction term €€€€32
a Cos@t + bD, as it is supposed to be, only when e t is small compared with unity. When e t is OH1L, the term that is supposed
to be a small correction becomes the same order as the principal term. Moreover, when e t > OH1L, the "small-correction"

such that e t < OH1L; that is, t < OHe-1 L. Consequently, we say such expansions are nonuniform or breakdown for long times
term becomes larger than the principal term. Hence, the above obtained straightforward expansion is valid only for times

and we call them pedestrian or naive expansions. The reason for the breakdown of the above expansion is the presence
of the term t Sin@t + bD, a product of algebraic and circular terms. Such terms are called mixed-secular terms. In subsequent
sections, we implement four methods that avoid secular terms, and hence yield uniform expansions.
Chapter2.nb 31

à 2.3 The Lindstedt-Poincaré Technique

The breakdown in the straightforward expansion is due to its failure to account for the nonlinear dependence of the
frequency of the system on the nonlinearity. A number of techniques that yield uniformly valid expansions have been
developed. Four of these techniques are discussed in this chapter. We start with the Lindstedt-Poincaré technique in this
section.

To account for the dependence of the frequency w of the system on the nonlinearity, we explicitly exhibit w in the govern -
ing differential equation. To this end, we introduce the transformation t = w t, where w is a constant that depends on e,
and obtain

DuffingEq = u² @tD + u@tD + e u@tD3 == 0;

eq23a = DuffingEq •. u -> Hu@w #D &L •. t -> t • w

u@tD + e u@tD3 + w2 u² @tD == 0

To determine a uniform expansion of the solution of eq23a, we expand both u@t; eD and w@eD in powers of e using uExpRule[k]
and

omgRule@maxOrder_D := w -> w0 + SumAei wi , 8i, maxOrder<E

where w0 is the linear natural frequency of the system. In the present case, the linear natural frequency was normalized to
unity. Hence, we set w0 = 1. The corrections to the linear frequency are determined in the course of the analysis by requiring
the expansion of u to be uniform for all t. Next, we show how one can determine w1 .

Substituting the expansions of u and w into eq23a, using the fact that w0 = 1, setting maxOrder = 1, expanding the result,
and discarding terms of order higher than e, we obtain

eq23b = ExpandAll@eq23a •. uExpRule@1D •. omgRule@1DD •. w0 -> 1 •. e n_•;n>1 -> 0

u0 @tD + e u0 @tD3 + e u1@tD + u²0@tD + 2 e w1 u²0 @tD + e u²1 @tD == 0

Equating coefficients of like powers of e on both sides of eq23b yields

eqEps = CoefficientList@eq23b@@1DD, eD == 0 •• Thread

8u0 @tD + u²0 @tD == 0, u0@tD3 + u1 @tD + 2 w1 u²0@tD + u²1@tD == 0<

The general solution of eqEps[[1]], the zeroth-order equation, can be expressed in polar form as

sol0 = DSolve@8eqEps@@1DD, u0@0D == a Cos@bD, u0 ¢@0D == -a Sin@bD<, u0, tD@@1DD

8u0 ® Ha Cos@bD Cos@#1D - a Sin@bD Sin@#1D &L<

where a and b are constants. We have expressed u 0 as a pure function so that we can evaluate its derivatives in the higher-
order problems.

Substituting for u 0@tD in eqEps[[2]], the first-order equation, yields

32 Chapter2.nb

order1Eq = eqEps@@2DD •. sol0 •• Simplify

a3 Cos@b + tD3 - 2 a Cos@b + tD w1 + u1 @tD + u²1 @tD == 0

whose particular solution is

u1pSol = DSolve@order1Eq, u1@tD, tD@@1DD •. C@_D -> 0 •• Simplify

9u1 @tD ®

€€€€€€€ a Ha2 H-6 Cos@b + tD + Cos@3 Hb + tLD - 12 t Sin@b + tDL + 16 HCos@b + tD + 2 t Sin@b + tDL w1L=
1
32

It is clear that u1pSol contains a mixed-secular term, which makes the expansion nonuniform. In contrast with the straightfor-
ward expansion, where the secular term cannot be annihilated unless u is trivial, in this case, we can choose the parameter w1
to eliminate the secular term according to

omg1Rule = Solve@Coefficient@u1pSol@@1, 2DD, Sin@b + tDD == 0, w1 D@@1DD

:w1 ® €€€€€€€€€€ >

3 a2
8

Substituting for w1 into u1pSol, we have

sol1 = u1pSol •. omg1Rule •• Simplify

9u1 @tD ® €€€€€€ a3 Cos@3 Hb + tLD=

1
32

We note that, to determine the condition for the elimination of the secular term from u 1, we do not need to determine the
particular solution first as done above. Instead, we only need to inspect order1Eq and choose w1 so that the coefficient of
Cos@b + tD, which produces secular terms in u 1 , is equal to zero. To this end, we calculate the nonhomogeneous part of
order1Eq and simplify the result using trigonometric identities as

expr23a = order1Eq@@1DD •. u1 -> H0 &L •• TrigReduce

€€€€ H3 a3 Cos@b + tD + a3 Cos@3 b + 3 tD - 8 a Cos@b + tD w1 L

1
4

Next, we choose w1 to annihilate the term in expr23a that produces secular terms, namely Cos@b + tD, and obtain

Solve@Coefficient@expr23a, Cos@b + tDD == 0, w1 D@@1DD

:w1 ® €€€€€€€€€€ >

3 a2
8

Substituting sol0 and sol1 into uExpRule[1] yields

u@tD = Hu@tD •. uExpRule@1D •. sol0 •. sol1 •• SimplifyL

a Cos@b + tD + €€€€€€€ a3 e Cos@3 Hb + tLD

1
32
Chapter2.nb 33

Replacing t with w t, substituting for w, using the fact that w0 = 1, and expanding the arguments of the trigonometric func-
tions, we obtain

solLP =
u@tD == Hu@tD •. t -> w t •. omgRule@1D •. omg1Rule •. w0 -> 1 •. f_@arg_D :> f@arg •• ExpandDL

3 1 9
u@tD == a CosAt + b + €€€€ a2 t eE + €€€€€€€ a3 e CosA3 t + 3 b + €€€€ a2 t eE
8 32 8

1 9
Clearly, the expansion solLP is free of secular terms and the correction term €€€€32€€€ a 3 eCosA3 t + 3 b + €€€€8€ e a 2 tE is small compared
3
with the principal term a CosAt + b + €€8€€ e a 2 tE for all t and hence it is uniformly valid.

à 2.4 The Method of Multiple Scales

We note from solLP that the functional dependence of u on t and e is not disjoint. In fact, u depends on the combination
e t as well as on the individual t and e. Carrying out the expansion to higher order, we find that u depends on the combina-
tions e t, e2 t, e3 t,º as well as on the individual t and e. Hence, u = u(e t, e2 t, e3 t,º; e). For small e, Tn = en t, for
1
n = 1, 2, 3,º, represent different time scales. For example, if e = €€€€
60
€€€ , variations on the time scales T0, T1, and T2 can be
observed, respectively, on the second, minute, and hour arms of a watch. Thus, T0 represents a fast time scale, T1 repre-
sents a slower time scale, T2 represents an even slower time scale, and so on. Since the dependence of u on t and e occurs
on different time scales, we imagine that we have a watch and attempt to observe the behavior of u using the different
scales of the watch. Therefore, instead of determining u as a function of t, we determine u as a function of T0 , T1, T2 , º.
To this end, we change the independent variable t in the governing equation from t to T0 , T1, T2 , º , which are symbolized
by

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

timeScales = 8T0 , T1, T2 <;

We define the derivative operators as

dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, maxOrder<E;

dt@2D@expr_D := Hdt@1D@dt@1D@exprDD •• ExpandL •. ei_•;i>maxOrder -> 0;

and treat the time scales T0, T1 , and T2 as independent variables.

Next, we use the method of multiple scales to determine an approximate solution for the DuffingEq. In Section 2.4.1, we treat
the DuffingEq in its second-order form; in Section 2.4.2, we transform it first into a system of two real-valued equations; and
in Section 2.4.3, we transform it first into a single complex-valued equation.

Ÿ 2.4.1 Second-Order Real-Valued System

DuffingEq = u² @tD + u@tD + e u@tD3 == 0;

34 Chapter2.nb

To determine a first-order uniform expansion of the solution of the Duffing equation, we use the two time scales T0 and T1
and hence let

maxOrder = 1;

Expressing the time derivative in DuffingEq in terms of the two time scales T0 and T1 , we have

eq241a = DuffingEq •. 8u@tD -> u@T0, T1 D, Derivative@n_D@uD@tD -> dt@nD@u@T0 , T1 DD<

u@T0 , T1D + e u@T0, T1D3 + 2 e uH1,1L@T0 , T1 D + uH2,0L@T0 , T1 D == 0

Consequently, we have replaced the original ordinary-differential equation by a partial-differential equation, and it appears
that, as a result, we have complicated the problem. This is partially true, but experience with this method has shown that
the disadvantages of introducing this complication are far outweighed by the advantages. Not only does this method
provide a uniform expansion, but it also provides all the various nonlinear resonance phenomena, as we shall see in
subsequent chapters.

We seek a uniform second-order expansion of the solution of eq241a in the form

solRule = u -> ISumAei ui @#1, #2D, 8i, 0, maxOrder<E &M;

Substituting this expansion into eq241a, expanding the result, and discarding terms of order higher than e, we obtain

eq241b = Heq241a •. solRule •• ExpandAllL •. e n_•;n>maxOrder -> 0

u0 @T0, T1 D + e u0 @T0 , T1D3 + e u1 @T0, T1 D + 2 e u0 @T0, T1 D + u0 @T0, T1 D + e u1 @T0, T1 D == 0

H1,1L H2,0L H2,0L

Equating coefficients of like powers of e in eq241b yields

eqEps = CoefficientList@eq241b@@1DD, eD == 0 •• Thread

9u0 @T0, T1 D + u0 @T0, T1 D == 0, u0 @T0, T1 D3 + u1 @T0 , T1 D + 2 u0 @T0 , T1 D + u1 @T0 , T1 D == 0=

H2,0L H1,1L H2,0L

The general solution of eqEps[[1]] can be written in a pure function form as

sol0 = u0 -> Ha@#2D Cos@#1 + b@#2DD &L;

where #1 stands for the scale T0 and #2 stands for the scale T1. The functional dependence of a and b on T1 is not known at
this level of approximation; it is determined at subsequent levels of approximation by eliminating the secular terms.

Substituting sol0 into eqEps[[2]] and moving the nonhomogeneous parts to the right-hand side of the resulting equation
yields

eq241c = HeqEps@@1, 1DD •. u0 -> u1 L - # & •ž Reverse@eqEps@@2DD •. sol0 •• SimplifyD

u1 @T0, T1 D + u1 @T0, T1 D ==
H2,0L

-a@T1D3 Cos@T0 + b@T1DD3 + 2 Sin@T0 + b@T1 DD a¢@T1 D + 2 a@T1 D Cos@T0 + b@T1DD b¢ @T1D

Next, we expand the right-hand side of eq241c in a Fourier series using trigonometric identities and obtain
Chapter2.nb 35

eq241c@@2DD = eq241c@@2DD •• TrigFactor •• Expand

- €€€€ a@T1 D3 Cos@T0 + b@T1 DD - €€€€ a@T1 D3 Cos@3 T0 + 3 b@T1 DD +

3 1

2 Sin@T0 + b@T1 DD a @T1 D + 2 a@T1 D Cos@T0 + b@T1 DD b¢@T1 D

4 4
¢

a 1 ¢@T1 D and b ¢@T1 D to annihilate each of the coefficients of Sin@T0 + b@T1DD and Cos@T0 + b@T1 DD. The result is
To produce a uniform expansion, we eliminate the terms that produce secular terms from eq241c[[2]]; that is, we choose

eq241d = Coefficient@eq241c@@2DD, #D == 0 & •ž 8Sin@T0 + b@T1 DD, Cos@T0 + b@T1DD<

92 a¢ @T1D == 0, - €€€€ a@T1 D3 + 2 a@T1D b¢ @T1D == 0=

3
4

Solving eq241d for a 1¢ @T1D and b ¢ @T1D when a@T1 D ¹ 0 yields

SCond = Solve@eq241d, 8a¢@T1 D, b¢ @T1D<D@@1DD

9a¢ @T1D ® 0, b¢ @T1 D ® €€€€ a@T1D2 =

3
8

The solution of SCond[[1]] is a = a 0 = constant. Then, the solution of SCond[[2]] is

betaRule = DSolve@eq241d@@2DD •. a@T1 D -> a0 , b@T1D, T1 D@@1DD •. C@1D -> b0

9b@T1 D ® €€€€ T1 a20 + b0 =

3
8

With the above solvability conditions, eq241c becomes

eq241e = eq241c •. SCond

u1 @T0, T1 D + uH2,0L @T0, T1 D == - €€€€ a@T1D3 Cos@3 T0 + 3 b@T1DD

1
1
4

The particular solution of eq241e can be obtained as

sol1 = DSolve@eq241e, u1@T0 , T1 D, 8T0, T1<D@@1DD •. C@_D -> H0 &L •• Simplify

9u1 @T0, T1 D ® €€€€€€€ a@T1 D3 Cos@3 HT0 + b@T1DLD=

1
32

eq241f = u@T0 , T1D == Hu@T0 , T1D •. solRule •. sol0 •. sol1 •. a@T1 D -> a0 •. betaRuleL

u@T0 , T1D == CosAT0 + €€€€ T1 a20 + b0 E a0 + €€€€€€ e CosA3 i

jT0 + €€€€ T1 a20 + b0 y
zE a30
k {
3 1 3
8 32 8

where a 0 and b 0 are constants to within the order of the error indicated. In terms of the original variable t, eq241f can be
expressed as

u@tD == eq241f@@2DD •. 8T0 -> t, T1 -> e t<

u@tD == CosAt + €€€€ t e a20 + b0E a0 + €€€€€€€ e CosA3 i

jt + €€€€ t e a20 + b0y
zE a30
k {
3 1 3
8 32 8
36 Chapter2.nb

in agreement with the expansion obtained by using the Lindstedt-Poincaré technique.

In the higher-order approximations, we include the scales T0 , T1, º, T N but we do not include the term O@eN D in an N th-
order expansion.

Before closing this section, we present an alternate representation of the solutions of the perturbation equations. Instead of
the real-valued form sol0, we represent the solution of eqEps[[1]] in a complex-valued form; that is,

sol0 = u0 -> IA@#2D Exp@I #1D + A@#2D Exp@-I #1D &M;

•

•••
where A is the complex conjugate of A and

ruleA = 9A -> i
j €€€€ a@#D Exp@I b@#DD &y A -> i
z, •• €€€€ a@#D Exp@-I b@#DD &y
•• j 1 z=;
k2
1
{ k2 {

Substituting sol0 into eqEps[[2]] and moving the nonhomogeneous terms to the right-hand side of the resulting equation, we
have

H2,0L
eq241g = u1 @T0, T1 D + u1 @T0, T1 D - # & •ž HeqEps@@2DD •. sol0L •• ExpandAll •• Reverse

u1 @T0, T1 D + u1 @T0, T1 D == -E3 IT0 A@T1 D3 - 3 EI T0 A@T1 D2 A@T1 D -

H2,0L •

A@T1 D A@T1 D - E-3 I T0 A@T1D - 2 I EI T0 A¢ @T1D + 2 I E-I T0 A @T1 D

• • •¢
3 E-I T0
2 3

We note that the terms proportional to Exp@I T0 D and Exp@-I T0 D produce secular terms in the particular solution of u 1 . Thus,
to produce a uniform expansion, we set the coefficient of each of these functions equal to zero; that is,

eq241h = Coefficient@eq241g@@2DD, #D == 0 & •ž 8Exp@I T0 D, Exp@-I T0 D<

:-3 A@T1D2 A@T1D - 2 I A¢@T1 D == 0, -3 A@T1 D A@T1D + 2 I A @T1 D == 0>

• • 2 •¢

These two equations are not independent because they are complex conjugates of each other. Hence, if one of them is
satisfied, the other is automatically satisfied.

To analyze the solutions of eq241h[[1]], we multiply it by Exp@-I b@T1DD and replace A with its polar form ruleA. The result is

eq241i = Expand@eq241h@@1, 1DD Exp@-I b@T1 DD •. ruleAD == 0

- €€€€ a@T1 D3 - I a¢@T1 D + a@T1 D b¢@T1 D == 0

3
8

Next, we separate the real and imaginary parts of eq241i. To accomplish this, we define the following rule:

realRule = 8Re@s_D -> s, Im@s_D -> 0<;

Using the realRule, we find that the imaginary part of eq241i is

ampEq = Im@eq241i@@1DDD == 0 •. realRule

-a¢ @T1 D == 0
Chapter2.nb 37

which governs the modulation of the amplitude. Similarly, the real part of eq241i is

phaseEq = Re@eq241i@@1DDD == 0 •. realRule

- €€€€ a@T1 D3 + a@T1D b¢ @T1D == 0

3
8

which governs the modulation of the phase. These modulation equations are in agreement with those obtained above by
expressing the solution in real form.

Comparing the complex-valued and the real-valued representations, we find it more convenient to use the complex-valued
form. Therefore, the complex-valued form is used in the remainder of this book.

Ÿ 2.4.2 First-Order Real-Valued System

We let v@tD = u ¢ @tD and transform the DuffingEq into the following set of two first-order equations:

eq242a = 9u¢ @tD == v@tD, v¢@tD + u@tD == -e u@tD3 =;

We seek a first-order uniform expansion of the solution of eq242a in terms of the two time scales T0 and T1 in the form

solRule = 9u -> ISumAej uj @#1, #2D, 8j, 0, 1<E &M, v -> ISumAej vj @#1, #2D, 8j, 0, 1<E &M=;

Substituting solRule into eq242a, transforming the derivative with respect to t in terms of the derivatives with respect to T0
and T1, and discarding terms of order higher than e, we obtain

eq242b = Heq242a •. 8u_¢@tD -> dt@1D@u@T0 , T1DD, u_@tD -> u@T0 , T1 D< •. solRule •• ExpandAllL •.
e n_•;n>1 -> 0

9e u0 @T0 , T1 D + u0 @T0 , T1 D + e u1 @T0 , T1 D == v0 @T0 , T1D + e v1@T0 , T1D,

H0,1L H1,0L H1,0L

u0 @T0, T1 D + e u1 @T0 , T1D + e @T0, T1D + v0 @T0, T1D + e v1 @T0, T1 D == -e u0@T0 , T1D3 =
H0,1L H1,0L H1,0L
v0

Equating coefficients of like powers of e, we have

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq242b •• Transpose

99-v0@T0 , T1 D + u0 @T0 , T1 D == 0, u0@T0 , T1 D + v0 @T0 , T1 D == 0=,

H1,0L H1,0L

9-v1@T0 , T1 D + @T0 , T1 D + u1 @T0 , T1 D == 0,

H0,1L H1,0L

u0 @T0, T1 D3 + u1 @T0, T1 D + v0 @T0, T1 D + v1 @T0, T1 D == 0==

u0
H0,1L H1,0L

Ÿ Zeroth-Order Problem: Linear System

The zeroth-order problem is given by the linear system

linearSys = #@@1DD & •ž eqEps@@1DD

9-v0@T0 , T1 D + uH1,0L
0 @T0 , T1 D, u0@T0 , T1 D + vH1,0L
0 @T0 , T1 D=
38 Chapter2.nb

To determine the solution of this linear system, we seek a solution proportional to Exp@I T0 D as

coefList = E-I T0 linearSys •. 9u0 -> IP EI # &M, v0 -> IQ EI # &M= •• Expand

8I P - Q, P + I Q<

Next, we determine the coefficient matrix as

coefMat = Outer@Coefficient, coefList, 8P, Q<D

88I, -1<, 81, I<<

To determine the eigenvalues and eigenvectors of coefMat, we define the following conjugate rule and Hermitian matrix:

conjugateRule = 9A -> A, A -> A, Complex@0, n_D -> Complex@0, -nD=;

• •

hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose

Then the left and right eigenvectors of coefMat are given by

leftVec = 81, c1< •. Solve@[email protected], c1<L@@1DD == 0, c1D@@1DD

81, I<

rightVec = 81, c1< •. [email protected], c1<L@@1DD == 0, c1D@@1DD

81, I<

To express the solution of the zeroth-order problem, we introduce the following basic function:

basicH = A@T1 D EI T0 ;

In terms of this function and the right eigenvector, the solution of the zeroth-order problem can be expressed in terms of

sol0Form = rightVec basicH

8EI T0 A@T1 D, I EI T0 A@T1D<

and its complex conjugate as

sol0 =
8u0 -> Function@8T0 , T1 <, sol0Form@@1DD + Hsol0Form@@1DD •. conjugateRuleL •• EvaluateD,
v0 -> Function@8T0 , T1 <, sol0Form@@2DD + Hsol0Form@@2DD •. conjugateRuleL •• EvaluateD<

8u0 ® Function@8T0 , T1 <, EI T0 A@T1D + E-I T0 A@T1DD,

•

v0 ® Function@8T0 , T1 <, I EI T0 A@T1D - I E-I T0 A@T1DD<

•

We have expressed the solution of the zeroth-order problem in function form so that its partial derivatives can be readily
evaluated.
Chapter2.nb 39

Ÿ First-Order Problem

Substituting the zeroth-order solution sol0 into the first-order problem eqEps[[2]] and moving the nonhomogeneous terms to
the right-hand sides of the resulting equations, we obtain

order1Eq = HlinearSys •. u_0 -> u1 L ==

HlinearSys •. u_0 -> u1 L - HSubtract žž # & •ž eqEps@@2DD •. sol0 •• ExpandL •• Thread

:-v1@T0 , T1 D + u1 @T0 , T1 D == -EI T0 A¢ @T1D - E-I T0 A @T1 D,

H1,0L •¢

u1 @T0, T1 D + v1 @T0, T1 D == -E3 IT0 A@T1 D3 - 3 EI T0 A@T1 D2 A@T1 D -

H1,0L •

A@T1D A@T1 D - E-3 I T0 A@T1D - I EI T0 A¢ @T1D + I E-I T0 A @T1 D>

• 2 • 3 •¢
3 E-I T0

Eliminating the terms that produce secular terms in order1Eq demands that their right-hand sides are orthogonal to every

Exp@I T0D as
solution of the adjoint homogeneous problem. To determine these conditions, we first calculate the coefficient vector of

STerms = Coefficient@#@@2DD, Exp@I T0DD & •ž order1Eq

8-A¢ @T1 D, -3 A@T1D2 A@T1D - I A¢@T1 D<

•

Then, we demand that this vector is orthogonal to the left eigenvector and obtain the solvability condition

SCond = Expand@[email protected] == 0

3 I A@T1 D2 A@T1 D - 2 A¢ @T1D == 0

•

coefficient vector of Exp@-I T0 D, we obtain the complexconjugate of SCond.

which is in agreement with that obtained by treating the second-order form of the DuffingEq. Repeating the procedure for the

Ÿ 2.4.3 First-Order Complex-Valued System

In this section, we determine a first-order uniform expansion of the solution of the DuffingEq by transforming it first into a
single first-order complex-valued equation. To this end, we introduce the transformation

transfRule = 9u@tD -> z@tD + z@tD, u¢ @tD -> I Iz@tD - z@tDM=;

• •

whose inverse is

zetaRule = SolveAtransfRule •. Rule -> Equal, 9z@tD, z@tD=E@@1DD

•

9z@tD ® €€€€ Hu@tD - I u¢@tDL, z@tD ® €€€€ Hu@tD + I u¢@tDL=

1 • 1
2 2

It follows from the DuffingEq that the acceleration is related to the displacement according to
40 Chapter2.nb

acceleration = Solve@DuffingEq, u²@tDD@@1DD

8u² @tD ® -u@tD - e u@tD3<

Differentiating z@tD, zetaRule[[1]], with respect to t and substituting for u@tD, u ¢ @tD, and u ²@tD using the acceleration and
transfRule, we obtain

eq243a = D@zetaRule@@1DD, tD •. acceleration •. transfRule •. Rule -> Equal •• ExpandAll

z¢ @tD == I z@tD + €€€€ I e z@tD3 + €€€€ I e z@tD2 z@tD + €€€€ I e z@tD z@tD + €€€€ I e z@tD
1 3 • 3 • 2 1 • 3

2 2 2 2

Next, we use the method of multiple scales to determine a first-order uniform expansion of the solution of eq243a in the form

solRule = 9z -> ISumAej zj @#1, #2D, 8j, 0, 1<E &M, z -> ISumAej zj @#1, #2D, 8j, 0, 1<E &M=;
• •

Substituting this expansion into eq243a, expanding the result, and discarding terms of order higher than e, we have

eq243b =
Ieq243a •. 9z@tD -> z@T0, T1D, z@tD -> z@T0 , T1D, z¢ @tD -> dt@1D@z@T0, T1 DD= •. solRule ••
• •

ExpandAllM •. e n_•;n>1 -> 0

@T0 , T1 D + z0 @T0 , T1 D + e z1 @T0 , T1 D == I z0 @T0 , T1D + €€€€ I e z0@T0 , T1 D3 +

H0,1L H1,0L H1,0L 1
e z0
2
I e z1 @T0, T1 D + €€€€ I e z0 @T0, T1D2 z0@T0 , T1D + €€€€ I e z0@T0 , T1 D z0@T0 , T1D + €€€€ I e z0 @T0 , T1D
3 • 3 • 2 1 • 3

2 2 2

Equating coefficients of like power of e in eq243b yields

eqEps = CoefficientList@Subtract žž eq243b, eD == 0 •• Thread

9-I z0@T0 , T1 D + z0 @T0 , T1 D == 0,

H1,0L

- €€€€ I z0 @T0 , T1 D3 - I z1@T0 , T1 D - €€€€ I z0 @T0, T1 D2 z0 @T0 , T1D - €€€€ I z0@T0 , T1 D z0@T0 , T1D -
1 3 • 3 • 2

2 2 2
€€€€ I z0 @T0, T1 D + z0 @T0 , T1 D + z1 @T0 , T1 D == 0=
1 • 3 H0,1L H1,0L

The solution of the zeroth-order problem, eqEps[[1]], can be expressed as

sol0Form = DSolve@eqEps@@1DD, z0@T0 , T1D, 8T0, T1 <D@@1DD •. C@1D -> A

8z0 @T0, T1 D ® EI T0 A@T1D<

which we express in function form as

sol0 = 8z0 -> Function@8T0 , T1 <, sol0Form@@1, 2DD •• EvaluateD<

8z0 ® Function@8T0 , T1 <, EI T0 A@T1DD<

To evaluate the complex conjugates of A and z, we define the conjugate rule

Chapter2.nb 41

conjugateRule = 9A -> A, A -> A, z -> z, z -> z, Complex@0, n_D -> Complex@0, -nD=;
• • • •

Substituting sol0 into the first-order problem, eqEps[[2]], using conjugateRule, and moving the nonhomogeneous terms to
the right-hand side of the resulting equation, we obtain

order1Eq = HeqEps@@1, 1DD •. z0 -> z1 L ==

HeqEps@@1, 1DD •. z0 -> z1L - HeqEps@@2, 1DD •. sol0 •. Hsol0 •. conjugateRuleLL

-I z1@T0 , T1 D + z1 @T0 , T1 D ==
H1,0L

€€€€ I E3 I T0 A@T1 D3 + €€€€ I EI T0 A@T1D2 A@T1D + €€€€ I E-I T0 A@T1D A@T1 D + €€€€ I E-3 I T0 A@T1D - EI T0 A¢ @T1D
1 3 • 3 • 2 1 • 3

2 2 2 2

Elimintaing the terms that lead to secular terms from order1Eq yields the solvability condition

SCond = CoefficientAorder1Eq@@2DD, EIT0 E == 0

€€€€ I A@T1 D2 A@T1 D - A¢ @T1D == 0

3 •
2

in agreement with that obtained by treating the second-order form of the DuffingEq as well as that obtained by transforming
it first into two first-order real-valued equations.

à 2.5 Variation of Parameters

In the next section, we use the method of averaging to determine a first-order uniform expansion of the solution of the
Duffing equation. To this end, we use the method of variation of parameters to transform it into a system of two first-order
equations. The Duffing equation is repeated here; that is,

DuffingEq = u² @tD + u@tD + e u@tD3 == 0;

When e = 0, the solution of this equation can be written as

usolEq = u@tD == a Cos@t + bD;

where a and b are constants, which are sometimes referred to as parameters. It follows from usolEq that

cond01 = D@usolEq, tD

u¢ @tD == -a Sin@t + bD

When e ¹ 0, we assume that the solution of DuffingEq is still given by usolEq but with time-varying a and b; that is,

tdepRule = 8a -> a@tD, b -> b@tD<;

and hence

usolEqt = usolEq •. tdepRule

u@tD == a@tD Cos@t + b@tDD

42 Chapter2.nb

In other words, we consider usolEq as a transformation from u@tD to a@tD and b@tD. This is why this approach is called the
method of variation of parameters. Using this view, we note that we have two equations, namely DuffingEq and usolEq, for
the three unknowns u@tD, a@tD, and b@tD. Hence, we have the freedom of imposing a third condition (third equation). This
condition is arbitrary except that it must be independent of DuffingEq and usolEq. This arbitrariness can be used to advan-

condition cond01, thereby assuming that u@tD as well as u ¢@tD have the same form as in the linear case. This condition leads to
tage, namely to produce a simple and convenient transformation. Out of all possible conditions, we choose to impose the

a convenient transformation; it leads to a set of first-order rather than second-order equations for a@tD and b@tD.

Differentiating usolEq with respect to t and recalling that a and b are functions of t, we have

cond02 = D@usolEqt, tD

u¢ @tD == Cos@t + b@tDD a¢ @tD - a@tD Sin@t + b@tDD H1 + b¢@tDL

Comparing cond02 with cond01, we conclude that

cond1 = Expand@cond02@@2DD - Hcond01@@2DD •. tdepRuleLD == 0

Cos@t + b@tDD a¢ @tD - a@tD Sin@t + b@tDD b¢@tD == 0

Differentiating cond01 with respect to t, we obtain

cond03 = D@cond01 •. tdepRule, tD

u² @tD == -Sin@t + b@tDD a¢@tD - a@tD Cos@t + b@tDD H1 + b¢ @tDL

Substituting for u@tD and u ²@tD from usolEqt and cond03 into DuffingEq, we have

cond2 = DuffingEq •. H8usolEqt, cond03< •. Equal -> RuleL •• ExpandAll

e a@tD3 Cos@t + b@tDD3 - Sin@t + b@tDD a¢ @tD - a@tD Cos@t + b@tDD b¢ @tD == 0

Solving cond1 and cond2 for a ¢@tD and b ¢@tD, we obtain the desired two first-order equations

transformedEq = HSolve@8cond1, cond2<, 8a¢@tD, b¢ @tD<D@@1DD •• SimplifyL •. Rule -> Equal

8a¢ @tD == e a@tD3 Cos@t + b@tDD3 Sin@t + b@tDD, b¢ @tD == e a@tD2 Cos@t + b@tDD4 <

if a T 0. Thus, the original second-order DuffingEq for u@tD has been replaced by two first-order equations for a@tD and b@tD.
We emphasize that no approximations have been made in deriving transformedEq. Comparing transformedEq with Duff-
ingEq, we find that the transformed equations are more nonlinear than the original equation. Then, the question arises
what is the value of this transformation? The answer depends on the value of e. If e is small, the major parts of a and b vary
more slowly with t than u. This fact can be used to advantage analytically and numerically. The analytical advantage is
utilized in the method of averaging, as discussed in the next section. Numerically, it is advantageous to solve the trans -
formed equations instead of the original equation because a large step size can be used in the integration. This is the
reason why astronomers numerically solve the variational equations rather than the original equations. Usually, astrono -
mers and celestial mechanicians refer to this approach as the special method of perturbations.
Chapter2.nb 43

à 2.6 The Method of Averaging

In this section, we determine a first-order approximation to the transformed equations obtained in the preceding section.
To this end, we rewrite transformedEq as

ampEq = TrigReduce •ž transformedEq@@1DD

a¢ @tD == €€€€ H2 e a@tD3 Sin@2 t + 2 b@tDD + e a@tD3 Sin@4 t + 4 b@tDDL

1
8

phaseEq = TrigReduce •ž transformedEq@@2DD

b¢ @tD == €€€€ H3 e a@tD2 + 4 e a@tD2 Cos@2 t + 2 b@tDD + e a@tD2 Cos@4 t + 4 b@tDDL

1
8

The major parts of a and b are slowly varying functions of time if e is small. Hence, they change very little during the time

@0, pD. Hence, we replace a@tD and b@tD on the right-hand sides of ampEq and phaseEq with time-independent a and b and
interval p (the period of the circular functions) and, to the first approximation, they can be considered constant in the interval

obtain

ampEq@@2DD = ampEq@@2DD •. HReverse •ž tdepRuleL

€€€€ H2 a3 e Sin@2 t + 2 bD + a3 e Sin@4 t + 4 bDL

1
8

phaseEq@@2DD = phaseEq@@2DD •. HReverse •ž tdepRuleL

€€€€ H3 a2 e + 4 a2 e Cos@2 t + 2 bD + a2 e Cos@4 t + 4 bDL

1
8

We average the right-hand sides of ampEq and phaseEq over the interval @0, pD and obtain

list26a = €€€€€ Integrate@#, 8t, 0, p<D & •ž 8ampEq@@2DD, phaseEq@@2DD< •• Expand

1
p
Integrate::gener : Unable to check convergence

Integrate::gener : Unable to check convergence

:0, €€€€€€€€€€€€€€ >

3 a2 e
8

It follows from the above result that

averagingEq = 8a¢ @tD, b¢@tD< == list26a •• Thread

:a¢ @tD == 0, b¢@tD == €€€€€€€€€€€€€€ >

3 a2 e
8

This averaging method is usually referred to as the Krylov-Bogoliubov or van der Pol technique.

Solving the averaged equations yields

44 Chapter2.nb

rule26a = DSolve@averagingEq, 8a@tD, b@tD<, tD@@1DD •. 8C@1D -> a0 , C@2D -> b0 <

9a@tD ® a0 , b@tD ® €€€€ H3 a2 t e + 8 b0L=

1
8

Substituting for a@tD and b@tD into usolEq, we obtain, to the first approximation, that

usolEq •. tdepRule •. rule26a •. Cos@arg_D :> Cos@Expand@argDD

u@tD == CosAt + €€€€ a2 t e + b0E a0

3
8

in agreement with the solutions obtained by using the Lindstedt-Poincaré technique and the method of multiple scales.

Before closing this section, we note that one can arrive at the final results in ampEq and phaseEq without going through the
averaging process. The right-hand sides of ampEq and phaseEq are the sum of two groups of terms — a group that is a linear

mation, a ¢@tD in ampEq is equal to the slowly varying group on its right-hand side, which is zero. And, to the first approxima -
combination of fast varying terms and a group that is a linear combination of slowly varying terms. Then, to the first approxi -

tion, b ¢ @tD in phaseEq is equal to the slowly varying group on its right-hand side, which is €€8€€ ea 2 .
3
Chapter 3
Systems with Quadratic and Cubic
Nonlinearities
Off@General::spell1D

à 3.1 Nondimensional Equation of Motion

We consider the free oscillations of a particle of mass m under the action of gravity and restrained by a nonlinear spring. The
equation of motion is

eq31a = m D@x* @t* D, 8t* , 2<D + f@x* @t*DD == m g

f@x* @t*DD + m Hx*L² @t*D == g m

where g is the gravitational acceleration and f @x* D is the restraining force of the spring. We assume that f @x* D is an odd cubic
function of x*; that is,

eq31b = eq31a •. f@a_D -> k1 a + k3 a3

k1 x*@t* D + k3 x*@t* D3 + m Hx*L² @t* D == g m

The equilibrium positions x* = x*s = constant can be obtained by dropping the acceleration term. The result is

eq31c = eq31b •. x* -> Hx*s &L

k1 HxsL* + k3 HHxsL* L3 == g m

In this chapter, we investigate small oscillations about one of the equilibrium positions. To this end, we shift this equilibrium
position to the origin by using the transformation

xRule = 8x* -> Hx*s + u*@#D &L<;

Substituting this transformation into eq31b and using the equilibrium condition eq31c, we obtain

eq31d = Expand@Subtract žž eq31b •. xRuleD == 0 •. Heq31c •. Equal -> RuleL

k1 u*@t* D + 3 k3 HHxsL* L2 u* @t*D + 3 k3 Hxs L* u* @t*D2 + k3 u* @t*D3 + m Hu* L²@t* D == 0

As before, we introduce the following dimensionless quantities:

dimenRule = 8u* -> Hx*s u@# • TD &L, t* -> T t<;

46 Chapter3.nb

into eq31d and obtain

eq31e = Ieq31d@@1DD T2 • Hm x*s L •. dimenRule •• ExpandM == 0

3 T2 k3 HHxs L*L2 u@tD 3 T2 k3 HHxsL* L2 u@tD2 T2 k3 HHxs L*L2 u@tD3

€€€€€€€€€€€€€€ + u²@tD == 0
T2 k1 u@tD
€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
m m m m

We rewrite eq31e as

eq31f = HCollect@eq31e@@1DD, u@tDD •. 8_ u@tD ^n_. -> a n u@tD ^n<L == 0

a1 u@tD + a2 u@tD2 + a3 u@tD3 + u² @tD == 0

where the a j are given by

alphas = Ia # -> CoefficientAeq31e@@1DD, u@tD # EM & •ž 81, 2, 3<

3 T2 k3 HHxs L*L2 3 T2 k3 HHxs L* L2 T2 k3 HHxs L*L2

:a1 ® €€€€€€€€€1€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€ >
T2 k
€€€€€€€€€€€€€€€€€€€€€ , a2 ® €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ , a3 ® €€€€€€€€€€€€€€€€
m m m m

We choose T so that a1 = 1 and hence the natural frequency is unity; that is,

TRule = Solve@a1 == 1 •. alphas, TD@@2DD

• !m!
:T ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ >
" ### #### #### #### ### #### ###
k1 + 3 k3 HHxs L*L2

Consequently, a2 and a3 can be rewritten as

alphas = alphas •. TRule •• Rest

3 k3 HHxs L*L2 k3 HHxs L*L2

:a2 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ >
k1 + 3 k3 HHxsL* L2 k1 + 3 k3 HHxsL* L2
€€€€€€€€ , a3 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

With a1 being unity, eq31f becomes

eq31g = eq31f •. a1 -> 1

u@tD + a2 u@tD2 + a3 u@tD3 + u²@tD == 0

In contrast with the DuffingEq, eq31g contains a quadratic as well as a cubic term.

In the next section, we determine a second-order straightforward expansion to the solutions of eq31g for small but finite
amplitudes. In Section 3.3, we determine a uniform second-order expansion by using the Lindstedt-Poincaré technique. In
Section 3.4, we determine a uniform second-order expansion by using the method of multiple scales. In Section 3.5, we show
that the first approximation obtained with the method of averaging yields an incomplete solution. In Section 3.6, we introduce
the generalized method of averaging and obtain a uniform second-order expansion. Finally, in Sections 3.7 and 3.8, we
introduce the Krylov-Bogoliubov-Mitropolsky technique and the method of normal forms, respectively.
Chapter3.nb 47

à 3.2 Straightforward Expansion

To carry out a straightforward expansion for small but finite amplitudes for eq31g, we need to introduce a small parameter e as
a bookkeeping parameter. In terms of this parameter, we seek a third-order expansion in the form

uExpRule@maxOrder_D := u -> ISumAei ui @#D, 8i, 1, maxOrder<E &M

where maxOrder is the order of the expansion sought. In what follows, we let maxOrder = 3 for an expansion of order three.
Clearly, e is a measure of the amplitude of oscillation and can be set equal to unity in the final solution if the amplitude is
taken to be small, as described below. Substituting this expansion into eq31g, expanding the result for small e, and discarding
terms of order higher than e3, we have

eq32a = Heq31g •. uExpRule@3D •• ExpandAllL •. e n_•;n>3 -> 0

e u1 @tD + e2 a2 u1 @tD2 + e3 a3 u1@tD3 + e2 u2@tD +

2 e3 a2 u1 @tD u2@tD + e3 u3 @tD + e u²1 @tD + e2 u²2@tD + e3 u²3 @tD == 0

Equating coefficients of like powers of e in eq32a yields

eqEps = TableACoefficientAeq32a@@1DD, ei E == 0, 8i, 3<E

8u1 @tD + u²1 @tD == 0, a2 u1 @tD2 + u2@tD + u²2@tD == 0, a3 u1 @tD3 + 2 a2 u1 @tD u2 @tD + u3 @tD + u²3@tD == 0<

which can be solved sequentially for u 1, u 2 , and u 3.

The general solution of the first-order equation, eqEps[[1]], can be expressed as

sol1 = DSolve@8eqEps@@1DD, u1@0D == a Cos@bD, u1 ¢@0D == -a Sin@bD<, u1@tD, tD@@1DD •• Simplify

8u1 @tD ® a Cos@t + bD<

where a and b are constants. Substituting sol1, u 1 @tD, into the second-order equation, eqEps[[2]], and solving the resulting
equation for u 2@tD, we obtain

sol2 = DSolve@eqEps@@2DD •. sol1, u2@tD, tD@@1DD •. C@_D -> 0 •• Simplify

9u2 @tD ® €€€€ a2 H-3 + Cos@2 Ht + bLDL a2=

1
6

As before, we do not include the solution of the homogeneous problem for u 2 @tD.

Substituting sol1 and sol2 into the third-order equation, eqEps[[3]], and solving the resulting equation u 3@tD, we have

sol3 = DSolve@eqEps@@3DD •. sol1 •. sol2, u3@tD, tD@@1DD •. C@_D -> 0 •• Simplify

9u3 @tD ® €€€€€€ a3 H2 H10 Cos@t + bD + Cos@3 Ht + bLD + 20 t Sin@t + bDL a22 +
1

3 H-6 Cos@t + bD + Cos@3 Ht + bLD - 12 t Sin@t + bDL a3 L=

96
48 Chapter3.nb

Simply letting C@_D ® 0 did not remove all of the solutions of the homogeneous equation in this case. Therefore, we clear it
one more time and obtain

sol3 = sol3 •. Cos@t + _D -> 0 •• ExpandAll

9u3 @tD ®

€€€€€€€ a3 Cos@3 t + 3 bD a22 + €€€€€€€ a3 t Sin@t + bD a22 + €€€€€€€ a3 Cos@3 t + 3 bD a3 - €€€€ a3 t Sin@t + bD a3=
1 5 1 3
48 12 32 8

Combining the first-, second-, and third-order solutions, we obtain, to the third approximation, that

sol = u@tD == Hu@tD •. uExpRule@3D •. Flatten@8sol1, sol2, sol3<DL

u@tD == a e Cos@t + bD + €€€€ a2 e2 H-3 + Cos@2 Ht + bLDL a2 +

1

3 i y
e j €€€€€€ a Cos@3 t + 3 bD a22 + €€€€€€ a3 t Sin@t + bD a22 + €€€€€€ a3 Cos@3 t + 3 bD a3 - €€€€ a3 t Sin@t + bD a3 z
6

k 48 {
1 3 5 1 3
12 32 8

We note that the dependence of u on e and a is in the combination e a. Thus, one can set e = 1 and consider a as the
perturbation parameter.

The straightforward expansion sol breaks down for t ³ OHe-1 a -1 L because the third-order term is the same order or larger
than the second-order term owing to the presence of the mixed-secular term. Next, we use the Lindstedt-Poincaré tech -
nique to determine a third-order uniform solution.

à 3.3 The Lindstedt-Poincaré Technique

To apply the Lindstedt-Poincaré technique to eq31g, we introduce the transformation t = w t and rewrite it as

eq33a = eq31g •. u -> Hu@w #D &L •. t -> t • w

u@tD + a2 u@tD2 + a3 u@tD3 + w2 u² @tD == 0

To determine a uniform expansion of order maxOrder of the solution of eq33a, we expand both w and u@tD in powers of e as

omgRule@maxOrder_D := w -> 1 + SumAei wi, 8i, maxOrder<E

uExpRule@maxOrder_D := u -> ISumAei ui @#D, 8i, maxOrder + 1<E &M

As discussed in the preceding chapter, the first term in the expansion of w is taken to be the linear natural frequency of the
system, which is unity in this case. We note that the order of the expansion for u is larger than that of w and the uniform
expansion required. As discussed below, we eliminate the terms that lead to secular terms from the equation at order max-
Order+1 but do not include in the solution of the resulting equation in the expansion.

For a uniform second order expansion, we let

maxOrder = 2;
Chapter3.nb 49

Substituting the expansions of u and w into eq33a, expanding the result for small e, and discarding terms of order higher
than e3 , we obtain

eq33b = Heq33a •. 8uExpRule@maxOrderD, omgRule@maxOrderD< •• ExpandAllL •. e n_•;n>maxOrder+1 -> 0

e u1 @tD + e2 a2 u1 @tD2 + e3 a3 u1@tD3 + e2 u2@tD + 2 e3 a2 u1@tD u2 @tD + e3 u3@tD + e u²1@tD +

2 e2 w1 u²1 @tD + e3 w21 u²1 @tD + 2 e3 w2 u²1 @tD + e2 u²2@tD + 2 e3 w1 u²2@tD + e3 u²3 @tD == 0

eqEps = TableACoefficientAeq33b@@1DD, ei E == 0, 8i, 3<E

8u1 @tD + u²1 @tD == 0, a2 u1 @tD2 + u2@tD + 2 w1 u²1 @tD + u²2 @tD == 0,
a3 u1@tD3 + 2 a2 u1@tD u2 @tD + u3 @tD + w21 u²1@tD + 2 w2 u²1 @tD + 2 w1 u²2@tD + u²3@tD == 0<

The general solution of the first-order equation, eqEps[[1]], can be expressed as

sol1 = DSolve@8eqEps@@1DD, u1@0D == a Cos@bD, u1 ¢@0D == -a Sin@bD<, u1, tD@@1DD

8u1 ® Ha Cos@bD Cos@#1D - a Sin@bD Sin@#1D &L<

where a and b are constants. Instead of using u 1 @tD, we use u 1 in the second argument to DSolve so that sol1 can be used in
the terms involved with derivatives.

Substituting sol1, u 1 @tD, into the second-order equation, eqEps[[2]], moving the nonhomogeneous terms to the right-hand
side of the resulting equation, we obtain

eq33c = u2 @tD + u²2 @tD - # & •ž Reverse@eqEps@@2DD •. sol1 •• SimplifyD

u2 @tD + u²2 @tD == -a2 Cos@b + tD2 a2 + 2 a Cos@b + tD w1

Expanding the right-hand side of eq33c in a Fourier series using trigonometric identities yields

eq33c@@2DD = eq33c@@2DD •• TrigReduce

€€€€ H-a2 a2 - a2 Cos@2 b + 2 tD a2 + 4 a Cos@b + tD w1L

1
2

Eliminating the terms, Cos@b + tD and Sin@b + tD, demands that w1 = 0. Then, the particular solution of eq33c can be
expressed as

sol2 = DSolve@eq33c •. w1 -> 0, u2 @tD, tD@@1DD •. C@_D -> 0 •• Simplify

9u2 @tD ® €€€€ a2 H-3 + Cos@2 Hb + tLDL a2=

1
6

or in a pure function format as

sol21 = u2 -> Function@t, sol2@@1, 2DD •• EvaluateD

u2 ® FunctionAt, €€€€ a2 H-3 + Cos@2 Hb + tLDL a2E

1
6

Substituting sol1 and sol21 into the third-order equation, eqEps[[3]], and using the fact that w1 = 0, we obtain
50 Chapter3.nb

eq33d = u3 @tD + u²3 @tD - # & •ž Reverse@eqEps@@3DD •. w1 -> 0 •. sol1 •. sol21 •• SimplifyD

u3 @tD + u²3 @tD == - €€€€ a3 Cos@b + tD H-3 + Cos@2 Hb + tLDL a22 - a3 Cos@b + tD3 a3 + 2 a Cos@b + tD w2
1
3

Expanding the right-hand side of eq33d in a Fourier series using trigonometric identities, we have

eq33d@@2DD = eq33d@@2DD •• TrigReduce

€€€€€€€ H10 a3 Cos@b + tD a22 - 2 a3 Cos@3 b + 3 tD a22 -

1

9 a3 Cos@b + tD a3 - 3 a3 Cos@3 b + 3 tD a3 + 24 a Cos@b + tD w2L

12

Eliminating the terms that lead to secular terms from eq33d[[2]] demands that

omg2Rule = Solve@Coefficient@eq33d@@2DD, Cos@t + bDD == 0, w2D@@1DD •• ExpandAll

:w2 ® - €€€€€€ a2 a22 + €€€€€€€€€€€€€€€€ >

5 3 a2 a3
12 8

As discussed above, for a second-order uniform expansion, we do not need to solve for u 3 @tD. Combining the first- and
second-order solutions, we obtain, to the second approximation, that

u@tD == Hu@tD •. uExpRule@2D •. sol1 •. sol21 •• Simplify •• ExpandL

u@tD == a e Cos@b + tD - €€€€ a2 e2 a2 + €€€€ a2 e2 Cos@2 Hb + tLD a2 + e3 u3 @tD

1 1
2 6

where

t == w t •. omgRule@2D •. omg2Rule •. w1 -> 0

i i 5 2 2 3 a2 a3 y y
t == t j
j1 + e2 j
j- €€€€€€ a a2 + €€€€€€€€€€€€€€€€ z
zz
z
k k 12 8 {{

The above expansion is uniform to second order because secular terms do not appear in it and the correction term (the term
proportional to e2) is small compared with the first term.

concluded that w1 = 0 before carrying out the expansion because the term e w1 in omgRule[2] creates secular terms at OHe2 L
Returning to sol in the previous section, we note that the first secular term appears at O(e3). Consequently, we could have

and not at OHe3L, as needed to eliminate the secular term from sol.

à 3.4 The Method of Multiple Scales

We use the method of multiple scales to attack directly eq31g in Section 3.4.1, the corresponding first-order real-valued
equations of eq31g in Section 3.4.2, and the corresponding first-order complex-valued equation of eq31g in Section 3.4.3. To
obtain a second-order uniform expansion by using the method of multiple scales, we need the three time scales
T0 = t, T1 = e t, and T2 = e2 t, which we symbolize by

Needs@"Utilities`Notation`"D
Chapter3.nb 51

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

and list as

timeScales = 8T0 , T1, T2 <;

In terms of the time scales Tn , the time derivatives become

dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, maxOrder<E;

dt@2D@expr_D := Hdt@1D@dt@1D@exprDD •• ExpandL •. ei_•;i>maxOrder -> 0;

For a uniform second-order expansion, we let

maxOrder = 2;

To represent some of the expressions in a more concise way, we introduce the following display rule:

displayRule =
9Derivative@a__DAu_i_E@__D :> SequenceFormATimes žž MapIndexedAD #1
#2@@1DD-1 &, 8a<E, ui E,

#2@@1DD &, 8a<E, AE,

Derivative@a__D@AD@__D :> SequenceFormATimes žž MapIndexedAD #1
Derivative@a__D@AD@__D :> SequenceFormATimes žž MapIndexedAD #1
#2@@1DD &, 8a<E, AE,
• •

u_i_ @__D -> ui , A@__D -> A, A@__D -> A=;

• •

Ÿ 3.4.1 Second-Order Real-Valued System

Using the derivative rule, we transform eq31g into the partial-differential equation

eq341a = eq31g •. 8u@tD -> u@T0, T1 , T2 D, Derivative@n_D@uD@tD -> dt@nD@u@T0 , T1, T2 DD<

u@T0 , T1, T2D + a2 u@T0 , T1, T2D2 + a3 u@T0, T1 , T2 D3 + e2 uH0,2,0L@T0 , T1 , T2D +

2 e2 uH1,0,1L @T0, T1 , T2 D + 2 e uH1,1,0L@T0 , T1, T2 D + uH2,0,0L @T0, T1 , T2 D == 0

Again, to determine a uniform expansion of order maxOrder, we expand u to order maxOrder+1, eliminate the secular terms
from the maxOrder+1 equation, but do not include its solution in the final approximate solution. Hence, we seek a uniform
expansion of the solution of eq341a in the form

solRule = u -> ISumAei ui @#1, #2, #3D, 8i, maxOrder + 1<E &M;

Substituting this expansion into eq341a, expanding the result for small e, and discarding terms of order higher than e3, we
have

eq341b = Heq341a •. solRule •• ExpandAllL •. e n_•;n>maxOrder+1 -> 0

e u1 @T0, T1 , T2 D + e2 a2 u1 @T0, T1 , T2 D2 + e3 a3 u1@T0 , T1 , T2D3 + e2 u2@T0 , T1 , T2D +

2 e3 a2 u1 @T0, T1 , T2 D u2@T0 , T1, T2 D + e3 u3 @T0, T1 , T2 D + e3 uH0,2,0L @T0, T1 , T2 D +
@T0, T1 , T2 D + 2 e2 u1 @T0, T1 , T2 D + 2 e3 u2 @T0, T1 , T2 D +
H1,0,1L H1,1,0L H1,1,0L
1
2 e3 u1
e uH2,0,0L
1 @T0 , T1, T2D + e2 uH2,0,0L
2 @T0 , T1, T2D + e3 uH2,0,0L
3 @T0 , T1 , T2D == 0
52 Chapter3.nb

Equating coefficients of like powers of e in eq341b, we obtain

eqEps = CoefficientList@eq341b@@1DD, eD == 0 •• Thread •• Rest

9u1 @T0, T1 , T2 D + u1 @T0 , T1 , T2D == 0,

H2,0,0L

a2 u1@T0 , T1, T2D + u2 @T0, T1, T2 D + 2 uH1,1,0L

2
@T0 , T1 , T2D + uH2,0,0L @T0, T1, T2 D == 0,
a3 u1@T0 , T1, T2D3 + 2 a2 u1@T0 , T1 , T2D u2 @T0, T1 , T2 D + u3 @T0, T1 , T2 D + u1 @T0 , T1, T2D +
H0,2,0L
1 2

@T0 , T1, T2D + 2 @T0, T1 , T2 D + @T0 , T1, T2D == 0=

H1,0,1L H1,1,0L H2,0,0L
2 u1 u2 u3

To place the linear operator on one side and the rest of the terms on the other, we define

eqOrder@i_D := HeqEps@@1, 1DD •. u1 -> ui L - # & •ž eqEps@@iDD •• Reverse

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

Array@eqOrder, 3D •. displayRule •• TableForm

D20 u1 + u1 == 0
D20 u2 + u2 == -2 HD0 D1u1 L - u21 a2
D20 u3 + u3 == -2 HD0 D1u2 L - D21 u1 - 2 HD0 D2 u1L - 2 u1 u2 a2 - u31 a3

where Dn u i = ¶ u i @T0, T1 , T2D • ¶ Tn .

The general solution of the first-order equation, eqOrder[1], can be expressed as

sol1 = 9u1 -> IA@#2, #3D Exp@I #1D + A@#2, #3D Exp@-I #1D &M=;
•

•••
where A is the complex conjugate of A defined by the conjugate rule

conjugateRule = 9A -> A, A -> A, Complex@0, n_D -> Complex@0, -nD=;

• •

Then, the second-order equation, eqOrder[2], becomes

eq341c = eqOrder@2D •. sol1 •• ExpandAll;

eq341c •. displayRule

D20 u2 + u2 == -2 I EI T0 HD1 AL + 2 I E-I T0 HD1AL - A2 E2 I T0 a2 - 2 A A a2 - E-2 I T0 A a2

• • •2

Eliminating the term EI T 0 that produces secular terms in u 2 from the right-hand side of eq341c demands that

SCond1 = Coefficient@eq341c@@2DD, Exp@I T0 DD == 0

-2 I AH1,0L @T1 , T2D == 0

or
Chapter3.nb 53

SCond1Rule = SolveASCond1, AH1,0L @T1 , T2DE@@1DD

H* either copy and paste AH1,0L @T1,T2 D from SCond1 or
use Derivative@1,0D@AD@T1,T2 D as the second argument to Solve *L

8AH1,0L@T1 , T2 D ® 0<

Eliminating the term E-I T 0 that produces secular terms in u 2 from the right-hand side of eq341c yields the complex conjugate
of SCond1Rule; that is,

ccSCond1Rule = SCond1Rule •. conjugateRule

•H1,0L
:A @T1 , T2 D ® 0>

It follows from SCond1 that A = A@T2 D. Substituting the solvability conditions into eq341c, we have

eq341d = eq341c •. SCond1Rule •. ccSCond1Rule

u2 @T0, T1 , T2 D + u2 @T0 , T1 , T2D ==

H2,0,0L

A@T1, T2 D a2 - 2 A@T1 , T2 D a2 A@T1 , T2D - E-2 I T0 a2 A@T1 , T2D

2 I T0 2 • • 2
-E

The particular solution of eq341d can be obtained by using DSolve; the result is

u2Sol =
DSolve@eq341d, u2 @T0 , T1 , T2D, timeScalesD@@1DD •. C@_D -> H0 &L •• TrigToExp •• ExpandAll

9u2 @T0, T1 , T2 D ® €€€€ E2 I T0 A@T1, T2 D2 a2 - 2 A@T1 , T2 D a2 A@T1 , T2D + €€€€ E-2 I T0 a2 A@T1 , T2D =
1 • 1 • 2

3 3

whose right-hand side can be used directly to express u 2 in a pure function form as

sol2 = 8u2 -> Function@8T0 , T1 , T2<, u2Sol@@1, 2DD •• EvaluateD<

9u2 ®

FunctionA8T0, T1, T2 <, €€€€ E2 I T0 A@T1 , T2D2 a2 - 2 A@T1, T2 D a2 A@T1 , T2 D + €€€€ E-2 I T0 a2 A@T1 , T2 D E=
1 • 1 • 2

3 3

Substituting the first- and second-order solutions, sol1 and sol2, into the third-order equation, eqOrder[3], we obtain

eq341e = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

eq341e •. displayRule

D20 u3 + u3 == -EI T0 HD21 AL - E-I T0 HD21 AL - 2 I EI T0 HD2AL + 2 I E-I T0 HD2AL -

• •

€€€€ I A E2 I T0 HD1 AL a2 + €€€€ I E-2I T0 A HD1 AL a2 - €€€€ A3 E3I T0 a22 + €€€€€€ A2 EI T0 A a22 +
8 8 • • 2 10 •
3 3 3 3
10 -I T0 •2 2 2 -3 I T0 •3 2 2 I T0 • •2 •3
€€€€€€€ A E A a2 - €€€€ E A a2 - A E3 3 I T0
a3 - 3 A E A a3 - 3 A E-I T0 A a3 - E-3 I T0 A a3
3 3

Eliminating the terms that produce secular terms in u 3 from the right-hand side of eq341e demands that
54 Chapter3.nb

SCond2 = Coefficient@eq341e@@2DD, Exp@I T0 DD == 0

€€€€€€€ A@T1 , T2 D2 a22 A@T1 , T2 D - 3 A@T1 , T2D2 a3 A@T1 , T2 D - 2 I AH0,1L@T1 , T2D - AH2,0L@T1 , T2D == 0
10 • •
3

Using SCond1Rule and the fact that A = A@T2D, we can rewrite SCond2 as

SCond = SCond2 •. D@SCond1Rule, T1D •. 9A -> HA@#2D &L, A -> IA@#2D &M=
• •

€€€€€€€ A@T2 D2 a22 A@T2D - 3 A@T2D2 a3 A@T2 D - 2 I A¢ @T2D == 0

10 • •
3

Expressing A in the polar form

ruleA = 9A -> i
j €€€€ a@#D Exp@I b@#DD &y A -> i
z, • j €€€€ a@#D Exp@-I b@#DD &y
z=;
k2
1
{ k2
1
{

where a and b are real and using the rule

realRule = 8Re@s_D -> s, Im@s_D -> 0<;

we write SCond as

eq341f = SCond@@1DD Exp@-I b@T2 DD •. ruleA •• Expand

€€€€€€€ a@T2 D3 a22 - €€€€ a@T2D3 a3 - I a¢@T2 D + a@T2 D b¢@T2 D

5 3
12 8

Separating the real and imaginary parts in eq341f, we obtain the modulation equations

ampEq = Im@eq341fD == 0 •. realRule

-a¢ @T2 D == 0

phaseEq = Re@eq341fD == 0 •. realRule

€€€€€€€ a@T2 D3 a22 - €€€€ a@T2D3 a3 + a@T2D b¢ @T2D == 0

5 3
12 8

The solution of ampEq is a = a 0 = const. Then, if a 0 ¹ 0, the solution of phaseEq is

betaRule = DSolve@phaseEq •. a@T2D -> a0 , b@T2D, T2 D@@1DD •. C@1D -> b0 •• ExpandAll

9b@T2 D ® - €€€€€€€ T2 a20 a22 + €€€€ T2 a20 a3 + b0=

5 3
12 8

where b 0 is a constant. Substituting ruleA into sol1 and sol2 and recalling that T0 = t and T2 = e2 t, we obtain, to the second
approximation, that
Chapter3.nb 55

u@tD == IIu@T0 , T1, T2 D •. solRule •. sol1 •. sol2 •. e3 -> 0 •. 9A@__D -> A@T2D, A@__D -> A@T2 D= •.
• •

ruleA •• ExpToTrigM •. 8a@T2 D -> a0 < •. betaRule •.

9T0 -> t, T2 -> e2 t= •. Cos@arg_D :> Cos@Collect@arg, tDDM

u@tD == e CosAt i j1 - €€€€€€€ e2 a20 a22 + €€€€ e2 a20 a3y

z + b0E a0 -
k {
5 3

€€€€ e a0 a2 + €€€€ e CosAt i j2 - €€€€ e2 a20 a22 + €€€€ e2 a20 a3y

z + 2 b0E a20 a2
12 8

k {
1 2 2 1 2 5 3
2 6 6 4

which is in full agreement with that obtained by using the Lindstedt-Poincaré technique.

Ÿ 3.4.2 First-Order Real-Valued System

v@tD = u ¢@tD and obtain

In this section, we first transform eq31g into a system of two real-valued first-order equations using the transformation

eq342a = 8u¢ @tD == v@tD, eq31g •. u² @tD -> v¢ @tD<

8u¢ @tD == v@tD, u@tD + a2 u@tD2 + a3 u@tD3 + v¢ @tD == 0<

To determine a second-order uniform expansion of the solution of eq342a using the method of multiple scales, we first
introduce

multiScales = 8u@tD -> u@T0, T1 , T2 D, v@tD -> v@T0, T1 , T2 D,

u¢ @tD -> dt@1D@u@T0 , T1 , T2DD, v¢ @tD -> dt@1D@v@T0, T1 , T2 DD<;

to transform the derivative with respect to t in terms of the derivatives with respect to the three time scales T0, T1 , and T2 and
then expand u and v in the form

solRule = 9u -> ISumAej uj @#1, #2, #3D, 8j, 3<E &M, v -> ISumAej vj @#1, #2, #3D, 8j, 3<E &M=;

Substituting multiScales and solRule into eq342a, expanding the result for small e, and discarding terms of order higher than
e3 , we obtain

eq342b = Heq342a •. multiScales •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e in eq342b yields

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq342b •• Transpose •• Rest;

To place the linear operator on one side and the rest of the terms on the other, we define

eqOrder@i_D := H#@@1DD & •ž eqEps@@1DD •. u_1 -> uiL ==

H#@@1DD & •ž eqEps@@1DD •. u_1 -> ui L - H#@@1DD & •ž eqEps@@iDDL •• Thread

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

56 Chapter3.nb

eqOrder@1D •. displayRule •• TableForm

eqOrder@2D •. displayRule •• TableForm
eqOrder@3D •. displayRule •• TableForm

D0 u1 - v1 == 0
D0 v1 + u1 == 0

D0 u2 - v2 == -HD1 u1L
D0 v2 + u2 == -HD1 v1L - u21 a2

D0 u3 - v3 == -HD1 u2L - D2u1

D0 v3 + u3 == -HD1 v2L - D2v1 - 2 u1 u2 a2 - u31 a3

Ÿ First-Order Equations: Linear System

To determine the solution of the first-order equations, eqOrder[1], we list their left-hand sides and obtain

linearSys = #@@1DD & •ž eqOrder@1D

9-v1@T0 , T1 , T2D + u1 @T0 , T1, T2 D, u1 @T0, T1 , T2 D + v1 @T0 , T1, T2D=

H1,0,0L H1,0,0L

Next, we seek a solution of linearSys in the form u 1 = P EI T 0 and v1 = Q EI T 0 and obtain

coefList = E-I T0 linearSys •. 9u1 -> IP EI # &M, v1 -> IQ EI # &M= •• Expand

8I P - Q, P + I Q<

The coefficient matrix of coefList is

coefMat = Outer@Coefficient, coefList, 8P, Q<D

88I, -1<, 81, I<<

and its adjoint is defined by

hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose

Hence, the right and left eigenvectors of coefMat are

rightVec = 81, c1< •. [email protected], c1<L@@1DD == 0, c1D@@1DD

81, I<

leftVec = 81, c1< •. Solve@[email protected], c1<L@@1DD == 0, c1D@@1DD

81, I<

Then the solution of the first-order equations can be expressed as

Chapter3.nb 57

sol1Form = rightVec A@T1, T2 D EI T0

8EI T0 A@T1 , T2D, I EI T0 A@T1 , T2 D<

and hence u 1 and v1 can be expressed in a pure function form as

order1Sol =
8u1 -> Function@8T0 , T1 , T2<, sol1Form@@1DD + Hsol1Form@@1DD •. conjugateRuleL •• EvaluateD,
v1 -> Function@8T0 , T1 , T2<, sol1Form@@2DD + Hsol1Form@@2DD •. conjugateRuleL •• EvaluateD<

8u1 ® Function@8T0 , T1 , T2<, EIT0 A@T1 , T2 D + E-I T0 A@T1 , T2DD,

•

v1 ® Function@8T0 , T1 , T2<, I EI T0 A@T1 , T2 D - I E-I T0 A@T1 , T2DD<

•

Ÿ Second-Order Equations

Substituting the first-order solution, order1Sol, into the second-order equations, eqOrder[2], we have

order2Eq = eqOrder@2D •. order1Sol •• ExpandAll;

order2Eq •. displayRule

:D0 u2 - v2 == -EI T0 HD1 AL - E-I T0 HD1 AL,

•

D0 v2 + u2 == -I EI T0 HD1 AL + I E-I T0 HD1AL - A2 E2 I T0 a2 - 2 A A a2 - E-2 I T0 A a2>

• • •2

To eliminate the terms that produce secular terms in u 2 and v2, we determine the vector proportional to EI T 0 in the right-hand
sides of order2Eq and obtain

STerms1 = Coefficient@#@@2DD, Exp@I T0DD & •ž order2Eq

8-AH1,0L@T1 , T2D, -I AH1,0L@T1 , T2 D<

Then, the condition for the elimination of the terms that produce secular terms, solvability condition, demands that STerms1
be orthogonal to the adjoint, left eigenvector. The result is

SCond1 = [email protected] == 0, AH1,0L @T1, T2 DE@@1DD

8AH1,0L@T1 , T2 D ® 0<

whose complex conjugate is

ccSCond1 = SCond1 •. conjugateRule

•H1,0L
:A @T1 , T2 D ® 0>

With this solvability condition, order2Eq becomes

58 Chapter3.nb

order2Eqm = order2Eq •. SCond1 •. ccSCond1;

order2Eqm •. displayRule

:D0 u2 - v2 == 0, D0v2 + u2 == -A2 E2 I T0 a2 - 2 A A a2 - E-2 IT0 A a2 >

• •2

whose particular solution can be expressed as

DSolve@order2Eqm, 8u2@T0 , T1 , T2D, v2 @T0 , T1, T2 D<, timeScalesD

DSolve::pde : Partial differential equation may not have

a general solution. Try loading Calculus`DSolveIntegrals` to find special solutions.

DSolveB:-v2@T0 , T1 , T2D + u2 @T0, T1 , T2 D == 0, u2@T0 , T1 , T2D + v2 @T0, T1, T2 D ==

H1,0,0L H1,0,0L

A@T1, T2 D2 a2 - 2 A@T1 , T2 D a2 A@T1 , T2D - E-2 I T0 a2 A@T1 , T2D >,

• • 2
-E2 I T0

8u2 @T0, T1 , T2 D, v2@T0 , T1 , T2D<, 8T0 , T1 , T2<F

Here, directly applying DSolve to solve for order2Eqm does not work. Since the differential operators on the left-hand sides
of order2Eqm are essentially ordinary-differential operators in terms of T0 , we can first transform order2Eqm into an ordinary-
differential-equation form and then use DSolve to obtain the particular solution as

sol2pForm =
DSolve@order2Eqm •. 8u2 -> Hu2 @#1D &L, v2 -> Hv2@#1D &L<, 8u2@T0 D, v2 @T0 D<, T0 D@@1DD •.
C@_D -> 0 •• TrigToExp •• ExpandAll

9u2 @T0D ® €€€€ E2 I T0 A@T1 , T2D2 a2 - 2 A@T1 , T2D a2 A@T1, T2 D + €€€€ E-2 I T0 a2 A@T1, T2 D ,
1 • 1 • 2

3 3

v2 @T0D ® €€€€ I E2 I T0 A@T1 , T2D2 a2 - €€€€ I E-2 I T0 a2 A@T1 , T2 D =

2 2 • 2

3 3

We can then express the solution of the second-order equations in a pure function form as

order2Sol = 8u2 -> Function@8T0 , T1 , T2<, sol2pForm@@1, 2DD •• EvaluateD,

v2 -> Function@8T0 , T1 , T2<, sol2pForm@@2, 2DD •• EvaluateD<

9u2 ®

FunctionA8T0, T1, T2 <, €€€€ E2 I T0 A@T1 , T2D2 a2 - 2 A@T1, T2 D a2 A@T1 , T2 D + €€€€ E-2 I T0 a2 A@T1 , T2 D E,
1 • 1 • 2

3 3

v2 ® FunctionA8T0 , T1 , T2<, €€€€ I E A@T1 , T2 D a2 - €€€€ I E a2 A@T1, T2 D E=

2 2 IT0 2 2 -2 I T0 • 2

3 3

Ÿ Third-Order Equations

Substituting the first- and second-order solutions into the third-order equations, eqOrder[3], yields
Chapter3.nb 59

order3Eq = eqOrder@3D •. order1Sol •. order2Sol •• ExpandAll;

order3Eq •. displayRule

9D0 u3 - v3 ==

-EI T0 HD2AL - E-I T0 HD2AL - €€€€ A E2 I T0 HD1AL a2 + 2 A HD1 AL a2 + 2 A HD1 AL a2 - €€€€ E-2 I T0 A HD1AL a2 ,
• 2 • • 2 • •
3 3
D0 v3 + u3 == -I EI T0 HD2 AL + I E-I T0 HD2AL - €€€€ I A E2 I T0 HD1AL a2 +
• 4
3

A HD1AL a2 - €€€€ A E
4 -2 I T0 • • 2 3 3 I T0 2 10 2 I T0 • 2 10 •2
€€€€ I E a2 + €€€€€€€ A E A a2 + €€€€€€ A E-I T0 A a22 -
3 3 3 3
A a3 - E-3 I T0 A a3 =
2 -3 I T0 •3 2 3 3 I T0 2 IT0 • -I T0 •2 •3
€€€€ E A a2 - A E a3 - 3 A E A a3 - 3 A E
3

To eliminate the terms that lead to secular terms from order3Eq, we calculate the vector proportional to EI T 0 in their right-
hand sides and obtain

STerms2 = Coefficient@#@@2DD, Exp@I T0DD & •ž order3Eq

9-AH0,1L@T1 , T2D, €€€€€€ A@T1, T2 D2 a22 A@T1, T2 D - 3 A@T1 , T2 D2 a3 A@T1, T2 D - I AH0,1L @T1, T2 D=
10 • •
3

Hence, the solvability condition at this order is

SCond2 = [email protected] == 0 •• ExpandAll

- €€€€€€ I A@T1, T2D2 a22 A@T1 , T2D + 3 I A@T1, T2 D2 a3 A@T1, T2D - 2 AH0,1L @T1, T2D == 0
10 • •
3

This solvability condition is the same as that obtained by attacking the second-order form of the governing equation.

Ÿ 3.4.3 First-Order Complex-Valued System

In this section, we first transform eq31g into a single first-order complex-valued equation using the transformation

transfRule = 9u@tD -> z@tD + z@tD, u¢ @tD -> I Iz@tD - z@tDM=;

• •

whose inverse is given by

zetaRule = SolveAtransfRule •. Rule -> Equal, 9z@tD, z@tD=E@@1DD

•

9z@tD ® €€€€ Hu@tD - I u¢@tDL, z@tD ® €€€€ Hu@tD + I u¢@tDL=

1 • 1
2 2
_
where z is the complex conjugate of z as defined by the complex conjugate rule

conjugateRule = 9A -> A, A -> A, z -> z, z -> z, Complex@0, n_D -> Complex@0, -nD=;
• • • •

It follows from eq31g that the acceleration is given by

60 Chapter3.nb

acceleration = Solve@eq31g, u²@tDD@@1DD

8u² @tD ® -u@tD - a2 u@tD2 - a3 u@tD3<

Differentiating zetaRule[[1]] once with respect to t and using acceleration and transfRule, we obtain the following first-
order complex-valued equation:

eq343a = D@zetaRule@@1DD, tD •. acceleration •. transfRule •. Rule -> Equal •• ExpandAll

z¢ @tD == I z@tD + €€€€ I a2 z@tD2 + €€€€ I a3 z@tD3 + I a2 z@tD z@tD +

1 1 •
2 2
€€€€ I a3 z@tD2 z@tD + €€€€ I a2 z@tD + €€€€ I a3 z@tD z@tD + €€€€ I a3 z@tD
3 • 1 • 2 3 • 2 1 • 3

2 2 2 2

To determine a second-order uniform expansion of the solution of eq343a using the method of multiple scales, we first
introduce

multiScales = 9z@tD -> z@T0, T1 , T2 D, z@tD -> z@T0, T1 , T2 D, z¢@tD -> dt@1D@z@T0 , T1, T2 DD=;
• •

••
and then expand z and z in the form

solRule = 9z -> ISumAej zj @#1, #2, #3D, 8j, 3<E &M, z -> ISumAej zj @#1, #2, #3D, 8j, 3<E &M=;
• •

Substituting multiScales and solRule into eq343a, expanding the result of small e, and discarding terms of order higher than
e3 , we obtain

eq343b = Heq343a •. multiScales •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e yields

eqEps = CoefficientList@Subtract žž eq343b, eD == 0 •• Thread •• Rest;

To place the linear operator on one side and the rest of the terms on the other, we define

eqOrder@i_D := HeqEps@@1, 1DD •. z1 -> zi L - # & •ž eqEps@@iDD •• Reverse

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

Array@eqOrder, 3D •. displayRule •• TableForm

D0 z1 - I z1 == 0
D0 z2 - I z2 == -HD1 z1L + €€€
1 • •2
I a2 z21 + I a2 z1 z1 + €€€12 I a2 z1

D0 z3 - I z3 == -HD1 z2L - D2z1 + €€€12 I a3 z31 + I a2 z1 z2 + €€€32 I a3 z21 z1 + I a2 z2 z1 + €€€32 I a3 z1 z1 + €€€12 I a3 z1 + I a

2
• • •2 •3

The general solution of the first-order equation, eqOrder[1], can be expressed in a pure function form as

sol1 = DSolve@eqOrder@1D, z1, timeScalesD@@1DD •. C@1D -> A

8z1 ® HEI #1 A@#2, #3D &L<

Chapter3.nb 61

whose complex conjugate is

ccsol1 = sol1 •. conjugateRule

8z1 ® HEComplex@0,-1D #1 A@#2, #3D &L<

• •

Substituting the first-order solution into the second-order equation, eqOrder[2], we have

order2Eq = eqOrder@2D •. sol1 •. ccsol1;

order2Eq •. displayRule

D0 z2 - I z2 == -EI T0 HD1 AL + €€€€ I A2 E2 IT0 a2 + I A A a2 + €€€€ I E-2 I T0 A a2

1 • 1 •2
2 2

Eliminating the terms that lead to seculr terms in z2 from order2Eq yields

SCond1 = CoefficientAorder2Eq@@2DD, EIT0 E == 0

-AH1,0L@T1 , T2D == 0

or

SCond1Rule = SolveASCond1, AH1,0L @T1 , T2DE@@1DD

8AH1,0L@T1 , T2 D ® 0<

With this solvability condition, the particular solution of order2Eq can be expressed in a pure function form as

sol2 = DSolve@order2Eq •. SCond1Rule, z2, timeScalesD@@1DD •. C@_D -> H0 &L

iEI #1 H0 &L@#2, #3D +

9z2 ® j
k
€€€€ E-2 I #1 a2 J3 E4 I #1 A@#2, #3D2 - 6 E2 I #1 A@#2, #3D A@#2, #3D - A@#2, #3D N &y
z=
• •
{
1 2

whose complex conjugate is

ccsol2 = sol2 •. conjugateRule;

Substituting the first- and second-order solutions in the third-order equation, eqOrder[3], we obtain

order3Eq = eqOrder@3D •. sol1 •. ccsol1 •. sol2 •. ccsol2 •• ExpandAll;

order3Eq •. displayRule

D0 z3 - I z3 == -EI T0 HD2 AL - A E2 I T0 HD1AL a2 + A HD1 AL a2 + A HD1AL a2 +

• •

€€€€ E-2 I T0 A HD1 AL a2 + €€€€ I A3 E3 I T0 a22 - €€€€ I A2 EI T0 A a22 - €€€€ I A E-I T0 A a22 +
1 • • 1 5 • 5 •2
3 3 3 3
1 •3 1 3 • 3 •2 1 •3
€€€€ I E-3 I T0 A a22 + €€€€ I A3 E3 I T0 a3 + €€€€ I A2 EI T0 A a3 + €€€€ I A E-I T0 A a3 + €€€€ I E-3 I T0 A a3
3 2 2 2 2

Eliminating the terms that produce secular terms in z3 from order3Eq demands the following condition
62 Chapter3.nb

SCond2 = CoefficientAorder3Eq@@2DD, EIT0 E == 0;

SCond2 •. displayRule

5 • 3 •
-HD2AL - €€€€ I A2 A a22 + €€€€ I A2 A a3 == 0
3 2

which is in full agreement with that obtained by attacking directly eq31g.

à 3.5 The Method of Averaging

As before, we use the method of variation of parameters to transform the dependent variable u into a and b according to

eq35a = u@tD == e a Cos@t + bD;

subject to the constraint

eq35b = D@eq35a, tD

u¢ @tD == -a e Sin@t + bD

where a and b are functions of t

tdependentRule = 8a -> a@tD, b -> b@tD<;

and e is a small dimensionless parameter that is a measure of the amplitude of oscillation. Differentiating eq35a with respect
to t yields

eq35c = D@eq35a •. tdependentRule, tD

u¢ @tD == e Cos@t + b@tDD a¢@tD - e a@tD Sin@t + b@tDD H1 + b¢ @tDL

Comparing eq35b and eq35c, we conclude that

cond35a = Expand@eq35c@@2DD - Heq35b@@2DD •. tdependentRuleLD == 0

e Cos@t + b@tDD a¢ @tD - e a@tD Sin@t + b@tDD b¢ @tD == 0

Differentiating eq35b once with respect to t yields

eq35d = D@eq35b •. tdependentRule, tD

u² @tD == -e Sin@t + b@tDD a¢@tD - e a@tD Cos@t + b@tDD H1 + b¢ @tDL

Substituting eq35a and eq35d into eq31g, we have

cond35b = eq31g •. H8eq35a •. tdependentRule, eq35d< •. Equal -> RuleL •• ExpandAll

e2 a@tD2 Cos@t + b@tDD2 a2 + e3 a@tD3 Cos@t + b@tDD3 a3 -

e Sin@t + b@tDD a¢ @tD - e a@tD Cos@t + b@tDD b¢ @tD == 0
Chapter3.nb 63

Solving cond35a and cond35b for a ¢@tD and b ¢ @tD, we have

transformedEq = Solve@8cond35a, cond35b<, 8a¢@tD, b¢ @tD<D@@1DD •. Rule -> Equal •• Simplify

8a¢ @tD == e a@tD2 Cos@t + b@tDD2 Sin@t + b@tDD Ha2 + e a@tD Cos@t + b@tDD a3 L,
b¢ @tD == e a@tD Cos@t + b@tDD3 Ha2 + e a@tD Cos@t + b@tDD a3 L<

where a is assumed to be different from zero in arriving at transformedEq[[2]]. Using trigonometric identities, we expand the
right-hand sides of transformedEq in Fourier series and obtain

ampEq1 = TrigReduce •ž transformedEq@@1DD

a¢ @tD == €€€€ H2 e a@tD2 Sin@t + b@tDD a2 + 2 e a@tD2 Sin@3 t + 3 b@tDD a2 +

1

2 e2 a@tD3 Sin@2 t + 2 b@tDD a3 + e2 a@tD3 Sin@4 t + 4 b@tDD a3 L

8

phaseEq1 = TrigReduce •ž transformedEq@@2DD

b¢ @tD == €€€€ H6 e a@tD Cos@t + b@tDD a2 + 2 e a@tD Cos@3 t + 3 b@tDD a2 +

1

3 e2 a@tD2 a3 + 4 e2 a@tD2 Cos@2 t + 2 b@tDD a3 + e2 a@tD2 Cos@4 t + 4 b@tDD a3 L

8

Since e is small, a ¢ @tD and b ¢ @tD are slowly varying functions of t. Then, according to the method of averaging, we can neglect
the fast-varying terms in ampEq1 and phaseEq1 and obtain the following averaged equations:

ampAvgEq = ampEq1 •. Sin@_. t + _D -> 0

a¢ @tD == 0

phaseAvgEq = phaseEq1 •. Cos@_. t + _D -> 0

b¢ @tD == €€€€ e2 a@tD2 a3

3
8

Whereas ampAvgEq agrees with ampEq, phaseAvgEq does not agree with phaseEq obtained in the preceding section by using
5 2 2 2
the method of multiple scales. There is a term €€€€12
€€€ e a a2 missing from phaseAvgEq. Following the details of the solution in
the preceding section, one finds that this term is the result of the interaction of the first- and second-order approximations.
This interaction was not taken into account in arriving at ampAvgEq and phaseAvgEq. To include the effect of this interac-
tion, we need to carry out the solutions of ampEq1 and phaseEq1 to higher order. This is accomplished by using the general-
ized method of averaging, which is discussed next, or its variant the Krylov-Bogoliubov-Mitropolsky technique, which is
discussed in Section 3.7.

à 3.6 The Generalized Method of Averaging

To apply this method, we introduce the variable f = t + b and rewrite ampEq1 and phaseEq1 as
64 Chapter3.nb

ampEq2 = ampEq1 •. Sin@n_. t + a_D :> Sin@Expand@n Hf@tD - b@tDL + aDD

a¢ @tD == €€€€ H2 e a@tD2 Sin@f@tDD a2 +

1

2 e a@tD2 Sin@3 f@tDD a2 + 2 e2 a@tD3 Sin@2 f@tDD a3 + e2 a@tD3 Sin@4 f@tDD a3 L

8

phaseEq2 = f¢ @tD == 1 + phaseEq1@@2DD •. Cos@n_. t + a_D :> Cos@Expand@n Hf@tD - b@tDL + aDD

f¢ @tD == 1 + €€€€ H6 e a@tD Cos@f@tDD a2 + 2 e a@tD Cos@3 f@tDD a2 +

1

3 e2 a@tD2 a3 + 4 e2 a@tD2 Cos@2 f@tDD a3 + e2 a@tD2 Cos@4 f@tDD a3 L

8

We seek approximate solutions to ampEq2 and phaseEq2 in the form

asol = a@tD == a0 @tD + SumAei ai @a0 @tD, f0@tDD, 8i, 2<E

a@tD == a0 @tD + e a1 @a0@tD, f0 @tDD + e2 a2@a0 @tD, f0@tDD

phisol = f@tD == f0 @tD + SumAei fi @a0@tD, f0 @tDD, 8i, 2<E

f@tD == f0 @tD + e f1 @a0@tD, f0 @tDD + e2 f2@a0 @tD, f0@tDD

where a 0 @tD and f0 @tD are expanded in power series in e as

a0Eq = a0 ¢@tD == SumAei Ai@a0 @tDD, 8i, 2<E

a¢0 @tD == e A1@a0 @tDD + e2 A2@a0 @tDD

phi0Eq = f0 ¢@tD == 1 + SumAei Fi @a0@tDD, 8i, 2<E

f¢0 @tD == 1 + e F1 @a0@tDD + e2 F2 @a0@tDD

The functions a 1, a 2 , …, a n and f1, f2 , …, fn are fast varying functions of f0 , while it follows from a0Eq and phi0Eq that a 0 ,
and hence the An and Fn are slowly varying functions of t.

To the second approximation, we differentiate a@tD and f@tD once with respect to t, use a0Eq and phi0Eq, expand the result for
small e, discard terms of order higher than e2, and obtain

aEq1 = HD@asol, tD •. H8a0Eq, phi0Eq< •. Equal -> RuleL •• ExpandAllL •. e n_•;n>2 -> 0

a¢ @tD == e A1@a0 @tDD + e2 A2@a0 @tDD + e a1 @a0 @tD, f0 @tDD +

H0,1L

e2 F1@a0 @tDD aH0,1L

1 @a0@tD, f0 @tDD + e2 aH0,1L
2 @a0 @tD, f0@tDD + e2 A1 @a0@tDD aH1,0L
1 @a0 @tD, f0 @tDD

phiEq1 = HD@phisol, tD •. H8a0Eq, phi0Eq< •. Equal -> RuleL •• ExpandAllL •. e n_•;n>2 -> 0

f¢ @tD == 1 + e F1 @a0@tDD + e2 F2 @a0@tDD + e fH0,1L @a0@tD, f0 @tDD +

e2 F1@a0 @tDD f1 @a0@tD, f0 @tDD + e2 f2 @a0 @tD, f0@tDD + e2 A1 @a0@tDD f1 @a0 @tD, f0 @tDD
H0,1L H0,1L H1,0L
1

OHe2L, and rewrite their right-hand sides as

Next, we substitute asol and phisol into ampEq2 and phaseEq2, expand the right-hand sides for small e keeping terms up to
Chapter3.nb 65

ampEqrhs = Series@ampEq2@@2DD •. H8asol, phisol< •. Equal -> RuleL, 8e, 0, 2<D •• Normal

€€€€ e H2 Sin@f0 @tDD a2 a0 @tD2 + 2 Sin@3 f0 @tDD a2 a0 @tD2L +

1
8

€€€€ e2 H2 Sin@2 f0@tDD a3 a0 @tD3 + Sin@4 f0@tDD a3 a0@tD3 +

1

2 a2 H2 Sin@f0 @tDD a0@tD a1 @a0@tD, f0 @tDD + Cos@f0 @tDD a0@tD2 f1@a0 @tD, f0@tDDL +
8

2 a2 H2 Sin@3 f0 @tDD a0@tD a1 @a0@tD, f0 @tDD + 3 Cos@3 f0 @tDD a0 @tD2 f1 @a0@tD, f0 @tDDLL

phaseEqrhs = Series@phaseEq2@@2DD •. H8asol, phisol< •. Equal -> RuleL, 8e, 0, 2<D •• Normal

1 + €€€€ e H6 Cos@f0@tDD a2 a0 @tD + 2 Cos@3 f0 @tDD a2 a0 @tDL +

1
8

€€€€ e H3 a3 a0 @tD2 + 4 Cos@2 f0@tDD a3 a0 @tD2 + Cos@4 f0@tDD a3 a0@tD2 +

1 2

6 a2 HCos@f0 @tDD a1@a0 @tD, f0 @tDD - Sin@f0@tDD a0 @tD f1@a0 @tD, f0@tDDL +
8

2 a2 HCos@3 f0 @tDD a1@a0 @tD, f0@tDD - 3 Sin@3 f0@tDD a0 @tD f1 @a0 @tD, f0@tDDLL

Equating coefficients of like powers of e from aEq1, ampEqrhs, phiEq1, and phaseEqrhs, we obtain

ampEq3 = CoefficientList@#, eD & •ž HaEq1@@2DD == ampEqrhsL •• Thread •• Rest

9A1 @a0@tDD + a1 @a0@tD, f0@tDD == €€€€ H2 Sin@f0@tDD a2 a0@tD2 + 2 Sin@3 f0 @tDD a2 a0 @tD2L,
H0,1L 1

A2 @a0@tDD + F1@a0 @tDD a1 @a0@tD, f0 @tDD + a2 @a0@tD, f0 @tDD +

H0,1L H0,1L
8

A1 @a0@tDD a1 @a0 @tD, f0@tDD == €€€€ H2 Sin@2 f0 @tDD a3 a0 @tD3 + Sin@4 f0 @tDD a3 a0 @tD3 +
H1,0L 1

2 a2 H2 Sin@f0 @tDD a0@tD a1 @a0@tD, f0 @tDD + Cos@f0 @tDD a0@tD2 f1@a0 @tD, f0@tDDL +
8

2 a2 H2 Sin@3 f0 @tDD a0@tD a1 @a0@tD, f0 @tDD + 3 Cos@3 f0@tDD a0 @tD2 f1 @a0@tD, f0 @tDDLL=

phaseEq3 = CoefficientList@#, eD & •ž HphiEq1@@2DD == phaseEqrhsL •• Thread •• Rest

9F1 @a0@tDD + f1 @a0@tD, f0@tDD == €€€€ H6 Cos@f0@tDD a2 a0@tD + 2 Cos@3 f0@tDD a2 a0 @tDL,
H0,1L 1

F2 @a0@tDD + F1@a0 @tDD f1 @a0@tD, f0 @tDD +

H0,1L
8

@a0 @tD, f0 @tDD + A1 @a0 @tDD f1 @a0 @tD, f0@tDD ==

H0,1L H1,0L
f2

€€€€ H3 a3 a0@tD + 4 Cos@2 f0 @tDD a3 a0 @tD2 + Cos@4 f0 @tDD a3 a0 @tD2 +

1 2

6 a2 HCos@f0 @tDD a1@a0 @tD, f0@tDD - Sin@f0@tDD a0 @tD f1@a0 @tD, f0@tDDL +
8

2 a2 HCos@3 f0 @tDD a1@a0 @tD, f0@tDD - 3 Sin@3 f0@tDD a0 @tD f1 @a0@tD, f0@tDDLL=

Next, we use the method of separation of variables to separate fast and slowly varying terms in the first-order equations,
ampEq3[[1]] and phaseEq3[[1]]. The slowly varying parts yield

cond36a = 8A1 @a0@tDD -> 0, F1 @a0@tDD -> 0<;

whereas the fast varying parts yield

66 Chapter3.nb

order1Eq = 8ampEq3@@1DD, phaseEq3@@1DD< •. cond36a

9a1 @a0 @tD, f0 @tDD == €€€€ H2 Sin@f0 @tDD a2 a0 @tD2 + 2 Sin@3 f0@tDD a2 a0@tD2 L,
H0,1L 1
8

@a0 @tD, f0 @tDD == €€€€ H6 Cos@f0 @tDD a2 a0 @tD + 2 Cos@3 f0 @tDD a2 a0 @tDL=
H0,1L 1
f1
8

equations, replace f0@tD with s, and obtain

To determine the solutions of order1Eq using DSolve, we transform them from partial-differential to ordinary-differential

order1EqTransf = order1Eq •. 8a1 -> Ha1 @#2D &L, f1 -> Hf1@#2D &L< •. f_@tD -> f •. f0 -> s

9a¢1 @sD == €€€€ H2 Sin@sD a20 a2 + 2 Sin@3 sD a20 a2 L, f¢1@sD == €€€€ H6 Cos@sD a0 a2 + 2 Cos@3 sD a0 a2 L=
1 1
8 8

Instead of replacing f0@tD, we could have symbolized it. Using DSolve, we find that a 1 and f1 are given by

rule36@1D = DSolve@order1EqTransf@@1DD, a1 @sD, sD@@1DD •. C@_D -> 0 •.

8s -> f0 @tD, a0 -> a0 @tD< •. a1@arg_D -> a1 @a0@tD, argD

9a1 @a0@tD, f0 @tDD ® €€€€€€ H-3 Cos@f0 @tDD a2 a0 @tD2 - Cos@3 f0@tDD a2 a0@tD2 L=
1
12

rule36@2D = DSolve@order1EqTransf@@2DD, f1 @sD, sD@@1DD •. C@_D -> 0 •.

8s -> f0 @tD, a0 -> a0 @tD< •. f1@arg_D -> f1 @a0@tD, argD

9f1 @a0@tD, f0 @tDD ® €€€€€€ H9 Sin@f0@tDD a2 a0@tD + Sin@3 f0 @tDD a2 a0 @tDL=

1
12

which can be combined into

cond36b = Join@rule36@1D, rule36@2DD;

Substituting the slow- and fast- varying components of the first-order solution, cond36a and cond36b, into the second-order
equations, we have
Chapter3.nb 67

order2Eq = 8ampEq3@@2DD, phaseEq3@@2DD< •. cond36a •. cond36b

9A2 @a0@tDD + a2 @a0@tD, f0@tDD == €€€€ i

j2 Sin@2 f0@tDD a3 a0@tD3 + Sin@4 f0 @tDD a3 a0 @tD3 +
H0,1L
8 k
1

i 1
2 a2 j €€€€€€ Cos@f0 @tDD a0@tD2 H9 Sin@f0@tDD a2 a0@tD + Sin@3 f0 @tDD a2 a0 @tDL +
k 12
€€€€ Sin@f0 @tDD a0 @tD H-3 Cos@f0 @tDD a2 a0 @tD2 - Cos@3 f0@tDD a2 a0@tD2 Ly z+
{
1

i
2 a2 j €€€€ Cos@3 f0 @tDD a0@tD H9 Sin@f0@tDD a2 a0@tD + Sin@3 f0 @tDD a2 a0 @tDL +
6

k4
1 2

yz
y,
€€€€ Sin@3 f0 @tDD a0 @tD H-3 Cos@f0 @tDD a2 a0 @tD2 - Cos@3 f0@tDD a2 a0 @tD2 Lz
{{
1

@a0@tD, f0@tDD == €€€€ i

j3 a3 a0@tD2 + 4 Cos@2 f0@tDD a3 a0@tD2 + Cos@4 f0 @tDD
6
F2 @a0@tDD + f2
H0,1L
8 k
1

i 1
a3 a0@tD2 + 6 a2 j- €€€€€€€ Sin@f0 @tDD a0@tD H9 Sin@f0@tDD a2 a0@tD + Sin@3 f0@tDD a2 a0@tDL +
k 12
y+
€€€€€€€ Cos@f0 @tDD H-3 Cos@f0@tDD a2 a0@tD2 - Cos@3 f0 @tDD a2 a0 @tD2Lz
{
1

i
2 a2 j- €€€€ Sin@3 f0 @tDD a0@tD H9 Sin@f0@tDD a2 a0@tD + Sin@3 f0 @tDD a2 a0 @tDL +
12

k 4
1

€€€€€€€ Cos@3 f0 @tDD H-3 Cos@f0@tDD a2 a0@tD2 - Cos@3 f0 @tDD a2 a0 @tD2Ly zy

z=
{{
1
12

Next, we use trigonometric identities to expand the right-hand sides of the second-order equations in Fourier series and
obtain

order2Eqrhs = H#@@2DD •• TrigReduce •• ExpandL & •ž order2Eq

9- €€€€€€ Sin@2 f0@tDD a22 a0@tD3 + €€€€€€€ Sin@4 f0 @tDD a22 a0 @tD3 +
9 5
32 24

Sin@6 f0 @tDD a22 a0 @tD3 + €€€€ Sin@2 f0@tDD a3 a0 @tD3 + €€€€ Sin@4 f0 @tDD a3 a0@tD3 ,
1 1 1
€€€€€€€
96 4 8

a2 a0@tD - €€€€€€€ Cos@2 f0 @tDD a2 a0 @tD + €€€€ Cos@4 f0@tDD a22 a0 @tD2 +
5 2 2 3 2 2 1
- €€€€€€
12 16 4
Cos@6 f0 @tDD a2 a0 @tD + €€€€ a3 a0 @tD + €€€€ Cos@2 f0@tDD a3 a0@tD2 + €€€€ Cos@4 f0 @tDD a3 a0 @tD2=
1 2 2 3 2 1 1
€€€€€€€
48 8 2 8

Since we are seeking an expansion valid to OHe2L, we do not need to solve for a 2 and f2. All we need to do is to investigate
the above expressions to determine the slowly varying parts and determine A2 and F2 . The result is

cond36c = 8A2 @a0@tDD, F2 @a0@tDD< -> Horder2Eqrhs •. 8Sin@_D -> 0, Cos@_D -> 0<L •• Thread

9A2 @a0@tDD ® 0, F2 @a0@tDD ® - €€€€€€€ a22 a0@tD2 + €€€€ a3 a0@tD2 =

5 3
12 8

Substituting the fast-varying components of the first-order solution, cond36b, into the expansions for a and f and discarding
terms of order higher than e, we obtain, to the second approximation, that

asolF = asol •. cond36b •. e2 -> 0

a@tD == a0 @tD + €€€€€€€ e H-3 Cos@f0@tDD a2 a0@tD2 - Cos@3 f0 @tDD a2 a0@tD2 L

1
12
68 Chapter3.nb

phisolF = phisol •. cond36b •. e2 -> 0

f@tD == €€€€€€€ e H9 Sin@f0 @tDD a2 a0@tD + Sin@3 f0@tDD a2 a0@tDL + f0@tD

1
12

Substituting the solw-varying components of the first- and second-order solutions, cond36a and cond36c, into a0Eq and
phi0Eq yields

a0Eq1 = a0Eq •. cond36a •. cond36c

a¢0 @tD == 0

phi0Eq1 = phi0Eq •. cond36a •. cond36c

f¢0 @tD == 1 + e2 i € a2 a0 @tD2 + €€€€ a3 a0 @tD2y

j- €€€€€€ z
k 12 {
5 2 3
8

It follows from the above equations that a 0 = constant and

phi0Rule = DSolve@phi0Eq1 •. a0@tD -> a0 , f0@tD, tD@@1DD •. C@1D -> b0

9f0 @tD ® t i
5 2 2 2 3 2 2 y
j1 - €€€€€€
€ e a0 a2 + €€€€ e a0 a3 z + b0=
k 12 8 {

where b 0 is a constant. Substituting these expansions into eq35a, we obtain the following second-order expansion:

eq36a = u@tD == e a@tD Cos@f@tDD •. H8asolF, phisolF< •. Equal -> RuleL

u@tD == e CosA €€€€€€ e H9 Sin@f0@tDD a2 a0@tD + Sin@3 f0 @tDD a2 a0 @tDL + f0 @tDE

1

i
ja0 @tD + €€€€€€€ e H-3 Cos@f0@tDD a2 a0@tD2 - Cos@3 f0 @tDD a2 a0@tD2 Ly
z
12

k {
1
12

To compare the present solution with those obtained by using the method of multiple scales and the Lindstedt-Poincaré
technique, we expand the circular functions in eq36a for small e about f0 and obtain

u@tD == HSeries@eq36a@@2DD, 8e, 0, 2<D •• Normal •• TrigReduce •• ExpandL

u@tD == e Cos@f0@tDD a0 @tD - €€€€ e2 a2 a0 @tD2 + €€€€ e2 Cos@2 f0@tDD a2 a0 @tD2

1 1
2 6

This expansion and phi0Rule are in full agreement with those obtained by using the method of multiple scales.

à 3.7 The Krylov-Bogoliubov-Mitropolsky Technique

In this section, we describe a variant of the generalized method of averaging, namely the Krylov-Bogoliubov-Mitropolsky
technique. It is often referred to as the asymptotic method. When the nonlinear terms are neglected, the solution of eq31g
is

linearSol = u == e a Cos@t + bD;

Chapter3.nb 69

where a and b are constants and e is a small dimensionless parameter that is a measure of the amplitude. When the nonlinear
terms are included, we consider linearSol to be the first term in an approximate solution of eq31g but with slowly varying
rather than constant a and b. Moreover, we introduce the fast scale f = t + b and use a to represent the slow variations.
Thus, we seek a second-order uniform expansion of the solution of eq31g in the form

uSol = u@a@tD, f@tDD == e a@tD Cos@f@tDD + SumAei ui @a@tD, f@tDD, 8i, 2, 3<E

u@a@tD, f@tDD == e a@tD Cos@f@tDD + e2 u2@a@tD, f@tDD + e3 u3 @a@tD, f@tDD

Again, terms of order e3 have been included in the expansion, but these terms will not be included in the final result. Since a
and b are slowly varying functions of t, we express them in power series of e in terms of a as

apEq = a '@tD == SumAei Ai@a@tDD, 8i, 2<E

a¢ @tD == e A1@a@tDD + e2 A2 @a@tDD

phipEq = f'@tD == 1 + SumAei Fi @a@tDD, 8i, 2<E

f¢ @tD == 1 + e F1 @a@tDD + e2 F2@a@tDD

In what follows, we need a ² @tD and f²@tD. To this end, we differentiate apEq and phipEq with respect to t, use these equations
to express a ¢ @tD and f¢ @tD in terms of the Ai @a@tDD and Fi@a@tDD, discard terms of order higher than e2 , and obtain

appEq = HD@apEq, tD •. HapEq •. Equal -> RuleL •• ExpandAllL •. e n_•;n>2 -> 0

a² @tD == e2 A1 @a@tDD A¢1@a@tDD

phippEq = HD@phipEq, tD •. HapEq •. Equal -> RuleL •• ExpandAllL •. e n_•;n>2 -> 0

f² @tD == e2 A1 @a@tDD F¢1@a@tDD

Thus, this method can be viewed as a multiple scales procedure with a and f being the scales.

Substituting the assumed expansion for u@tD, expressed in function form, and the expansions for the derivatives of a@tD and
f@tD into eq31g, expanding the result for small e, and keeping terms up to order e3 , we have

eq37a = Heq31g •. u -> Function@t, uSol@@2DD •• EvaluateD •.

H8apEq, phipEq, appEq, phippEq< •. Equal -> RuleL •• ExpandAllL •. e n_•;n>3 -> 0

e2 a@tD2 Cos@f@tDD2 a2 + e3 a@tD3 Cos@f@tDD3 a3 - 2 e2 Sin@f@tDD A1 @a@tDD -

2 e3 Sin@f@tDD A2 @a@tDD + e2 u2@a@tD, f@tDD + 2 e3 a@tD Cos@f@tDD a2 u2 @a@tD, f@tDD +
e3 u3@a@tD, f@tDD - 2 e2 a@tD Cos@f@tDD F1@a@tDD - 2 e3 Sin@f@tDD A1@a@tDD F1 @a@tDD -
e3 a@tD Cos@f@tDD F1 @a@tDD2 - 2 e3 a@tD Cos@f@tDD F2@a@tDD + e3 Cos@f@tDD A1@a@tDD A¢1 @a@tDD -
e3 a@tD Sin@f@tDD A1 @a@tDD F¢1@a@tDD + e2 uH0,2L@a@tD, f@tDD +
2 e3 F1@a@tDD u2 @a@tD, f@tDD + e3 u3 @a@tD, f@tDD + 2 e3 A1 @a@tDD u2 @a@tD, f@tDD == 0
H0,2L H0,2L H1,1L
2

Equating coefficients of like powers of e in eq37a yields

70 Chapter3.nb

eqEps = TableACoefficientAeq37a@@1DD, ei E == 0, 8i, 2, 3<E •. f_@tD -> f

9a2 Cos@fD2 a2 - 2 Sin@fD A1@aD + u2@a, fD - 2 a Cos@fD F1@aD + u2 @a, fD == 0,

H0,2L

a Cos@fD a3 - 2 Sin@fD A2@aD + 2 a Cos@fD a2 u2 @a, fD + u3 @a, fD -

3 3

2 Sin@fD A1@aD F1 @aD - a Cos@fD F1 @aD2 - 2 a Cos@fD F2 @aD + Cos@fD A1 @aD A¢1 @aD -
a Sin@fD A1@aD F¢1 @aD + 2 F1 @aD u2 @a, fD + u3 @a, fD + 2 A1 @aD u2 @a, fD == 0=
H0,2L H0,2L H1,1L

Next, we use trigonometric identities to expand the nonhomogeneous terms in eqEps[[1]] in a Fourier series and obtain

order2Eq = TrigReduce •ž eqEps@@1DD

€€€€ Ia2 a2 + a2 Cos@2 fD a2 - 4 Sin@fD A1 @aD + 2 u2@a, fD - 4 a Cos@fD F1@aD + 2 u2 @a, fDM == 0
1 H0,2L
2

Eliminating the terms that produce secular terms in u 2 from order2Eq demands that

rule37a =
Solve@Coefficient@order2Eq@@1DD, 8Cos@fD, Sin@fD<D == 0 •• Thread, 8F1 @aD, A1@aD<D@@1DD

8F1 @aD ® 0, A1@aD ® 0<

Then, the solution of the resulting order2Eq can be written as

rule37b = DSolve@order2Eq •. rule37a, u2@a, fD, 8a, f<D@@1DD •. C@_D -> H0 &L •• Simplify

9u2 @a, fD ® €€€€ a2 H-3 + Cos@2 fDL a2 =

1
6

which can be expressed in function form as

rule37c = 8u2 -> Function@8a, f<, rule37b@@1, 2DD •• EvaluateD<

9u2 ® FunctionA8a, f<, €€€€ a2 H-3 + Cos@2 fDL a2 E=

1
6

Substituting rule37a and rule37c into eqEps[[2]], we obtain

eq37b = eqEps@@2DD •. rule37a •. rule37c

€€€€ a3 Cos@fD H-3 + Cos@2 fDL a22 + a3 Cos@fD3 a3 -

1

2 Sin@fD A2@aD + u3@a, fD - 2 a Cos@fD F2@aD + uH0,2L @a, fD == 0

3
3

Eliminating the terms that produce secular terms in u 3 from eq37b demands that

eq37c = Coefficient@eq37b@@1DD •• TrigReduce, 8Cos@fD, Sin@fD<D == 0 •• Thread

:- €€€€ a3 a22 + €€€€€€€€€€€€€€€€ - 2 a F2 @aD == 0, -2 A2 @aD == 0>

5 3 a3 a3
6 4

or
Chapter3.nb 71

rule37d = Solve@eq37c, 8A2@aD, F2 @aD<D@@1DD

9A2 @aD ® 0, F2@aD ® - €€€€€€€ a2 H10 a22 - 9 a3 L=

1
24

Substituting rule37a and rule37d into apEq and phipEq yields

apEq •. Hrule37a~ Join~rule37d •. a -> a@tDL

a¢ @tD == 0

phipEq •. Hrule37a~ Join~rule37d •. a -> a@tDL •• ExpandAll

f¢ @tD == 1 - €€€€€€€ e2 a@tD2 a22 + €€€€ e2 a@tD2 a3

5 3
12 8

which are in full agreement with those obtained by using the method of multiple scales and the generalized method of
averaging.

à 3.8 The Method of Normal Forms

In this section, we use the method of normal forms to determine a second-order uniform expansion of the solution of eq31g.
To this end, we start with the corresponding first-order complex-valued equation

eq343a = z¢ @tD == I z@tD + €€€€ I a2 z@tD2 + €€€€ I a3 z@tD3 + I a2 z@tD z@tD +

1 1 •
2 2

€€€€ I a3 z@tD z@tD + €€€€ I a2 z@tD + €€€€ I a3 z@tD z@tD + €€€€ I a3 z@tD ;
3 2• 1 • 2 3 • 2 1 • 3

2 2 2 2

According to the method of normal forms, we introduce the near-identity transformation

basicTerms = 8h@tD, •
h@tD<;

zetaRule = 9
z -> FunctionAt, e h@tD + SumAej hj žž basicTerms, 8j, 2, 3<E •• EvaluateE,
h@tD + SumAej hj žž basicTerms, 8j, 2, 3<E •• EvaluateE=
• •
z -> FunctionAt, e •

8z ® Function@t, e h@tD + e2 h2@h@tD, h@tDD + e3 h3 @h@tD, h@tDDD,

• •

z ® Function@t, e h@tD + e2 h2@h@tD, h@tDD + e3 h3 @h@tD, h@tDDD<

• • • • • •

that results in the simplest possible equation

etaRule = 9h¢ @tD -> I h@tD + SumAej gj @tD, 8j, 2<E=

8h¢ @tD ® I h@tD + e g1 @tD + e2 g2 @tD<

where the overbar denotes the complex conjugate

conjugateRule = 8h -> •
h, •
h -> h, g -> •
g, •
g -> g, Complex@0, n_D -> Complex@0, -nD<;
72 Chapter3.nb

Substituting the expansion for z, the zetaRule, into eq343a, using etaRule, expanding the result for small e, and discarding
terms of order higher than e3, we have

eq38a =
Heq343a •. zetaRule •. etaRule •. HetaRule •. conjugateRuleL •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e in eq38a yields

eqEps = TableACoefficientASubtract žž eq38a, eiE == 0, 8i, 2, 3<E

9- €€€€ I a2 h@tD2 - I a2 h@tD h@tD - €€€€ I a2 h@tD2 + g1@tD -

1 • 1 •

I h2 @h@tD, h@tDD - I h@tD h2 @h@tD, h@tDD + I h@tD h2 @h@tD, h@tDD == 0,

H0,1L H1,0L
2 2
• • • •

- €€€€ I a3 h@tD3 - €€€€ I a3 h@tD2 h@tD - €€€€ I a3 h@tD h@tD2 - €€€€ I a3 h@tD3 + g2 @tD -
1 3 • 3 • 1 •

I a2 h@tD h2 @h@tD, h@tDD - I a2 h@tD h2@h@tD, h@tDD - I h3@h@tD, h@tDD -

2 2 2 2
• • • •

I a2 h@tD h2 @h@tD, h@tDD - I a2 h@tD h2@h@tD, h@tDD + g1@tD h2 @h@tD, h@tDD -

• • • • • • H0,1L •

I h@tD h3 @h@tD, h@tDD + g1@tD h2 @h@tD, h@tDD + I h@tD h3 @h@tD, h@tDD == 0=

• H0,1L • H1,0L • H1,0L •

Ÿ Second-Order Solution

We choose the h i to eliminate as many terms from eqEps, thereby reducing them into their simplest possible form. It turns out
that we can eliminate all nonresonance terms. To determine the resonance terms in eqEps[[1]], we note that all of the possible
forms of its nonhomogeneous terms are

possibleQTerms = Outer@Times, basicTerms, basicTermsD •• Flatten •• Union

9h@tD2 , h@tD h@tD, h@tD2=

• •

It follows from the linear parts of the etaRule that

form = 9h@tD -> EI t, •

h@tD -> E-I t =;

Hence, the possible resonance terms are given by

ResonantQTerm = IE-I t possibleQTerms •. form •. E _t -> 0M possibleQTerms •• Union •• Rest

8<

Consequently, there are no resonance terms in the second-order problem and the nonresonance terms can be defined as

NRQT = Complement@possibleQTerms, ResonantQTermD

9h@tD2 , h@tD h@tD, h@tD2=

• •

We associate with each of them a coefficient according to

Chapter3.nb 73

coeffsQ = Table@Gj, 8j, Length@NRQTD<D

8G1 , G2, G3<

Therefore, h 2 and its complex conjugate have the form

hFormQ = 9h2 -> [email protected] •. Thread@basicTerms -> 8#1, #2<DD &L,

h2 -> [email protected] •. conjugateRule •. Thread@basicTerms -> 8#1, #2<DD &L=
•

8h2 ® H#12 G1 + #1 #2 G2 + #22 G3 &L, h2 ® H#22 G1 + #1 #2 G2 + #12 G3 &L<

•

Substituting for h 2 in eqEps[[1]], equating the coefficients of the possible nonresonance terms to zero, and solving the
resulting equations for the Gi, we obtain

coeffsQRule = Solve@Coefficient@eqEps@@1, 1DD •. hFormQ, NRQTD == 0 •• Thread, coeffsQD@@1DD

:G1 ® €€€€€€ , G2 ® -a2, G3 ® - €€€€€€ >

a2 a2
2 6

We choose g 1 to eliminate the resonance terms in eqEps[[1,1]] according to

gRuleQ = g1 @tD -> - Coefficient@eqEps@@1, 1DD, ResonantQTermD.ResonantQTerm

g1 @tD ® 0

In this case, there are on resonance terms and hence g 1 = 0.

Ÿ Third-Order Equations

Substituting the second-order results into the third-order equation yields

order3expr =
eqEps@@2, 1DD •. hFormQ •. coeffsQRule •. gRuleQ •. HgRuleQ •. conjugateRuleL •• Expand

- €€€€ I a22 h@tD3 - €€€€ I a3 h@tD3 + €€€€ I a22 h@tD2 h@tD - €€€€ I a3 h@tD2 h@tD +
1 1 5 • 3 •
3 2 3 2
€€€€ I a22 h@tD h@tD2 - €€€€ I a3 h@tD h@tD2 - €€€€ I a22 h@tD3 - €€€€ I a3 h@tD3 + g2 @tD -
5 • 3 • 1 • 1 •

I h3 @h@tD, h@tDD - I h@tD h3 @h@tD, h@tDD + I h@tD h3 @h@tD, h@tDD

H0,1L H1,0L
3 2 3 2
• • • •

The nonhomogeneous terms in order3expr are proportional to

possibleCTerms = Outer@Times, possibleQTerms, basicTermsD •• Flatten •• Union

9h@tD3 , h@tD2 h@tD, h@tD h@tD2 , h@tD3=

• • •

Next, we determine the resonance terms according to

74 Chapter3.nb

ResonantCTerm = IE-I t possibleCTerms •. form •. E _t -> 0M possibleCTerms •• Union •• Rest

8h@tD2 h@tD<
•

Then, the nonresonance terms are the complement of these resonance terms; that is,

NRCT = Complement@possibleCTerms, ResonantCTermD

9h@tD3 , h@tD h@tD2, h@tD3 =

• •

Again, we associate coefficients with these terms as follows:

coeffsC = Table@Lj, 8j, Length@NRCTD<D

8L1 , L2, L3<

Consequently, h 3 and its complex conjugate have the forms

hFormC = 9h3 -> [email protected] •. Thread@basicTerms -> 8#1, #2<DD &L,

h3 -> [email protected] •. conjugateRule •. Thread@basicTerms -> 8#1, #2<DD &L=
•

8h3 ® H#13 L1 + #1 #22 L2 + #23 L3 &L, h3 ® H#23 L1 + #12 #2 L2 + #13 L3 &L<

•

Substituting for h 3 into order3expr, equating the coefficient of each possible nonresonance term to zero, and solving the
resulting algebraic equations for the Li, we obtain

coeffsCRule = Solve@Coefficient@order3expr •. hFormC, NRCTD == 0 •• Thread, coeffsCD@@1DD

9L1 ® €€€€€€ H2 a22 + 3 a3L, L2 ® €€€€€€€ H10 a22 - 9 a3 L, L3 ® €€€€€€€ H-2 a22 - 3 a3 L=
1 1 1
12 12 24

Choosing g 2 to eliminate the resonance terms from order3expr yields

gRuleC = g2 @tD -> - Coefficient@order3expr, ResonantCTermD.ResonantCTerm

i 5 I a22 3 I a3 y
g2 @tD ® -j
j z h@tD2 •
j €€€€€€€€€€€€€€ - €€€€€€€€€€€€€ z
z h@tD
k 3 2 {

Combining etaRule, gRuleQ, and gRuleC and letting h = A@tDEI t , we obtain the modulation equation

moduEq = 2 I E-I t Hh¢ @tD - Hh¢ @tD •. etaRule •. gRuleQ •. gRuleCLL == 0 •.

9h -> IA@#D EI # &M, • h -> IA@#D E-I # &M= •• ExpandAll
•

- €€€€€€ e2 A@tD2 a22 A@tD + 3 e2 A@tD2 a3 A@tD + 2 I A¢@tD == 0

10 • •
3

which is in full agreement with that obtained by using the other techniques.
Chapter 4
Forced Oscillations of the Duffing
Equation
Off@General::spell1D

In contrast with the preceding chapters, which deal with free oscillations, this chapter and the next two deal with forced
oscillations. We consider

FDuffingEq = u² @tD + w2 u@tD + 2 m u¢@tD + d u@tD2 + a u@tD3 == F Cos@W tD;

`

where m` is a positive constant. In this chapter, we determine second-order solutions to FDuffingEq beginning with the
straightforward expansion in the next section. We investigate this straightforward expansion and determine under what
conditions it breaks down. This leads to the so-called resonance values of W. In Section 4.2, we use the method of multiple
scales to determine second-order uniform expansions of the solutions of FDuffingEq for all resonance cases, including the
effect of light viscous damping. In Sections 4.3 and 4.4, we use the generalized method of averaging and the method of
normal forms, respectively, to determine second-order uniform expansions for the case of subharmonic resonance of order
one-half; that is, W » 2 w.

à 4.1 Straightforward Expansion

We seek a straightforward expansion for the solutions of FDuffingEq in the form

uSol = u -> ISumAei ui @#D, 8i, 3<E &M;

Substituting uSol into FDuffingEq , letting F = e f and `m = e m, expanding the result for small e, and discarding terms of
order higher than e3, we have

eq41a = HFDuffingEq •. uSol •. 8F -> e f, m -> e m< •• ExpandAllL •. e n_•;n>3 -> 0

`

e w2 u1@tD + d e2 u1 @tD2 + a e3 u1 @tD3 + e2 w2 u2@tD + 2 d e3 u1 @tD u2@tD +

e3 w2 u3 @tD + 2 e2 m u¢1@tD + 2 e3 m u¢2@tD + e u²1@tD + e2 u²2 @tD + e3 u²3@tD == f e Cos@t WD

Equating coefficients of like powers of e on both sides yields

eqEpsa = CoefficientList@Subtract žž eq41a, eD == 0 •• Thread •• Rest

8-f Cos@t WD + w2 u1@tD + u²1@tD == 0, d u1 @tD2 + w2 u2 @tD + 2 m u¢1 @tD + u²2 @tD == 0,
a u1 @tD3 + 2 d u1@tD u2 @tD + w2 u3@tD + 2 m u¢2 @tD + u²3 @tD == 0<

The particular solution of the first-order equation, linear and undamped problem, can be expressed as
76 Chapter4.nb

sol1p = DSolve@eqEpsa@@1DD, u1@tD, tD@@1DD •. C@_D -> 0 •• TrigToExp •• Simplify

E-I t W H1 + E2 It W L f
:u1 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ >
2 H-w2 + W2 L

We note that sol1p consists of a term whose denominator is very small when W » w. Such a term is called a small-divisor
term. Moreover, u@tD tends to infinity as W ® w, and the excitation is referred to as a resonance excitation. Because the
small-divisor term appears in the first-order problem, we speak of a primary or main resonance. When the small-divisor
terms appear in the higher-order problems, we speak of secondary resonances. In the case of primary resonance, the
scaling F ® e f is not valid. Physically, as soon as the motion becomes large, the damping and nonlinearity are activated
to counter the effect of the resonance. Consequently, to obtain a uniform expansion in this case, we rescale F and `m, as
discussed in Section 4.2, so that the influence of the damping and nonlinearity balances the influence of the primary
resonance.

When W is away from w, we add the homogeneous solution to sol1p and obtain the general solution of the first-order
problem as

fRule = 9f -> 2 L IW2 - w2 M=;

sol1 = u1 -> FunctionAt, A EI w t + A E-I w t + sol1p@@1, 2DD •. fRule •• Expand •• EvaluateE

•

u1 ® Function@t, A EIt w - E-I t W L - EI t W L + E-I t w AD

•

where A is a complex-valued constant. Substituting sol1 into the second-order equation, eqEpsa[[2]], yields

order2Eqa = u2 ²@tD + w2 u2 @tD - # & •ž eqEpsa@@2DD •. sol1 •• ExpandAll •• Reverse

w2 u2@tD + u²2@tD ==
-A2 E2 I t w d + 2 A EI t w-It W d L + 2 A EI t w+I t W d L - 2 d L2 - E-2 I t W d L2 - E2I t W d L2 - 2 I A EIt w m w -
• • • • •2
2 I E-I t W L m W + 2 I EI t W L m W - 2 A d A + 2 E-I t w-I t W d L A + 2 E-I t w+I t W d L A + 2 I E-I t w m w A - E-2 I t w d A

Solving for the particular solution of order2Eqa, we obtain

sol2p = Hu2 @tD •. DSolve@order2Eqa@@1DD == #, u2 @tD, tD@@1DD •. C@_D -> 0 •• TrigToExp ••

SimplifyL & •ž order2Eqa@@2DD

A2 E2 I t w d 2 d L2 A EI t w m HI + 2 t wL 2 A EI t Hw-WL d L
H2 w - WL W
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
3 w2 w2 2w
2 A EI tHw+WL d L 2 I EI t W L m W 2 I E-I t W L m W E-2I t W d L2 E2 I t W d L2
W H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ -
w2 - W2 -w2 + W2 -w2 + 4 W2 -w2 + 4 W2

E-I t w m H-I + 2 t wL A 2 E-I t Hw-WL d L A 2 E-I t Hw+WL d L A

• • • • •2
E-2 I t w d A
H2 w - WL W W H2 w + WL
2AdA
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€
w2 2w 3 w2

breaks down when W » 2 w, W » w • 2, or W » 0 because they produce small-divisor terms and hence secondary resonances.
Clearly, sol2p breaks down because it contains secular terms proportional to the damping coefficient m. Moreover, sol2p

As discussed in Section 4.2, the case W » 2 w is called subharmonic resonance of order one-half and the case W » w •2 is
called superharmonic resonance of order two.
Chapter4.nb 77

When W is away from w, 2 w, w • 2, or 0, sol2p is valid only if we rescale the damping term so that it first appears in the third-
order equation; that is, `m -> e2 m. Using this scaling, we rewrite eq41a as

eq41b = eq41a •. m -> e m •. e4 -> 0

e w2 u1@tD + d e2 u1 @tD2 + a e3 u1 @tD3 + e2 w2 u2@tD + 2 d e3 u1 @tD u2@tD +

e3 w2 u3 @tD + 2 e3 m u¢1@tD + e u²1 @tD + e2 u²2 @tD + e3 u²3@tD == f e Cos@t WD

eqEpsb = CoefficientList@Subtract žž eq41b, eD == 0 •• Thread •• Rest

8-f Cos@t WD + w2 u1@tD + u²1@tD == 0, d u1 @tD2 + w2 u2 @tD + u²2 @tD == 0,

a u1 @tD3 + 2 d u1@tD u2 @tD + w2 u3@tD + 2 m u¢1 @tD + u²3 @tD == 0<

Comparing eqEpsb to eqEpsa, we immediately obtain its solution fromsol2p as

sol2pNew = sol2p •. m -> 0

A2 E2 I t w d 2 d L2 2 A EI t Hw-WL d L 2 A EI t Hw+WL d L E-2 I t W d L2

H2 w - WL W W H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ +
3 w2 w2 -w2 + 4 W2

2 E-It Hw-WL d L A 2 E-I t Hw+WL d L A

• • • •2
E-2 I t w d A
H2 w - WL W W H2 w + WL
E2 I tW d L2 2AdA
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€
-w2 + 4 W2 w2 3 w2

Hence, the general solution of eqEpsb[[2]] can be written as

sol2New = u2 -> FunctionAt, B EI w t + B E-I w t + sol2pNew •• EvaluateE

•

A2 E2I t w d 2 d L2 2 A EIt Hw-WL d L 2 A EI t Hw+WL d L E-2I t W d L2

H2 w - WL W W H2 w + WL
u2 ® FunctionBt, B EIt w + €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€ +
3 w2 w2 -w2 + 4 W2

2 E-It Hw-WL d L A 2 E-I t Hw+WL d L A

€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ + E-I t w BF
• • • • 2
E-2 I t w d A
H2 w - WL W W H2 w + WL
E2 I tW d L2 2AdA •
-w2 + 4 W2 w2 3 w2

where B is a complex-valued constant.

Substituting sol1 and sol2New into eqEpsb[[3]] yields

78 Chapter4.nb

order3Eqb = u3 ²@tD + w2 u3 @tD - # & •ž eqEpsb@@3DD •. sol1 •. sol2New •• ExpandAll •• Reverse

w2 u3@tD + u²3@tD == -A3 E3 I t w a - 2 A B E2 I t w d + 3 A2 E2 I t w-It W a L + 3 A2 E2 It w+I t W a L +

2 B EI t w-I tW d L + 2 B EI t w+I t W d L - 6 A EI t w a L2 - 3 A EIt w-2 I t W a L2 - 3 A EI t w+2 I t W a L2 +
2 A3 E3 I t w d2 2 A2 E2 I t w-I t W d2 L
3 E-I t W a L3 + 3 EI t W a L3 + E-3 I t W a L3 + E3 It W a L3 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ +
3 w2 3 w2
2 A2 E2 I t w+I tW d2 L 4 A EI t w d2 L2 4 E-I t W d2 L3 4 EI t W d2 L3
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 I A EI t w m w -
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 w2 w2 w2 w2
2 2 I t w-I tW 2 Itw 2 2
4A E d L 4AE d L 4 A EI t w-2 It W d2 L2
2 I E-I t W L m W + 2 I EI t W L m W - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ +
2 w W - W2 2 w W - W2 2 w W - W2
4 A2 E2 I t w+I tW d2 L 4 A EI t w d2 L2 4 A EI t w+2 It W d2 L2 2 A EIt w+2 I t W d2 L2 2 EI t W d2 L3
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 w W + W2 2 w W + W2 2 w W + W2 -w2 + 4 W2 -w2 + 4 W2
2E 3 It W d L
2 3 2AE I t w d L
2 2 2E -I tW d L
2 3 2E I tW d L2 3
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
-w2 + 4 W2 -E2 I t W w2 + 4 E2I t W W2 -E2 I tW w2 + 4 E2 I t W W2 -E2 I t W w2 + 4 E2 I t W W2
• • • • • •
3 A2 EIt w a A - 2 B d A + 6 A E-I t W a L A + 6 A EI t W a L A - 6 E-I t w a L2 A - 3 E-I t w-2 I t W a L2 A -
• • • •
• 10 A2 EI t w d2 A 4 A E-I t W d2 L A 4 A EI t W d2 L A 4 E-I tw d2 L2 A
3 E-I t w+2 It W a L2 A + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
3 w2 w2 w2 w2
-I t W 2 • ItW 2 • -It w+2 I t W 2 2 • •
• 4AE d LA 4AE d LA 2E d L A 2 E-I tw d2 L2 A
2 I E-I t w m w A - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ -
2 w W - W2 2 w W + W2 -w2 + 4 W2 -E2 I t W w2 + 4 E2 I t W W2
• • •
4 A EI tw d2 L A 4 E-I t W d2 L2 A 4 EI t W d2 L2 A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 EI tw-I t W w W - EI t w-I tW W2 2 EI t w-I t W w W - EIt w-I t W W2 2 EI t w-I t W w W - EI t w-I tW W2
• • •
4 A EI tw d2 L A 4 E-I t W d2 L2 A 4 EI t W d2 L2 A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2 EI tw+I t W w W + EI t w+I tW W2 2 EI t w+I t W w W + EIt w+I t W W2 2 EI t w+I t W w W + EI t w+I tW W2
•2 •2
•2 •2 •2 10 A E-I t w d2 A 2 E-2 I tw-I t W d2 L A
3 A E-I t w a A + 3 E-2 I t w-It W a L A + 3 E-2 It w+I t W a L A + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€ +
3 w2 3 w2
• 2 • 2 • 2
2 E-2 It w+I t W d2 L A 4 E-I t w d2 L A 4 E-I t w d2 L A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
3 w2 2 EI t w-I tW w W - EI t w-I t W W2 2 EIt w+I t W w W + EI t w+It W W2
• 3
•3 2 E-3 I tw d2 A • • • ••
E-3 I t w a A - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 A d B + 2 E-I t w-I t W d L B + 2 E-I t w+I t W d L B - 2 E-2I t w d A B
3 w2

The particular solution of order3Eqb is given by

Chapter4.nb 79

Hsol3p = Hu3 @tD •. DSolve@order3Eqb@@1DD == #, u3 @tD, tD@@1DD •. C@_D -> 0 •• TrigToExp ••

SimplifyL & •ž order3Eqb@@2DDL •• Timing

A EI tw d2 L2 H1 - 2 I t wL
:64.803 Second, €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
A3 E3 I t w d2 A3 E3 I t w a 2 A B E2 I tw d
€€€€€€€€€€€€€€€€€€€€€€€ +

3 I A EIt w a L2 HI + 2 t wL A EI t w m HI + 2 t wL
12 w4 8 w2 3 w2 w4
3 A EI t Hw-2 WL a L2 2 B EI t Hw-WL d L
4 Hw - WL W H2 w - WL W
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 w2 2w
A EI tw d2 L2 H1 - 2 I t wL 3 A EI t Hw+2 WL a L2 2 B EI t Hw+WL d L I A EI t w d2 L2 HI + 2 t wL
w2 H2 w - WL W 4 W Hw + WL W H2 w + WL w2 W H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ +

A EI t Hw+2 WL d2 L2 3 A2 EI t H2 w+WL a L 2 A2 EI t H2w+WL d2 L 4 A2 EI t H2 w+WL d2 L

W2 Hw + WL H2 w + WL Hw + WL H3 w + WL 3 w2 Hw + WL H3 w + WL W Hw + WL H2 w + WL H3 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

E-3 It W a L3 E3 I t W a L3 A EI t Hw-2 WL d2 L2 A EI t Hw+2 WL d2 L2 3 E-I t W a L3

2 Hw - WL W Hw - 4 W L 2 W Hw + WL Hw - 4 W L
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ +
w -9W
2 2 w -9W
2 2 2 2 2 2 w2 - W2
3 EI tW a L3 2 I EI t W L m W 4 E-I t W d2 L3 4 EI t W d2 L3 2 I E-I t W L m W 3 A2 EI t H2 w-WL a L
w2 H-w2 + W2 L w2 H-w2 + W2L
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w2 - W2 w2 - W2 -w2 + W2 3 w2 - 4 w W + W2
2 A2 EI t H2 w-WL d2 L A EI t Hw-2WL d2 L2 4 A2 EIt H2 w-WL d2 L
3 w H3 w - 4 w W + W L W H2 w - 3 w W + W L W H-6 w3 + 11 w2 W - 6 w W2 + W3L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2 2 2 2 2 2
•
2 E-It W d2 L3 2 EI t W d2 L3 2 E-3 I t W d2 L3 2 E3 I t W d2 L3 2BdA
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€ - €€€€€€€€€€€€€€€ +

d H1 - 2 I t wL A d L H1 + 2 I t wL A a L H-I + 2 t wL A
w4 - 5 w2 W2 + 4 W4 w4 - 5 w2 W2 + 4 W4 w4 - 13 w2 W2 + 36 W4 w4 - 13 w2 W2 + 36 W4 w2
2 I t w 2
• -I t w 2 2
• -I t w 2
•
5A E E 3IE
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

E-I tw m H-I + 2 t wL A 3 I A2 EI t w a HI + 2 t wL A E-I t w d2 L2 H1 + 2 I t wL A

6w 4 w 4 2 w2
• • •

w2 H2 w - WL W
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ +
2w 4 w2
• • •
3 E-I tHw+2 WL a L2 A 4 A E-I t W d2 L A 4 A EI t W d2 L A
4 W Hw + WL W H-2 w + WL H-w + WL Hw + WL W H-2 w + WL H-w + WL Hw + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

d L H-I + 2 t wL A -I t Hw+2 WL 2 2 •
-I tw 2 2 • •
4 A E-It W d2 L A
w2 W H2 w + WL W2 Hw + WL H2 w + WL W H-w + WL Hw + WL H2 w + WL
IE E d L A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

-I t Hw-2 WL 2 2 •
E-I t Hw+2 WL d2 L2 A
ItW 2 • •

W H-w + WL Hw + WL H2 w + WL 2 Hw - WL W Hw2 - 4 W2L 2 W Hw + WL Hw2 - 4 W2 L

4AE d LA E d L A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€ +

E-I t Hw-2 WL d2 L2 A
• • • • •
6 A E-It W a L A 6 A EI t W a L A 4 A E-I t W d2 L A 4 A EI t W d2 L A
w2 H-w2 + W2 L w2 H-w2 + W2 L W2 H2 w2 - 3 w W + W2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ +
w2 - W2 w2 - W2

5 A E-I t w d2 H1 + 2 I t wL A 3 I A E-I t w a H-I + 2 t wL A

• •2 •2
3 E-I tw+2 I t W a L2 A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
-4 w W + 4 W2 6 w4 4 w2
H2 H2
-I w+WL
•2 -I w+WL
•2 -It w+WL
•2

Hw + WL H3 w + WL 3 w2 Hw + WL H3 w + WL W Hw + WL H2 w + WL H3 w + WL
3E tH2 aL A 2E t d LA2 4E d LA
2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• 2 • 2 •2
3 E-2 It w+I t W a L A 2 E-2 I t w+I t W d2 L A 4 E-2 I t w+I tW d2 L A
3 w2 H3 w2 - 4 w W + W2 L W H-6 w3 + 11 w2 W - 6 w W2 + W3 L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
3 w2 - 4 w W + W2

2 E-I t Hw-WL d L B 2 E-I t Hw+WL d L B

€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
•3 •3 • • • ••
E-3 It w d2 A E-3I t w a A 2 E-2 I t w d A B
H2 w - WL W W H2 w + WL
2AdB
12 w4 8 w2 w2 3 w2

W » w • 2, W » 3 w, and W » w • 3. As shown in the next section, the small-divisor terms arising from W » 3 w produce a
Clearly, sol3p breaks down because it contains secular terms and small-divisor terms when W » w, W » 0, W » 2 w,
80 Chapter4.nb

subharmonic reosonance of order one-third and those arising from W » w • 3 produce a superharmonic resonance of order
three. Carrying out the expansion to higher order, one finds that other resonances may occur.

We note that the resonances that occur depend on the order of the nonlinearity. Quadratic nonlinearities produce (a)
subharmonic resonances of order one-half and superharmonic resonances of order two at second order and (b) subhar-
monic resonances of order one-third and superharmonic resonances of order three at third order. On the other hand, cubic
nonlinearities produce subharmonic resonances of order one-third and superharmonic resonances of order three at third
order. For a given system and order of approximation, the resonances produced can be easily identified by carrying out a
straightforward expansion as done above.

In the next three sections, we use the method of multiple scales, generalized method of averaging, and method of normal
forms to determine second-order uniform expansions for the solutions of FDuffingEq that do not contain secular or small-
divisor terms.

à 4.2 The Method of Multiple Scales

Ÿ 4.2.1 Preliminaries

For a uniform second-order expansion, we need three time scales, which we symbolize and list as follows:

Needs@"Utilities`Notation`"D

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

timeScales = 8T0 , T1, T2 <;

The maximum order of the expansion is related, in general, to the number of time scales by

maxOrder = Length@timeScalesD - 1;

In what follows, we need the complex conjugates of A, z, and G, which we define by

conjugateRule = 9Complex@0, n_D -> Complex@0, -nD, A -> A, A -> A, z -> z, z -> z, G -> G, G -> G=;
• • • • • •

In terms of the time scales Tn , the time derivatives become

dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, maxOrder<E;

dt@2D@expr_D := Hdt@1D@dt@1D@exprDD •• ExpandL •. ei_•;i>maxOrder -> 0;

To represent some of the expressions in a more concise way, we introduce the following display rule:
Chapter4.nb 81

displayRule =
9Derivative@a__DAz_i_E@__D :> SequenceFormATimes žž MapIndexedAD #1
#2@@1DD-1 &, 8a<E, zi E,

#2@@1DD &, 8a<E, AE,

Derivative@a__D@AD@__D :> SequenceFormATimes žž MapIndexedAD #1
Derivative@a__D@AD@__D :> SequenceFormATimes žž MapIndexedAD #1
#2@@1DD &, 8a<E, AE,
• •

z_i_ @__D -> zi , A@__D -> A, A@__D -> A=;

• •

Using the time scales T0, T1 , and T2 , we transform FDuffingEq from an ordinary-differential equation into a partial-differen-
tial equation according to the rule

multiScalesRule =
8u@tD -> u@T0 , T1, T2 D, Derivative@n_D@uD@tD :> dt@nD@u@T0 , T1 , T2DD, t -> T0 <;

We seek a second-order uniform expansion of the solution of FDuffingEq in the form

solRule = u -> ISumAei ui @#1, #2, #3D, 8i, maxOrder + 1<E &M;

For a uniform expansion, we need to eliminate the secular and small-divisor terms. To accomplish this, we need to distinguish
between primary and secondary resonances. They are treated separately beginning with the case of primary resonance.

Ÿ 4.2.2 Primary Resonance

We scale the damping `m and forcing F so that the damping term and resonance terms appear at the same order at which the
secular and / or small-divisor terms generated by the nonlinearities appear. In this case, they appear at order e3, and hence we
scale the forcing and damping as

scaling = 9F -> e3 f, m -> e2 m=;

`

Substituting the solRule into FDuffingEq, using the scaling and multiScalesRule, expanding the result for small e, and
discarding terms of order higher than e3, we obtain

eq422a =
HFDuffingEq •. scaling •. multiScalesRule •. solRule •• ExpandAllL •. e n_•;n>maxOrder+1 -> 0;
eq422a •. displayRule

2 e3 m HD0u1 L + e HD20 u1L + e2 HD20 u2L + e3 HD20 u3L + 2 e2 HD0 D1u1 L + 2 e3 HD0 D1 u2L + e3 HD21 u1 L +
2 e3 HD0 D2 u1L + e w2 u1 + d e2 u21 + a e3 u31 + e2 w2 u2 + 2 d e3 u1 u2 + e3 w2 u3 == f e3 Cos@T0 WD

Equating coefficients of like powers of e in eq422a, we obtain

eqEps = CoefficientList@Subtract žž eq422a, eD == 0 •• Thread •• Rest •• TrigToExp;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := HeqEps@@1, 1DD •. u1 -> ui L - # & •ž eqEps@@iDD •• Reverse

Using eqOrder[i] and the displayRule, we rewrite eqEps in a concise way as

82 Chapter4.nb

Array@eqOrder, 3D •. displayRule

9D20 u1 + w2 u1 == 0, D20 u2 + w2 u2 == -2 HD0 D1u1 L - d u21 ,

D20 u3 + w2 u3 == €€€€ HE-I T0 W + EI T0 WL f - 2 m HD0u1 L - 2 HD0 D1u2 L - D21 u1 - 2 HD0 D2 u1L - a u31 - 2 d u1 u2 =
1
2

The general solution of eqOrder[1] can be expressed as

sol1 = u1 -> FunctionA8T0 , T1 , T2<, A@T1 , T2 D Exp@I w T0 D + A@T1 , T2 D Exp@-I w T0 DE;

•

Using sol1, we rewrite eqOrder[2] as

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

D20 u2 + w2 u2 == -A2 E2 I T0 w d - 2 A d A - E-2 IT0 w d A - 2 I EI T0 w w HD1 AL + 2 I E-I T0 w w HD1AL

• •2 •

Eliminating the terms that produce secular terms in order2Eq demands that

SCond1 = Coefficient@order2Eq@@2DD, Exp@I w T0 DD == 0

-2 I w AH1,0L @T1, T2D == 0

Solving SCond1 for AH1,0L@T1 , T2 D yields

SCond1Rule = SolveASCond1, AH1,0L @T1 , T2DE@@1DD

8AH1,0L@T1 , T2 D ® 0<

Hence, A = A@T2 D.

Substituting the solvability condition into order2Eq, we have

order2Eqm = order2Eq •. SCond1Rule •. HSCond1Rule •. conjugateRuleL;

order2Eqm •. displayRule
• •2
D20 u2 + w2 u2 == -A2 E2 I T0 w d - 2 A d A - E-2 IT0 w d A

The particular solution of order2Eqm can be obtained by using DSolve; the result is

sol2p = DSolve@order2Eqm, u2 žž timeScales, timeScalesD@@1DD •. C@_D -> H0 &L •• TrigToExp ••

ExpandAll;
sol2p •. displayRule

:u2 ® €€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€ >

• •2
A2 E2 I T0 w d 2AdA E-2 I T0 w d A
3 w2 w2 3 w2

Next, we use sol2p[[1,2]] directly to express u 2 in a pure function form as

sol2 = u2 -> Function@8T0 , T1 , T2<, sol2p@@1, 2DD •• EvaluateD;

Chapter4.nb 83

Substituting the first- and second-order solutions, sol1 and sol2, into the third-order equation, eqOrder[3], we obtain

order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

order3Eq •. displayRule

1 1 2 A3 E3 I T0 w d2
D20 u3 + w2 u3 == €€€€ E-I T0 W f + €€€€ EIT0 W f - A3 E3 I T0 w a - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2 2 3 w2
•
• 10 A2 EI T0 w d2 A • •2
2 I A EI T0 w m w - 3 A2 EI T0 w a A + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 I E-I T0 w m w A - 3 A E-I T0 w a A +
3 w2
8 I A E2 I T0 w d HD1AL
•2 •3
10 A E-IT0 w d2 A •3 2 E-3 I T0 w d2 A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - E-3I T0 w a A - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ +

8 I E-2I T0 w d A HD1 AL
3 w2 3 w2 3w

€€€€€€€€€€€€€€€ - EI T0 w HD21AL - E-IT0 w HD21 AL - 2 I EI T0 w w HD2 AL + 2 I E-I T0 w w HD2AL

• •
• •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3w

To express the nearness of W to w, we introduce a detuning parameter s = OH1L defined by

OmgRule = 9W -> w + e2 s=;

Using the OmgRule, we convert the small-divisor term arising from the excitation into a secular term. Then, eliminating the
secular terms from u3 demands that

expRule1 = Exp@a_D :> ExpAExpand@a •. OmgRuleD •. e2 T0 -> T2 E;

SCond2 = Coefficient@order3Eq@@2DD •. expRule1, Exp@I w T0 DD == 0;

SCond2 •. displayRule

€€€€ EI T2 s f - 2 I A m w - 3 A2 a A + €€€€€€€€€€€€€€€€€€€€€€ - D21A - 2 I w HD2 AL == 0

•
1 • 10 A2 d2 A
2 3 w2

Using the fact that A = A@T2 D from SCond1Rule, we rewrite SCond2 as

SCond = SCond2 •. D@SCond1Rule, T1D •. 9A -> HA@#2D &L, A -> IA@#2D &M=
• •

10 d2 A@T2 D2 A@T2 D
€€€€ EI T2 s f - 2 I m w A@T2 D - 3 a A@T2 D2 A@T2 D + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - 2 I w A¢@T2 D == 0
•
1 •
2 3 w2

Ÿ 4.2.3 Secondary Resonances Due to Cubic Nonlinearities

In this case W is away from w, 2 w, and w • 2 and small-divisor terms first appear at OHe3 L. We scale the damping m` and forcing
F so that the damping term and resonance terms appear at the same order as the cubic nonlinearity according to

scaling = 9F -> e f, m -> e2 m=;

`
84 Chapter4.nb

eq423a =
HFDuffingEq •. scaling •. multiScalesRule •. solRule •• ExpandAllL •. e n_•;n>maxOrder+1 -> 0;
eq423a •. displayRule

2 e3 m HD0u1 L + e HD20 u1L + e2 HD20 u2L + e3 HD20 u3L + 2 e2 HD0 D1u1 L + 2 e3 HD0 D1 u2L + e3 HD21 u1 L +
2 e3 HD0 D2 u1L + e w2 u1 + d e2 u21 + a e3 u31 + e2 w2 u2 + 2 d e3 u1 u2 + e3 w2 u3 == f e Cos@T0 WD

Equating coefficients of like powers of e, we obtain

eqEps = CoefficientList@Subtract žž eq423a, eD == 0 •• Thread •• Rest •• TrigToExp;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := HeqEps@@1, 1DD •. f -> 0 •. u1 -> ui L - # & •ž eqEps@@iDD •• Reverse

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

Array@eqOrder, 3D •. displayRule •• TableForm

D20 u1 + w2 u1 == €€€12 HE-I T0 W + EI T0 WL f

D20 u2 + w2 u2 == -2 HD0 D1 u1L - d u21
D20 u3 + w2 u3 == -2 m HD0 u1 L - 2 HD0 D1 u2 L - D21 u1 - 2 HD0 D2 u1L - a u31 - 2 d u1 u2

The particular solution of eqOrder[1] can be expressed as

sol1p = DSolve@eqOrder@1D, u1 žž timeScales, timeScalesD@@1DD •. C@_D -> H0 &L •• Simplify

E-I T0 W H1 + E2 I T0 WL f
:u1 @T0, T1 , T2 D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ >
2 H-w2 + W2 L

The general solution of eqOrder[1] can be written as

fRule = 9f ® -2 L I-w2 + W2 M=;

sol1 = u1 -> FunctionA8T0 , T1 , T2<,

A@T1 , T2D Exp@I w T0 D + A@T1 , T2D Exp@-I w T0D + sol1p@@1, 2DD •. fRule •• Expand •• EvaluateE
•

u1 ® Function@8T0 , T1 , T2<, E-I T0 W L + EI T0 W L + EI T0 w A@T1 , T2D + E-I T0 w A@T1, T2 DD

•

Substituting sol1 into eqOrder[2] yields

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

D20 u2 + w2 u2 == -A2 E2 I T0 w d - 2 A EI T0 w-I T0 W d L - 2 A EI T0 w+I T0 W d L - 2 d L2 - E-2 I T0 W d L2 - E2 I T0 W d L2 -

2 A d A - 2 E-I T0 w-IT0 W d L A - 2 E-I T0 w+I T0 W d L A - E-2 I T0 w d A - 2 I EI T0 w w HD1 AL + 2 I E-I T0 w w HD1 AL
• • • •2 •
Chapter4.nb 85

Ÿ The Case W » 3 w

To express the nearness of W to 3 w, we introduce a detuning parameter s = OH1L defined by

OmgRule = 9W -> 3 w + e2 s=;

expRule1 = 9Exp@arg_D :> ExpAExpand@arg •. OmgRuleD •. e2 T0 -> T2 E=;

Eliminating the secular terms from u 2 demands that

SCond1 = Coefficient@order2Eq@@2DD •. expRule1, Exp@I w T0 DD == 0

-2 I w AH1,0L @T1, T2D == 0

SCond1Rule = SolveASCond1, AH1,0L @T1 , T2DE@@1DD

8AH1,0L@T1 , T2 D ® 0<

whose complex conjugate is

ccSCond1Rule = SCond1Rule •. conjugateRule

•H1,0L
:A @T1 , T2 D ® 0>

Substituting the solvability condition, SCond1Rule, and its complex conjugate, into order2Eq, we have

Horder2Eqm = order2Eq •. SCond1Rule •. ccSCond1RuleL •. displayRule

D20 u2 + w2 u2 == -A2 E2 I T0 w d - 2 A EI T0 w-I T0 W d L - 2 A EI T0 w+I T0 W d L - 2 d L2 -

• • • •2
E-2 I T0 W d L2 - E2 I T0 W d L2 - 2 A d A - 2 E-I T0 w-I T0 W d L A - 2 E-I T0 w+I T0 W d L A - E-2I T0 w d A

In order to efficiently use DSolve to determine the particular solution of order2Eqm, we first transform the partial-differential
equation to an ordinary-differential equation. Then we solve for the particular solution of the resulting equation and obtain

sol2p =
Hu2 @T0D •. DSolve@Horder2Eqm@@1DD •. u2 -> Hu2 @#1D &LL == #, u2 @T0D, T0 D@@1DD •. C@_D -> 0 ••
TrigToExp •• ExpandL & •ž order2Eqm@@2DD;
sol2p •. displayRule

A2 E2 I T0 w d 2 d L2 2 A EI T0 Hw-WL d L 2 A EI T0 Hw+WL d L E-2 IT0 W d L2

H2 w - WL W W H2 w + WL H-w + 2 WL Hw + 2 WL
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ +
3 w2 w2

2 E-I T0 Hw-WL d L A 2 E-I T0 Hw+WL d L A

• • • •2
E2 I T0 W d L2 E-2 I T0 w d A
H-w + 2 WL Hw + 2 WL H2 w - WL W W H2 w + WL
2AdA
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ - €€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€
w2 3 w2

which can be used directly to express u 2 in a pure function form as

sol2 = u2 -> Function@8T0 , T1 , T2<, sol2p •• EvaluateD;

Substituting sol1 and sol2 into the third-order equation, eqOrder[3], we obtain
86 Chapter4.nb

order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

Eliminating the secular terms from u3 demands that

SCond2 =
Coefficient@order3Eq@@2DD •. expRule1 •. W -> 3 w, Exp@I w T0 DD == 0 •. D@SCond1Rule, T1 D;
SCond2 •. displayRule

-6 A a L2 + €€€€€€€€€€€€€€€€€€€€€€ - 2 I A m w - 3 A2 a A + €€€€€€€€€€€€€€€€€€€€€€ - 3 EI T2 s a L A - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 I w HD2 AL == 0

• •2
12 A d2 L2 • 10 A2 d2 A •2 2 EI T2 s d2 L A
5 w2 3 w2 w2

1
Ÿ The Case W » €€3€€€ w

To express the nearness of W to €€13€€ w, we introduce a detuning parameter s = OH1L defined by

OmgRule = 9W -> €€€€ Iw + e2 sM=;

1
3

expRule1 = 9Exp@arg_D :> ExpAExpand@arg •. OmgRuleD •. e2 T0 -> T2 E=;

Eliminating the secular terms from u2 demands that

SCond1 = Coefficient@order2Eq@@2DD •. expRule1, Exp@I w T0 DD == 0

-2 I w AH1,0L @T1, T2D == 0

or

SCond1Rule = SolveASCond1, AH1,0L @T1 , T2DE@@1DD

8AH1,0L@T1 , T2 D ® 0<

whose complex conjugate is

ccSCond1Rule = SCond1Rule •. conjugateRule;

Substituting the solvability condition, SCond1Rule, and its complex conjugate, into order2Eq, we have

order2Eqm = order2Eq •. SCond1Rule •. ccSCond1Rule;

order2Eqm •. displayRule

D20 u2 + w2 u2 == -A2 E2 I T0 w d - 2 A EI T0 w-I T0 W d L - 2 A EI T0 w+I T0 W d L - 2 d L2 -

• • • •2
E-2 I T0 W d L2 - E2 I T0 W d L2 - 2 A d A - 2 E-I T0 w-I T0 W d L A - 2 E-I T0 w+I T0 W d L A - E-2I T0 w d A

Again, we first transform the partial-differential equation into an ordinary-differential equation and then use DSolve to obtain
the particular solution of order2Eqm as
Chapter4.nb 87

Hsol2p =
Hu2 @T0D •. DSolve@Horder2Eqm@@1DD •. u2 -> Hu2 @#1D &LL == #, u2 @T0D, T0 D@@1DD •. C@_D -> 0 ••
TrigToExpL & •ž order2Eqm@@2DDL •. displayRule

A2 E2 I T0 w d 2 d L2 2 A EI T0 Hw-WL d L 2 A EI T0 Hw+WL d L E-2 IT0 W d L2

H2 w - WL W W H2 w + WL H-w + 2 WL Hw + 2 WL
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ +
3 w2 w2

2 E-I T0 Hw-WL d L A 2 E-I T0 Hw+WL d L A

• • • •2
E2 I T0 W d L2 E-2 I T0 w d A
H-w + 2 WL Hw + 2 WL H2 w - WL W W H2 w + WL
2AdA
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ - €€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€
w2 3 w2

sol2 = u2 -> Function@8T0 , T1 , T2<, sol2p •• EvaluateD;

Substituting sol1 and sol2 into the third-order equation, eqOrder[3], we obtain

order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

Converting the terms that produce small-divisor terms into terms that produce secular terms and then eliminating the secular
terms from u3 demands that

SCond2 =

CoefficientAorder3Eq@@2DD •. expRule1 •. W -> €€€€ w, Exp@I w T0 DE == 0 •. D@SCond1Rule, T1D;

1

SCond2 •. displayRule
3

-6 A a L2 - EI T2 s a L3 + €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 I A m w - 3 A2 a A + €€€€€€€€€€€€€€€€€€€€€€ - 2 I w HD2 AL == 0

•
212 A d2 L2 18 EI T2 s d2 L3 • 10 A2 d2 A
35 w2 5 w2 3 w2

Ÿ 4.2.4 Secondary Resonances Due to Quadratic Nonlinearities

In this case W is away from w and small-divisor terms first appear at OHe2 L. We scale the damping `m and forcing F so that the
damping term and resonance terms appear at the same order as the quadratic nonlinearity; that is,

scaling = 8F -> e f, m -> e m<;

`

eq424a =
HFDuffingEq •. scaling •. multiScalesRule •. solRule •• ExpandAllL •. e n_•;n>maxOrder+1 -> 0;
eq424a •. displayRule

2 e2 m HD0u1 L + 2 e3 m HD0 u2 L + e HD20 u1L + e2 HD20 u2L + e3 HD20 u3L +

2 e3 m HD1 u1 L + 2 e2 HD0 D1 u1L + 2 e3 HD0 D1u2 L + e3 HD21u1 L + 2 e3 HD0 D2 u1L +
e w2 u1 + d e2 u21 + a e3 u31 + e2 w2 u2 + 2 d e3 u1 u2 + e3 w2 u3 == f e Cos@T0 WD

Equating coefficients of like powers of e, we obtain

eqEps = CoefficientList@Subtract žž eq424a, eD == 0 •• Thread •• Rest •• TrigToExp;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := HeqEps@@1, 1DD •. f -> 0 •. u1 -> ui L - # & •ž eqEps@@iDD •• Reverse

88 Chapter4.nb

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

Array@eqOrder, 3D •. displayRule •• TableForm

D20 u1 + w2 u1 == €€€12 HE-I T0 W + EI T0 WL f

D20 u2 + w2 u2 == -2 m HD0 u1 L - 2 HD0 D1 u1 L - d u21
D20 u3 + w2 u3 == -2 m HD0 u2 L - 2 m HD1u1 L - 2 HD0 D1 u2 L - D21u1 - 2 HD0 D2u1 L - a u31 - 2 d u1 u2

The particular solution of eqOrder[1] can be expressed as

sol1p = DSolve@eqOrder@1D, u1 žž timeScales, timeScalesD@@1DD •. C@_D -> H0 &L •• Simplify

E-I T0 W H1 + E2 I T0 WL f
:u1 @T0, T1 , T2 D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ >
2 H-w2 + W2 L

fRule = f -> -2 I-w2 + W2 M L;

The general solution of eqOrder[1] can be written in pur function form as

sol1 = u1 -> FunctionA8T0 , T1 , T2<,

A@T1 , T2D Exp@I w T0 D + A@T1 , T2D Exp@-I w T0D + sol1p@@1, 2DD •. fRule •• Expand •• EvaluateE
•

u1 ® Function@8T0 , T1 , T2<, E-I T0 W L + EI T0 W L + EI T0 w A@T1 , T2D + E-I T0 w A@T1, T2 DD

•

Substituting sol1 into eqOrder[2] yields

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

D20 u2 + w2 u2 == -A2 E2 I T0 w d - 2 A EI T0 w-I T0 W d L - 2 A EI T0 w+I T0 W d L - 2 d L2 - E-2 I T0 W d L2 -

• •
E2 I T0 W d L2 - 2 I A EI T0 w m w + 2 I E-I T0 W L m W - 2 I EI T0 W L m W - 2 A d A - 2 E-I T0 w-IT0 W d L A -
2 E-I T0 w+I T0 W d L A + 2 I E-I T0 w m w A - E-2 I T0 w d A - 2 I EI T0 w w HD1AL + 2 I E-I T0 w w HD1 AL
• • •2 •

Ÿ The Case W » 2 w

To express the nearness of W to 2 w, we introduce a detuning parameter s = OH1L defined by

OmgRule = 8W -> 2 w + e s<;

sigRule = Solve@OmgRule •. Rule -> Equal, sD@@1DD

9s ® - €€€€€€€€€€€€€€€ =
2w-W
e

expRule2 = 8Exp@arg_D :> Exp@Expand@arg •. OmgRuleD •. e T0 -> T1D<;

Converting the terms that produce small-divisor terms into terms that produce secular terms and then eliminating the secular
terms from u2 demands that
Chapter4.nb 89

SCond1 = Coefficient@order2Eq@@2DD •. expRule2, Exp@I w T0 DD == 0;

SCond1 •. displayRule

-2 I A m w - 2 EI T1 s d L A - 2 I w HD1 AL == 0
•

or

SCond1Rule = ISolveASCond1, AH1,0L @T1 , T2DE@@1DD •• ExpandAllM;

SCond1Rule •. displayRule

:D1 A ® -A m + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

•
I EI T1 s d L A
w

Substituting sigRule into SCond1Rule yields

SCond1Rulem = SCond1Rule •. Exp@a_D :> Exp@a •. sigRule •. T1 -> e T0 •• ExpandD;

SCond1Rulem •. displayRule

:D1 A ® -A m + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ >
•
I E-2 I T0 w+I T0 W d L A
w

whose complex conjugate is

ccSCond1Rulem = SCond1Rulem •. conjugateRule;

Substituting these conditions into order2Eq yields

order2Eqm = order2Eq •. SCond1Rulem •. ccSCond1Rulem •• ExpandAll;

order2Eqm •. displayRule

D20 u2 + w2 u2 == -A2 E2 I T0 w d - 2 A EI T0 w+I T0 W d L - 2 d L2 - E-2 IT0 W d L2 -

• • •2
E2 I T0 W d L2 + 2 I E-I T0 W L m W - 2 I EI T0 W L m W - 2 A d A - 2 E-I T0 w-I T0 W d L A - E-2 IT0 w d A

Transforming order2Eqm into an ordinary-differential equation, we can obtain the particular solution as

sol2p =
Hu2 @T0D •. DSolve@Horder2Eqm@@1DD •. u2 -> Hu2 @#1D &LL == #, u2 @T0D, T0 D@@1DD •. C@_D -> 0 ••
TrigToExpL & •ž order2Eqm@@2DD;
sol2p •. displayRule

A2 E2 I T0 w d 2 d L2 2 I E-I T0 W L m W 2 I EI T0 W L m W 2 A EI T0 Hw+WL d L
H-w + WL Hw + WL H-w + WL Hw + WL W H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
3w 2 w 2

2 E-I T0 Hw+WL d L A
• • •2
E-2 I T0 W d L2 E2 I T0 W d L2 E-2I T0 w d A
H-w + 2 WL Hw + 2 WL H-w + 2 WL Hw + 2 WL W H2 w + WL
2AdA
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€
w2 3 w2

sol2 = u2 -> Function@8T0 , T1 , T2<, sol2p •• EvaluateD;

Then, using sol1 and sol2 in eqOrder[3], we have

order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

90 Chapter4.nb

Converting the terms that produce small-divisor terms into terms that produce secular terms and then eliminating the secular
terms from u3 demands that

SCond2 =
Coefficient@order3Eq@@2DD •. expRule2 •. W -> 2 w, Exp@I w T0 DD == 0 •. D@SCond1Rule, T1 D •.
SCond1Rule •. HSCond1Rule •. conjugateRuleL •• ExpandAll;
SCond2 •. displayRule

-6 A a L2 + A m2 + €€€€€€€€€€€€€€€€€€€€ - 3 A2 a A + €€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 I w HD2 AL == 0

• • •
5 A d2 L2 • 10 A2 d2 A 8 I EI T1 s d L m A EI T1 s d L s A
2 w2 3 w2 3w w

or

SCond2Rule = SolveASCond2, AH0,1L @T1 , T2DE@@1DD •• ExpandAll;

SCond2Rule •. displayRule

:D2 A ® - €€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ >

• • • •
5 I A d2 L2 3 I A a L2 I A m2 5 I A2 d2 A 4 EI T1 s d L m A I EI T1 s d L s A 3 I A2 a A
4 w3 w 2w 3 w3 3 w2 2 w2 2w

The two partial-differential equations, SCond1Rule and SCond2Rule, can be reconstituted to obtain an ordinary-differential
equation governing A. The result is

H2 I w A¢ == H2 I w dt@1D@A@T1 , T2DD •. SCond1Rule •. SCond2Rule •• Expand •• Collect@#, eD &LL •.

displayRule

2 I w A¢ == e H-2 I A m w - 2 EI T1 s d L AL +
•

i EI T1 s d L s A y
e j-6 A a L + A m + €€€€€€€€€€€€€€€€€€€€ - 3 A a A + €€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
j
j z
• • •
z
5 A d2 L2 2 • 10 A2 d2 A 8 I EI T1 s d L m A

k {
2 2 2
2 w2 3 w2 3w w

Ÿ 4.2.5 First-Order Real-Valued System

As in Section 4.2.4 , we determine a second-order uniform expansion of the solution of the FDuffingEq for the case of subhar-
monic resonance of order one-half. However, instead of treating the second-order form of this equation, we treat its corre -
sponding first-order form; that is, we first transform it into a system of two real-valued first-order equations. To this end, we
introduce the transformation

vRule = u¢ @tD -> v@tD;

In order that the influence of the nonlinearity and damping balance the subharmonic resonance, we scale the damping
coefficient `m and forcing amplitude F as

scaling = 8F -> e f, m -> e m<;

`

Using the scaling and vRule, we transform the FDuffingEq into the following system of two real-valued first-order equations:

eq425a = 8vRule •. Rule -> Equal, FDuffingEq •. scaling •. vRule •. D@vRule, tD<

8u¢ @tD == v@tD, w2 u@tD + d u@tD2 + a u@tD3 + 2 e m v@tD + v¢@tD == f e Cos@t WD<
Chapter4.nb 91

To determine a second-order uniform expansion of the solution of eq425a using the method of multiple scales, we first
transform these equations from ordinary-differential equations into partial-differential equations in terms of the three time
scales T0 , T1 , and T2 according to

multiScales =
8u@tD -> u@T0 , T1, T2 D, v@tD -> v@T0, T1, T2 D, u_¢ @tD -> dt@1D@u@T0 , T1 , T2DD, t -> T0 <;

Next, we expand u and v in the form

solRule = 9u -> ISumAej uj @#1, #2, #3D, 8j, 3<E &M, v -> ISumAej vj @#1, #2, #3D, 8j, 3<E &M=;

Substituting multiScales and solRule into eq425a, expanding the result for small e, and discarding terms of order higher than
e3 , we obtain

eq425b = Heq425a •. multiScales •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e in eq425b yields

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq425b •• Transpose •• Rest ••

TrigToExp;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

linearSys = #@@1DD & •ž eqEps@@1DD •. f -> 0;

linearSys •. displayRule

8D0 u1 - v1, D0v1 + w2 u1<

eqOrder@i_D :=
HlinearSys •. u_1 -> ui L == HlinearSys •. u_1 -> ui L - H#@@1DD & •ž eqEps@@iDDL •• Thread

Using eqOrder[i] and the displayRule, we rewrite eqEps in a concise way as

eqOrder@1D •. displayRule •• TableForm

eqOrder@2D •. displayRule •• TableForm
eqOrder@3D •. displayRule •• TableForm

D0 u1 - v1 == 0
D0 v1 + w2 u1 == €€€12 HE-I T0 W + EI T0 WL f

D0 u2 - v2 == -HD1 u1L
D0 v2 + w2 u2 == -HD1v1 L - d u21 - 2 m v1

D0 u3 - v3 == -HD1 u2L - D2u1

D0 v3 + w2 u3 == -HD1v2 L - D2 v1 - a u31 - 2 d u1 u2 - 2 m v2

Ÿ First-Order Equations: Linear System

We seek a solution of the linearSys in the form u 1 = P EI w T 0 and v1 = Q EI w T 0 and obtain

92 Chapter4.nb

coefList = E-I w T0 linearSys •. 9u1 -> IP EI w # &M, v1 -> IQ EI w # &M= •• Expand

8-Q + I P w, I Q w + P w2<

The coefficient matrix of coefList is

coefMat = Outer@Coefficient, coefList, 8P, Q<D

88I w, -1<, 8w2 , I w<<

whose adjoint is defined by

hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose

Hence, the right and left eigenvectors of coefMat are

rightVec = 81, c1< •. [email protected], c1<L@@1DD == 0, c1D@@1DD

81, I w<

leftVec = 8c1, 1< •. Solve@[email protected], 1<L@@1DD == 0, c1D@@1DD

8-I w, 1<

The complex conjugate of leftVec is

ccleftVec = leftVec •. conjugateRule

8I w, 1<

Thefore, the homogeneous solution of eqOrder[1] is given by

sol1h = # + H# •. conjugateRuleL & •ž IrightVec A@T1, T2 D EI w T0 M

8EI T0 w A@T1 , T2D + E-I T0 w A@T1, T2 D, I EI T0 w w A@T1, T2 D - I E-I T0 w w A@T1 , T2D<

• •

Transforming eqOrder[1] into ordinary-differential equations and using DSolve, we obtain the particular solution of
eqOrder[1] as

sol1p = DSolve@eqOrder@1D •. 8u1 -> Hu1 @#1D &L, v1 -> Hv1@#1D &L<, 8u1@T0 D, v1@T0 D<, T0 D@@1DD •.
C@_D -> 0 •• TrigToExp •• Simplify

E-I T0 W H1 + E2 I T0 W L f I E-IT0 W H-1 + E2 IT0 WL f W

:u1 @T0D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ , v1 @T0D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ >
2 H-w2 + W2L 2 H-w2 + W2L

Then, the general solution of the first-order equations can be expressed in pure function form as

fRule = f -> -2 I-w2 + W2 M L;

Chapter4.nb 93

sol1 =
8u1 -> Function@8T0 , T1 , T2<, sol1h@@1DD + Hu1@T0 D •. sol1p •. fRule •• ExpandL •• EvaluateD,
v1 -> Function@8T0 , T1 , T2<, sol1h@@2DD + Hv1@T0 D •. sol1p •. fRule •• ExpandL •• EvaluateD<

8u1 ® Function@8T0 , T1 , T2<, E-I T0 W L + EI T0 W L + EI T0 w A@T1 , T2D + E-I T0 w A@T1, T2 DD,

•

v1 ® Function@8T0 , T1 , T2<, -I E-I T0 W L W + I EI T0 W L W + I EI T0 w w A@T1 , T2 D - I E-I T0 w w A@T1, T2 DD<

•

Ÿ Second-Order Equations

Substituting the first-order solution, sol1, into the second-order equations, eqOrder[2], we have

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

:D0 u2 - v2 == -EI T0 w HD1 AL - E-I T0 w HD1AL,

•

D0 v2 + w2 u2 == -A2 E2 I T0 w d - 2 A EI T0 w-I T0 W d L - 2 A EI T0 w+I T0 W d L - 2 d L2 - E-2 I T0 W d L2 -

• •
E2 I T0 W d L2 - 2 I A EI T0 w m w + 2 I E-I T0 W L m W - 2 I EI T0 W L m W - 2 A d A - 2 E-I T0 w-IT0 W d L A -
2 E-I T0 w+I T0 W d L A + 2 I E-I T0 w m w A - E-2 I T0 w d A - I EI T0 w w HD1AL + I E-I T0 w w HD1AL>
• • •2 •

To express quantitatively the nearness of the subharmonic resonance of order one-half, we introduce the detuning parameter
s defined by

OmgRule = 8W -> 2 w + e s<;

To convert the terms that produce small-divisor terms into terms that produce secular terms, we use the rule

expRule1 = Exp@arg_D :> Exp@Expand@arg •. OmgRuleD •. e T0 -> T1D;

To eliminate the terms that produce secular terms in u 2 and v2 (i.e., the solvability condition of order2Eq), we first determine
the vector proportional to EI w T 0 in the right-hand sides of order2Eq and obtain

STerms1 = Coefficient@#@@2DD •. expRule1, Exp@I w T0 DD & •ž order2Eq

8-AH1,0L@T1 , T2D, -2 I m w A@T1, T2D - 2 EIT1 s d L A@T1 , T2 D - I w AH1,0L@T1 , T2D<

•

Then, the solvability condition of order2Eq demands that STerms1 be orthogonal to the solution of the adjoint; that is,
ccleftVec. The result is

SCond1 = SolveAccleftVec.STerms1 == 0, AH1,0L @T1, T2 DE@@1DD

m w A@T1, T2 D - I EI T1 s d L A@T1 , T2D

:AH1,0L@T1 , T2 D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€ >
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w

whose complex conjugate is

ccSCond1 = SCond1 •. conjugateRule

•H1,0L I E-I T1 s d L A@T1 , T2 D + m w A@T1 , T2D

:A @T1 , T2 D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€ >
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w
94 Chapter4.nb

Next, we use the solvability condition to eliminate D1 A and its complex conjugate from order2Eq and then find their particu-
lar solution. To simplify the resulting equations using Mathematica , we need to replace T1 with e T0 . To this end, we express
the detuning parameter s in terms of w and W as

sigRule = Solve@OmgRule •. Rule -> Equal, sD@@1DD

9s ® - €€€€€€€€€€€€€€€ =
2w-W
e

and define the rule

expRule2 = Exp@a_D :> Exp@a •. T1 -> e T0 •. sigRule •• ExpandD;

This rule enables us to rewrite SCond1 and ccSCond1 as

SCond1f = Join@SCond1, ccSCond1D •. expRule2;

SCond1f •. displayRule

:D1 A ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ >

• •
A m w - I E-2 I T0 w+I T0 W d L A • I A E2I T0 w-I T0 W d L + m w A
€€€€€€€€€€€€€€€€€€€€€€€€€€ , D1A ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w w

With this form of the solvability condition, order2Eq becomes

order2Eqm = order2Eq •. SCond1f •• ExpandAll;

order2Eqm •. displayRule

:D0 u2 - v2 == A EI T0 w m + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + E-I T0 w m A - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

•
I A EI T0 w-I T0 W d L • I E-I T0 w+I T0 W d L A
€€€€€€€€€ ,
w w
D0 v2 + w2 u2 == -A2 E2 I T0 w d - A EI T0 w-IT0 W d L - 2 A EI T0 w+I T0 W d L - 2 d L2 -
E-2 I T0 W d L2 - E2 I T0 W d L2 - I A EI T0 w m w + 2 I E-IT0 W L m W - 2 I EI T0 W L m W -
2 A d A - 2 E-I T0 w-I T0 W d L A - E-I T0 w+I T0 W d L A + I E-I T0 w m w A - E-2 I T0 w d A >
• • • • •2

Next, we use the method of undetermined coefficients to determine the particular solution of order2Eqm. To accomplish this,
we first identify the form of their nonhomogeneous terms, which we will refer to as possible terms, and then seek the solution
as a linear combination of them. Substituting this solution into order2Eqm and equating the coefficient of each possible term
to zero, we obtain a set of pairs of algebraic equations that need to be solved for the unknown coefficients. The solutions of
these pairs are unique except for the coefficients of EI w T 0 and E-I w T 0 . To determine a unique solution corresponding to
these terms, we use two different approaches.

Approach 1:

In the first approach, we replace w with another symbol, say w0 , find the solution of the system of algebraic equations, and
then take the limit as w0 tends to w. To accomplish this, we define

collectForm = 9EI T0 w A@T1, T2 D, E-I T0 w A@T1, T2D, EI T0 W L, E-I T0 W L=;

•

Then, the form of the possible terms on the right-hand sides of order2Eqm can be determined as
Chapter4.nb 95

possibleTerms =
HcollectForm •. w -> w0L ~Join~ HOuter@Times, collectForm, collectFormD •• Flatten •• UnionL

:EI T0 w0 A@T1, T2 D, E-I T0 w0 A@T1 , T2 D, EI T0 W L, E-I T0 W L, L2, E-2 I T0 W L2 ,

•

E2 I T0 W L2 , EI T0 w-I T0 W L A@T1, T2 D, EI T0 w+I T0 W L A@T1 , T2D, E2 I T0 w A@T1 , T2D2 ,

E-I T0 w-I T0 W L A@T1 , T2D, E-I T0 w+IT0 W L A@T1 , T2 D, A@T1, T2 D A@T1, T2D, E-2 I T0 w A@T1, T2D >
• • • • 2

Next, we associate with each possible term an undetermined coefficient in u 2 by

symbolList1 = Table@Gi, 8i, Length@possibleTermsD<D

8G1 , G2, G3, G4 , G5 , G6, G7 , G8 , G9, G10, G11 , G12, G13 , G14 <

and an undetermined coefficient in v2 by

symbolList2 = symbolList1 •. G -> L

8L1 , L2, L3, L4 , L5 , L6, L7 , L8 , L9, L10, L11 , L12, L13 , L14 <

Then, we seek a particular solution of order2Eqm in the form

sol2Rule = 8u2 -> Function@8T0 , T1 , T2<, symbolList1.possibleTerms •• EvaluateD,

v2 -> Function@8T0 , T1 , T2<, symbolList2.possibleTerms •• EvaluateD<;

Substituting sol2Rule into order2Eqm and replacing E±I w T 0 on the right-hand sides of order2Eqm by E±I w0 T 0 , we have

order2Eqf = Subtract žž # & •ž order2Eqm •. sol2Rule •.

8Exp@I w T0D -> Exp@I w0 T0D, Exp@-I w T0D -> Exp@-I w0 T0 D< •• Expand;

Applying directly the function Coefficient to collect the coefficients of all possibleTerms could result in extra terms that are
functions of T0 ,which we eliminate. The result is

algEqs = Flatten@Coefficient@order2Eqf, #D & •ž possibleTermsD == 0 •. Exp@_ T0D -> 0 •• Thread

9-m - L1 + I G1 w0 == 0, I m w + w2 G1 + I L1 w0 == 0, -m - L2 - I G2 w0 == 0,
-I m w + w2 G2 - I L2 w0 == 0, I W G3 - L3 == 0, 2 I m W + w2 G3 + I W L3 == 0, -I W G4 - L4 == 0,
-2 I m W + w2 G4 - I W L4 == 0, -L5 == 0, 2 d + w2 G5 == 0, -2 I W G6 - L6 == 0,
Id
d + w2 G6 - 2 I W L6 == 0, 2 I W G7 - L7 == 0, d + w2 G7 + 2 I W L7 == 0, - €€€€€€€€ + I w G8 - I W G8 - L8 == 0,
w
d + w2 G8 + I w L8 - I W L8 == 0, I w G9 + I W G9 - L9 == 0, 2 d + w2 G9 + I w L9 + I W L9 == 0,
2 I w G10 - L10 == 0, d + w2 G10 + 2 I w L10 == 0, -I w G11 - I W G11 - L11 == 0,
Id
2 d + w2 G11 - I w L11 - I W L11 == 0, €€€€€€€€ - I w G12 + I W G12 - L12 == 0, d + w2 G12 - I w L12 + I W L12 == 0,
w
-L13 == 0, 2 d + w G13 == 0, -2 I w G14 - L14 == 0, d + w2 G14 - 2 I w L14 == 0=
2

Using the function Solve, we directly solve the system of algebraic equations algEqs for these undetermined coefficients and
obtain

var = Join@symbolList1, symbolList2D;

96 Chapter4.nb

coefs = Solve@algEqs, varD@@1DD;

Substituting coefs into sol2Rule and changing w0 back to w, we obtain the solution of the second-order equations in pure
function form as

sol2a = sol2Rule •. Function@8T0, T1 , T2 <, b_D :>

Function@8T0, T1, T2 < •• Evaluate, b •. coefs •. w0 -> w •• EvaluateD;
sol2a •. displayRule

:u2 ® FunctionB8T0 , T1 , T2<, - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

2 d L2 E-2 I T0 W d L2 E2 I T0 W d L2 2 I E-I T0 W L m W
w2 w2 - 4 W2 w2 - 4 W2 w2 - W2
IT W LmW I T w mA IT w-I T W d LA I T w+I T W dLA E2 I T0 w d A2
W H2 w + WL
2IE 0 IE 0 E 0 0 2E 0 0
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€ +
w2 - W2 2w wW 3 w2

€€€€€€€€€ - €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€ F,

• • • • •2
I E-IT0 w m A E-I T0 w+I T0 W d L A 2 E-IT0 w-I T0 W d L A E-2 IT0 w d A
W H2 w + WL
2dAA
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2w wW w2 3 w2

v2 ® FunctionB8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

2 I E-2 I T0 W d L2 W 2 I E2 I T0 W d L2 W 2 E-IT0 W L m W2

d L Hw + WL A
w -4 W
2 2 w -4W
2 2 w2 - W2
I T0 W I T0 w-I T0 W I T0 w+I T0 W
2
2 I E2I T0 w d A2
W H2 w + WL
2E Lm W 1 IE dLA 2IE
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€ EI T0 w m A - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w2 - W2 2 W 3w

2 I E-I T0 w-I T0 W d L Hw + WL A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ F>
• • • 2
• I E-I T0 w+I T0 W d L A 2 I E-2 I T0 w d A
W H2 w + WL
1
€€€€ E-I T0 w m A + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 W 3w

Approach 2:

As discussed above, solutions of the algebraic equations resulting from applying the method of undetermined coefficients
are unique except those corresponding to EI w T 0 and E-I w T 0 . To render them unique, we demand that they be orthogonal to
solutions of the corresponding adjoint problems; that is, we damand that the coefficient vector of EI w T 0 be 8u 2 , v2 < be
orthogonal to ccleftVec. To accomplish this, we first determine the form of all possible terms on the right-hand sides of
order2Eqm and then identify those among them corresponding to the resonance terms and those corresponding to nonreso -
nance terms.

possibleTerms =
collectForm ~ Join~HOuter@Times, collectForm, collectFormD •• Flatten •• UnionL

:EI T0 w A@T1 , T2D, E-I T0 w A@T1 , T2D, EIT0 W L, E-I T0 W L, L2 , E-2 I T0 W L2,
•

E2 I T0 W L2 , EI T0 w-I T0 W L A@T1, T2 D, EI T0 w+I T0 W L A@T1 , T2D, E2 I T0 w A@T1 , T2D2 ,

E-I T0 w-I T0 W L A@T1 , T2D, E-I T0 w+IT0 W L A@T1 , T2 D, A@T1, T2 D A@T1, T2D, E-2 I T0 w A@T1, T2D >
• • • • 2

Next, we identify the resonance terms among them by

RT = I# •. 8a_ •; a =!= 0 -> 1< & •ž IE-I wT0 possibleTerms •. expRule1 •. Exp@_ T0 + _.D -> 0MM
possibleTerms •• Union •• Rest

8EI T0 w A@T1 , T2D, E-I T0 w+IT0 W L A@T1 , T2 D<

•
Chapter4.nb 97

Hence, the nonresonance part of possibleTerms is the complement of RT; that is,

NRT = Complement@possibleTerms, Join@RT, RT •. conjugateRuleDD

:E-I T0 W L, EI T0 W L, L2 , E-2 I T0 W L2, E2 I T0 W L2 , EI T0 w+I T0 W L A@T1, T2 D,

E2 I T0 w A@T1, T2D2 , E-I T0 w-I T0 W L A@T1, T2 D, A@T1 , T2D A@T1 , T2D, E-2I T0 w A@T1 , T2 D >
• • • 2

Next, we associate with each possible resonance term an undetermined coefficient by using the rule

RTsymbolList = Table@Gj, 8j, Length@RTD<D

8G1 , G2<

and we associate with each possible nonresonance term an undetermined coefficient by using the rule

NRTsymbolList@i_D = Table@Li,j , 8j, Length@NRTD<D

8Li,1 , Li,2, Li,3 , Li,4, Li,5 , Li,6, Li,7 , Li,8, Li,9 , Li,10<

Hence, the coefficients of RT in the nonhomogeneous terms in order2Eqm are

coefRT = Coefficient@#@@2DD, RTD & •ž order2Eqm •. Exp@_ T0 + _.D -> 0

99m, - €€€€€€€€ =, 8-I m w, -d<=

Id
w

It follows from

ccleftVec.coefRT === 80, 0<

True

that coefRT is orthogonal to the solution of the adjoint, ccleftVec, as it should.

Imposing the condition that {u 2 , v2 } is orthogonal to ccleftVec, we seek u 2 and v2 in pure function form as

sol2Form = 8
u2 -> Function@8T0 , T1 , T2<, RTsymbolList.RT +
HRTsymbolList.RT •. conjugateRuleL + [email protected] •• EvaluateD, v2 ->
Function@8T0, T1, T2 <, -I w RTsymbolList.RT + H-I w RTsymbolList.RT •. conjugateRuleL +
[email protected] •• EvaluateD<;

Substituting sol2Form into order2Eqm and collecting coefficients of NRT, we obtain the following set of algebraic equations:
98 Chapter4.nb

algEqs1 = HCoefficient@Subtract žž #, NRTD & •ž Horder2Eqm •. sol2FormL •. Exp@_ T0 D -> 0 ••

FlattenL == 0 •• Thread

8-I W L1,1 - L2,1 == 0, I W L1,2 - L2,2 == 0, -L2,3 == 0, -2 I W L1,4 - L2,4 == 0,

2 I W L1,5 - L2,5 == 0, I w L1,6 + I W L1,6 - L2,6 == 0, 2 I w L1,7 - L2,7 == 0,
-I w L1,8 - I W L1,8 - L2,8 == 0, -L2,9 == 0, -2 I w L1,10 - L2,10 == 0, -2 I m W + w2 L1,1 - I W L2,1 == 0,
2 I m W + w2 L1,2 + I W L2,2 == 0, 2 d + w2 L1,3 == 0, d + w2 L1,4 - 2 I W L2,4 == 0,
d + w2 L1,5 + 2 I W L2,5 == 0, 2 d + w2 L1,6 + I w L2,6 + I W L2,6 == 0, d + w2 L1,7 + 2 I w L2,7 == 0,
2 d + w2 L1,8 - I w L2,8 - I W L2,8 == 0, 2 d + w2 L1,9 == 0, d + w2 L1,10 - 2 I w L2,10 == 0<

Solving algEqs1 for the undetermined coefficients NRTsymbolList yields

coef1 = Solve@algEqs1, Array@NRTsymbolList, 2D •• FlattenD@@1DD

:L2,9 ® 0, L1,3 ® - €€€€€€€€ , L1,9 ® - €€€€€€€€ , L2,3 ® 0, L1,1 ® €€€€€€€€€€€€€€€€ , L1,4 ® - €€€€€€€€€€€€€€€€€€€€ ,
2d 2d 2ImW d
w 2 w 2 w -W
2 2 w - 4 W2
2
2
2ImW d 2mW d 2 m W2
L1,2 ® - €€€€€€€€€€€€€€€€ , L1,5 ® - €€€€€€€€€€€€€€€€€€€€ , L2,1 ® €€€€€€€€€€€€€€€€ , L1,7 ® €€€€€€€€€€ , L2,2 ® €€€€€€€€€€€€€€€€ ,
w2 - W2 w2 - 4 W2 w2 - W2 3 w2 w2 - W2
2IdW d 2Id 2IdW 2Id
L2,5 ® - €€€€€€€€€€€€€€€€€€€€ , L1,10 ® €€€€€€€€€€ , L2,10 ® - €€€€€€€€€€€€ , L2,4 ® €€€€€€€€€€€€€€€€€€€€ , L2,7 ® €€€€€€€€€€€€ ,

2 I d Hw + WL 2 I d Hw + WL
w -4W
2 2 3w 2 3w w -4W
2 2 3w

L2,6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L1,6 ® €€€€€€€€€€€€€€€€€€€€€€€€ , L1,8 ® €€€€€€€€€€€€€€€€€€€€€ , L2,8 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

W H2 w + WL W H2 w + WL
2d 2d
2 w W + W2 2 w W + W2

Substituting sol2Form into either of the two order2Eqm and collecting the coefficients of RT, we obtain the following set of
algebraic equations:

algEqs2 =
Coefficient@Subtract žž order2Eqm@@1DD •. sol2Form, RTD == 0 •. Exp@_ T0 D -> 0 •• Thread

9-m + 2 I w G1 == 0, €€€€€€€€ + I W G2 == 0=
Id
w

Solving algEqs2 for the undetermined coefficients RTsymbolList yields

coef2a = Solve@algEqs2, RTsymbolListD@@1DD

9G1 ® - €€€€€€€€ , G2 ® - €€€€€€€€ =

Im d
2w wW

coef2 = Join@coef2a, coef2a •. conjugateRuleD

9G1 ® - €€€€€€€€ , G2 ® - €€€€€€€€ , G1 ® €€€€€€€€ , G2 ® - €€€€€€€€ =

Im d • Im • d
2w wW 2w wW

Substituting coef1 and coef2 into sol2Form, we obtain the solution of the second-order equations in pure function form as
Chapter4.nb 99

sol2b = sol2Form •. Function@8T0, T1 , T2 <, b_D :>

Function@8T0, T1, T2 < •• Evaluate, b •. coef1 •. coef2 •• Expand •• EvaluateD;
sol2b •. displayRule

:u2 ® FunctionB8T0 , T1 , T2<, - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

2 d L2 E-2 I T0 W d L2 E2 I T0 W d L2 2 I E-I T0 W L m W
w2 w2 - 4 W2 w2 - 4 W2 w2 - W2
W w w-I W w+I W E2 I T0 w d A2
W H2 w + WL
2IE IT 0 LmW IE I T 0 mA EIT 0 T0 d LA 2E I T0 T0 dLA
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€ +
w2 - W2 2w wW 3 w2

€€€€€€€€€ - €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€ F,

• • • • •2
I E-IT0 w m A E-I T0 w+I T0 W d L A 2 E-IT0 w-I T0 W d L A 2dAA E-2 IT0 w d A
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2w wW 2 w W + W2 w2 3 w2

v2 ® FunctionB8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

2 I E-2 I T0 W d L2 W 2 I E2 I T0 W d L2 W 2 E-IT0 W L m W2
w2 - 4 W2 w2 - 4 W2 w2 - W2
I T0 W 2 I T0 w-I T0 W I T0 w+I T0 W
2E Lm W 1 IE dLA 2IE dLA
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€ EI T0 w m A - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ +
w2 - W2 2 W 2 w +W
•
2 I EIT0 w+I T0 W d L w A 2 I E2 I T0 w d A2 1 -IT0 w • I E-I T0 w+IT0 W d L A
W H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€ E m A + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ -
3w 2 W

€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ F>

• • •2
2 I E-I T0 w-I T0 W d L w A 2 I E-I T0 w-I T0 W d L W A 2 I E-2 I T0 w d A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w W + W2 2 w W + W2 3w

These two approaches generate the same results because

Subtract žž Hu2 @T0, T1, T2 D •. 8sol2a, sol2b<L •• Simplify

Subtract žž Hv2 @T0, T1, T2 D •. 8sol2a, sol2b<L •• Simplify

The second approach is more convenient to use for a general system of equations. Hence, we will adopt this approach in the
remainder of this book.

Ÿ Third-Order Equations

Substituting the first- and second-order solutions into the third-order equations, eqOrder[3], yields

order3Eq = eqOrder@3D •. sol1 •. sol2b •• ExpandAll;

Using the expRule1 to convert the terms that produce small-divisor terms into terms that produce secular terms in the right-
hand sides of order3Eq and collecting the terms that could produce secular terms, we have
100 Chapter4.nb

STerms2 = Coefficient@#@@2DD •. expRule1 •. W -> 2 w, Exp@I w T0 DD & •ž order3Eq;

STerms2 •. displayRule

I m HD1AL EI T1 s d L HD1AL
: €€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - D2 A, -6 A a L2 + A m2 + €€€€€€€€€€€€€€€€€€€€ - 3 A2 a A +
•
9 A d2 L2 •

I EI T1 s d L HD1 AL
2w 2w 2 2w 2

€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€ m HD1 AL - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - I w HD2AL>

• • •
10 A2 d2 A 14 I EI T1 s d L m A 1
3 w2 3w 2 2w

Then, the solvability condition demands that STerms2 be orthogonal to the solution of the adjoint ccleftVec. The result is

SCond2 = SolveAccleftVec.STerms2 == 0, AH0,1L @T1, T2 DE@@1DD •• ExpandAll;

SCond2 •. displayRule

:D2 A ® - €€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ >

• • •
9 I A d2 L2 3 I A a L2 I A m2 5 I A2 d2 A 7 EI T1 s d L m A 3 I A2 a A
4 w3 w 2w 3 w3 3 w2 2w

The two partial-differential equations, SCond1 and SCond2, can be reconstituted to obtain an ordinary-differential equation
governing A. The result is

H2 I w A¢ == H2 I w dt@1D@A@T1 , T2DD •. SCond1 •. SCond2 •• Expand •• Collect@#, eD &LL •.

displayRule

2 I w A¢ ==
i
j 14 I EI T1 s d L m A y
e H-2 I A m w - 2 EI T1 s d L AL + e2 j €€ z
z
• •
j z
• 9 A d2 L2 2 • 10 A2 d2 A

k {
2 2
-6 A a L + A m + €€€€€€€€
€ €€€€€€€
€ €€
€ - 3 A a A + €€€€€€€€€€€€€€€€
€€€€
€ € - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€
2 w 2 3 w 2 3 w

Ÿ 4.2.6 First-Order Complex-Valued System

In this section, we first transform FDuffingEq into a single first-order complex-valued equation using the transformation

transformRule = 9u -> Iz + zM, u¢ -> I w Iz - zM=;

• •

gRule = g -> - 2 e m u¢ - d u2 - a u3 + e f Cos@W tD;

then FDuffingEq becomes

eq426a = z¢ == I w z - €€€€€€€€ g •. gRule •. transformRule •• ExpandAll

I
2w

I d z2 I a z3
z¢ == -e z m + €€€€€€€€€€€€€€ + €€€€€€€€€€€€€€ + I z w -
2w 2w
• • •2 •2 •3
I f e Cos@t WD • Idzz 3 I a z2 z Idz 3Iazz Iaz
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + e m z + €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€
2w w 2w 2w 2w 2w

To determine a second-order uniform expansion of the solution of eq426a using the method of multiple scales, we first
introduce

multiScales = 9z -> z@T0, T1, T2 D, z -> z@T0, T1, T2 D, z¢ -> dt@1D@z@T0 , T1 , T2DD, t -> T0 =;
• •
Chapter4.nb 101

••
and then expand z and z in the form

solRule =
9z -> ISumAej zj @#1, #2, #3D, 8j, 1, 3<E &M, z -> ISumAej zj @#1, #2, #3D, 8j, 1, 3<E &M=;
• •

Substituting multiScales and solRule into eq426a, expanding the result for small e, and discarding terms of order higher than
e3 , we obtain

eq426b = Heq426a •. multiScales •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e yields

eqEps = CoefficientList@Subtract žž eq426b, eD == 0 •• Thread •• Rest •• TrigToExp;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := HeqEps@@1, 1DD •. f -> 0 •. z1 -> zi L - # & •ž eqEps@@iDD •• Reverse

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

Array@eqOrder, 3D •. displayRule •• TableForm

I HE -I T0 W
+E Lf
I T0 W
D0 z1 - I w z1 == - €€€€€€€€€€€€€€€€
€€€€€€€€
4w
€€€€€€€€
€

D0 z2 - I w z2 == -HD1 z1 L - m z1 + €€€€€€€€
• • •2
I d z21 I d z1 z1 I d z1
2w
€€ + m z1 + €€€€€€€€
w
€ €€€
€ € + €€€€€€€€
2w
€€

D0 z3 - I w z3 == -HD1 z2 L - D2z1 + €€€€€€€€

• • •2 •3 • •
I a z31 I d z1 z2 3I a z21 z1 I d z2 z1 3 I a z1 z1 I a z1 I d z1 z2
€€ - m z2 + €€€€€€€€
€€€€€€€ + €€€€€€€€
€€€€€€€€€ + €€€€€€€€
€€€€€€€ + €€€€€€€€
€€€€€€€€€ + €€€€€€€€
€€ + m z2 + €€€€€€€€
€€€€€€€
2w w 2w w 2w 2w w

The general solution of the first-order equation, eqOrder[1], can be expressed in pure function form as

sol1Form =
HDSolve@eqOrder@1D, z1@T0 , T1 , T2D, timeScalesD@@1, 1, 2DD •. C@1D -> A •• ExpandL •.
Exp@a_D :> Exp@a •• ExpandD ••. c1_ Exp@x_D + c2_ Exp@x_D :> Factor@c1 + c2D Exp@xD

€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + EI T0 w A@T1 , T2D

EI T0 W f E-I T0 W f
4 w Hw - WL 4 w Hw + WL

sol1 = 9z1 -> Function@8T0 , T1 , T2<, sol1Form •• EvaluateD,

z1 -> Function@8T0 , T1 , T2<, sol1Form •. conjugateRule •• EvaluateD=
•

:z1 ® FunctionB8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + EI T0 w A@T1 , T2 DF,

EI T0 W f E-I T0 W f
4 w Hw - WL 4 w Hw + WL

z1 ® FunctionB8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + E-IT0 w A@T1 , T2 DF>

• E-I T0 W f EI T0 W f •
4 w Hw - WL 4 w Hw + WL

Substituting the first-order solution into the second-order equation, eqOrder[2], we have
102 Chapter4.nb

order2Eq = Expand •ž HeqOrder@2D •. sol1L;

order2Eq •. displayRule

D0 z2 - I w z2 ==
I A2 E2 I T0 w d I f2 d I E-2 I T0 W f2 d I E2 IT0 W f2 d I A EI T0 w-I T0 W f d
16 w3 Hw - WL2 32 w3 Hw - WL2 32 w3 Hw - WL2 4 w2 Hw - WL
-A EI T0 w m + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2w
I A EI T0 w+I T0 W f d E-IT0 W f m EI T0 W f m I f2 d I E-2 IT0 W f2 d I E2 I T0 W f2 d
4 w2 Hw - WL 4 w Hw - WL 4 w Hw - WL 16 w3 Hw + WL2 32 w3 Hw + WL2 32 w3 Hw + WL2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

I A EI T0 w-I T0 W f d I A EI T0 w+I T0 W f d E-I T0 W f m EI T0 W f m I f2 d

4 w2 Hw + WL 4 w2 Hw + WL 4 w Hw + WL 4 w Hw + WL 8 w3 Hw - WL Hw + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ +
• •
I E-2 IT0 W f2 d I E2 I T0 W f2 d I E-I T0 w-I T0 W f d A
16 w3 Hw - WL Hw + WL 16 w3 Hw - WL Hw + WL 4 w2 Hw - WL
-I T0 w • IAdA
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ + E m A + €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ +
w

€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - EI T0 w HD1 AL

• • • •2
I E-I T0 w+I T0 W f d A I E-I T0 w-I T0 W f d A I E-I T0 w+I T0 W f d A I E-2 I T0 w d A
4 w2 Hw - WL 4 w2 Hw + WL 4 w2 Hw + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2w

Eliminating the terms that lead to secular terms in z2 from order2Eq yields

SCond1 = Coefficient@order2Eq@@2DD •. expRule1, Exp@I w T0 DD == 0

I EI T1 s f d A@T1, T2 D I EI T1 s f d A@T1 , T2 D
-m A@T1, T2 D + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ - AH1,0L @T1 , T2D == 0
• •

4 w2 Hw - WL 4 w2 Hw + WL
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

or

SCond1Rule =
ISolveASCond1, AH1,0L@T1 , T2DE@@1DD •• ExpandAllM •. c1_ m + c2_ m :> Factor@c1 + c2D m

I EIT1 s f d A@T1 , T2D

:A @T1 , T2 D ® -m A@T1 , T2D + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ >
•
H1,0L
2 w3 - 2 w W2

With this solvability condition, we rewrite order2Eq as

order2Eqm = Expand •ž
Horder2Eq •. SCond1Rule •. Exp@arg_D :> Exp@arg •. T1 -> e T0 •. sigRule •• ExpandDL;

There are at least two approaches we can use to solve for the particular solution of order2Eqm. In the first approach, we
transform order2Eqm into an ordinary-differential equation and then use DSolve and Map to determine the following particu-
lar solution:
Chapter4.nb 103

Hsol2a =
Hz2 @T0D •. DSolve@Horder2Eqm@@1DD •. z2 -> Hz2 @#1D &LL == #, z2 @T0D, T0 D@@1DD •. C@_D ->
0L & •ž order2Eqm@@2DD •. Exp@a_D :> Exp@Expand@aDD •• ExpandL •• Timing

:6.629 Second, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

f2 d I EI T0 W f m E2 I T0 W f2 d f2 d
16 w4 Hw - WL2 4 w Hw - WL2 32 w3 Hw - 2 WL Hw - WL2 16 w4 Hw + WL2
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

I E-I T0 W f m E2 IT0 W f2 d f2 d I E-I T0 W f m

4 w Hw + WL 32 w3 Hw - 2 WL Hw + WL 8 w Hw - WL Hw + WL 4 w Hw - WL Hw + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ +
2 2 4

I EI T0 W f m E2 I T0 W f2 d E-2 I T0 W f2 d
4 w Hw - WL Hw + WL 16 w3 Hw - 2 WL Hw - WL Hw + WL 32 w3 Hw - WL2 Hw + 2 WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ -

E-2 I T0 W f2 d E-2 I T0 W f2 d EIT0 w-I T0 W f d A@T1 , T2 D

32 w3 Hw + WL2 Hw + 2 WL 16 w3 Hw - WL Hw + WL Hw + 2 WL 4 w2 Hw - WL W
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ +

EI T0 w+I T0 W f d A@T1, T2 D EI T0 w-I T0 W f d A@T1, T2 D EIT0 w+I T0 W f d A@T1 , T2 D

4 w2 Hw - WL W 4 w2 W Hw + WL 4 w2 W Hw + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ +

E2 I T0 w d A@T1 , T2 D2 I E-I T0 w m A@T1, T2 D E-IT0 w+I T0 W f d A@T1 , T2 D

• •

4 w2 Hw - WL H2 w - WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ -
2 w2 2w
E-I T0 w+I T0 W f d A@T1, T2 D E-IT0 w-I T0 W f d A@T1 , T2 D E-I T0 w-IT0 W f d A@T1 , T2D
• • •

4 w2 H2 w - WL Hw + WL 4 w2 Hw - WL H2 w + WL 4 w2 Hw + WL H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ +

E-I T0 w+I T0 W f d A@T1, T2 D d A@T1 , T2 D A@T1, T2 D E-2I T0 w d A@T1 , T2D

€€€€€€€€€€€€€€€ >
• • • 2

2 w H2 w - WL Hw2 - W2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w2 6 w2

In the second approach, we multiply order2Eqm by an integrating factor and then use the function Integrate. To reduce the
computation time, we define first the rule

intRule = int@a_ + b_, arg_D -> int@a, argD + int@b, argD;

Then, using this rule and an integrating factor, we find a particular solution of order2Eqm as
104 Chapter4.nb

Hsol2b = HHint@order2Eqm@@2DD * Exp@-I T0 wD •• Expand, T0D ••. intRule •. int -> IntegrateL *
Exp@I T0 wD •• ExpandL •. Exp@a_D :> Exp@Expand@aDDL •• Timing

:0.841 Second, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

f2 d I EI T0 W f m E2 I T0 W f2 d f2 d
16 w4 Hw - WL2 4 w Hw - WL2 32 w3 Hw - 2 WL Hw - WL2 16 w4 Hw + WL2
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

I E-I T0 W f m E2 IT0 W f2 d f2 d I E-I T0 W f m

4 w Hw + WL2 32 w3 Hw - 2 WL Hw + WL2 8 w4 Hw - WL Hw + WL 4 w Hw - WL Hw + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ +

I EI T0 W f m E2 I T0 W f2 d E-2 I T0 W f2 d
4 w Hw - WL Hw + WL 16 w Hw - 2 WL Hw - WL Hw + WL 32 w3 Hw - WL2 Hw + 2 WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ -
3

E-2 I T0 W f2 d E-2 I T0 W f2 d EIT0 w-I T0 W f d A@T1 , T2 D

32 w3 Hw + WL2 Hw + 2 WL 16 w3 Hw - WL Hw + WL Hw + 2 WL 4 w2 Hw - WL W
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ +

EI T0 w+I T0 W f d A@T1, T2 D EI T0 w-I T0 W f d A@T1, T2 D EIT0 w+I T0 W f d A@T1 , T2 D

4 w2 Hw - WL W 4 w2 W Hw + WL 4 w2 W Hw + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ +

E2 I T0 w d A@T1 , T2 D2 I E-I T0 w m A@T1, T2 D E-IT0 w+I T0 W f d A@T1 , T2 D

• •

4 w2 Hw - WL H2 w - WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ -
2 w2 2w
E-I T0 w+I T0 W f d A@T1, T2 D E-IT0 w-I T0 W f d A@T1 , T2 D E-I T0 w-IT0 W f d A@T1 , T2D
• • •

4 w2 H2 w - WL Hw + WL 4 w2 Hw - WL H2 w + WL 4 w2 Hw + WL H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ +

E-I T0 w+I T0 W f d A@T1, T2 D d A@T1 , T2 D A@T1, T2 D E-2I T0 w d A@T1 , T2D

€€€€€€€€€€€€€€€ >
• • • 2

2 w H2 w - WL Hw2 - W2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w2 6 w2

The CPU time is decreased by a factor of

%%%@@1DD • %@@1DD

7.88228

sol2a === sol2b

True

Then, the particular solution of order2Eqm can be expressed in pure function form as

sol2 = 9z2 -> Function@8T0 , T1 , T2<, sol2b •• EvaluateD,

z2 -> Function@8T0 , T1 , T2<, sol2b •. conjugateRule •• EvaluateD=;
•

Substituting the first- and second-order solutions into the third-order equation, eqOrder[3], we obtain

order3Eq = Expand •ž HeqOrder@3D •. sol1 •. sol2L;

Eliminating the terms that produce secular terms in z3 from order3Eq yields
Chapter4.nb 105

SCond2Rule =
ISolveACoefficient@order3Eq@@2DD •. expRule1, Exp@I w T0 DD == 0, AH0,1L @T1, T2DE@@1DD •.
SCond1Rule •. HSCond1Rule •. conjugateRuleL •• ExpandAllM ••.
9c1_ m + c2_ m :> Factor@c1 + c2D m, c1_ f2 d2 + c2_ f2 d2 :> Factor@c1 + c2D f2 d2,
c1_ f2 a + c2_ f2 a :> Factor@c1 + c2D f2 a=;
SCond2Rule •. displayRule

I A f2 d2 H5 w + 2 WL
:D2 A ® - €€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
I A m2 3 I A f2 a
4 w Hw - WL2 Hw + WL2 4 w3 Hw - WL2 Hw + WL2 H2 w + WL
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2w
EI T1 s f d m H2 w3 + w2 W - 6 w W2 - W3 L A
€€€€€€€€€€€€€€€€€€€ >
• • •
5 I A2 d2 A 3 I A2 a A
4 w2 Hw - WL2 W Hw + WL2
€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 w3 2w

or

fRule = 9f -> 2 Iw2 - W2 M L=;

SCond2Rule •. fRule •. W -> 2 w •. displayRule

:D2 A ® - €€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ >

• • •
9 I A d2 L2 3 I A a L2 I A m2 5 I A2 d2 A 7 EI T1 s d L m A 3 I A2 a A
4 w3 w 2w 3 w3 3 w2 2w

which is in agreement with that obtained in the preceding section.

Ÿ 4.2.7 The Function MMS1

Collecting the steps described in Section 4.2.5 for a system of two real-valued first-order equations, we can build a function
named MMS1 (Method of Multiple Scales for 1DOF system) specifically for FDuffingEq. A more general function (a Package)
can be similarly created by considering as arguments the governing equation, symbols for the dependent variable, indepen-
dent variable, excitation amplitudes and frequencies, and all other related quantities which allow the program to identify their
respective meanings. We then use MMS1 to solve for different resonance cases.

MMS1@scaling_List, ResonanceCond : 8__Equal<D :=

ModuleA8<,
vRule = u¢@tD -> v@tD;
OmgRule = Solve@ResonanceCond, WD@@1DD;
eqa = 8vRule •. Rule -> Equal, FDuffingEq •. scaling •. vRule •. D@vRule, tD<;
multiScales =
8u@tD -> u@T0 , T1, T2 D, v@tD -> v@T0, T1, T2 D, u_¢ @tD -> dt@1D@u@T0 , T1, T2DD, t -> T0 <;
solRule = 9u -> ISumAej uj @#1, #2, #3D, 8j, 3<E &M,
v -> ISumAej vj @#1, #2, #3D, 8j, 3<E &M=;
eqb = Heqa •. multiScales •. solRule •• TrigToExp •• ExpandAllL •. e n_•;n>3 -> 0;
eqEps =
Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eqb •• Transpose •• Rest;

H* First-Order Problem *L
106 Chapter4.nb

linearSys = #@@1DD & •ž eqEps@@1DD •. f -> 0;

eqOrder@i_D :=
HlinearSys •. u_1 -> ui L == HlinearSys •. u_1 -> ui L - H#@@1DD & •ž eqEps@@iDDL •• Thread;
coefList = E-I wT0 linearSys •. 9u1 -> IP EI w # &M, v1 -> IQ EI w # &M= •• Expand;
coefMat = Outer@Coefficient, coefList, 8P, Q<D;
hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose;
rightVec = 81, c1< •. [email protected], c1<L@@1DD == 0, c1D@@1DD;
leftVec = 8c1, 1< •. Solve@[email protected], 1<L@@1DD == 0, c1D@@1DD;
ccleftVec = leftVec •. conjugateRule;
sol1h = # + H# •. conjugateRuleL & •ž IrightVec A@T1, T2 D EI wT0 M;
sol1p =
DSolve@eqOrder@1D •. 8u1 -> Hu1 @#1D &L, v1 -> Hv1@#1D &L<, 8u1@T0 D, v1@T0 D<, T0 D@@1DD •.
C@_D -> 0 •• TrigToExp •• Simplify;
fRule = f -> -2 I-w2 + W2 M L;
sol1 =
8u1 -> Function@8T0 , T1 , T2<, sol1h@@1DD + Hu1@T0 D •. sol1p •. fRule •• ExpandL •• EvaluateD,
v1 -> Function@8T0 , T1 , T2<,
sol1h@@2DD + Hv1 @T0D •. sol1p •. fRule •• ExpandL •• EvaluateD<;

H* Second-Order Problem *L
order2Eq = eqOrder@2D •. sol1 •• ExpandAll;
expRule1 =
Exp@arg_D :> Exp@Expand@arg •. OmgRuleD •. e n_. T0 :> timeScales@@n + 1DDD;
STerms1 = Coefficient@#@@2DD •. expRule1, Exp@I w T0 DD & •ž order2Eq;
SCond1 = SolveAccleftVec.STerms1 == 0, AH1,0L @T1, T2 DE@@1DD;
ccSCond1 = SCond1 •. conjugateRule;
sigRule = Solve@OmgRule •. Rule -> Equal, sD@@1DD;
expRule2 = Exp@a_D :> ExpAa •. 9T1 -> e T0, T2 -> e2 T0= •. sigRule •• ExpandE;
SCond1f = Join@SCond1, ccSCond1D •. expRule2;
order2Eqm = order2Eq •. SCond1f •• ExpandAll;
collectForm = JoinA9EIT0 w A@T1 , T2 D, E-I T0 w A@T1 , T2 D=,
•

IfAHF • f •. scalingL === e, 9EI T0 W L, E-I T0 W L=, 8<EE;

possibleTerms = collectForm ~Join~
HOuter@Times, collectForm, collectFormD •• Flatten •• UnionL;
RT = I# •. 8a_ •; a =!= 0 -> 1< & •ž IE-I w T0 possibleTerms •. expRule1 •.
Exp@_ T0 + _.D -> 0MM possibleTerms •• Union •• Rest;
NRT = Complement@possibleTerms, Join@RT, RT •. conjugateRuleDD;
RTsymbolList = Table@Gj, 8j, Length@RTD<D;
NRTsymbolList@i_D = Table@Li,j , 8j, Length@NRTD<D;
sol2Form = 8
u2 -> Function@8T0 , T1, T2 <,
RTsymbolList.RT + HRTsymbolList.RT •. conjugateRuleL + [email protected] ••
EvaluateD, v2 -> Function@8T0 , T1, T2<, -I w RTsymbolList.RT +
•• D<
Chapter4.nb 107

H-I w RTsymbolList.RT •. conjugateRuleL + [email protected] •• EvaluateD<;

algEqs1 = HCoefficient@Subtract žž #, NRTD & •ž Horder2Eqm •. sol2FormL •.
Exp@_ T0D -> 0 •• FlattenL == 0 •• Thread;
coef1 = Solve@algEqs1, Array@NRTsymbolList, 2D •• FlattenD@@1DD;
algEqs2 =
Coefficient@Subtract žž order2Eqm@@1DD •. sol2Form, RTD == 0 •. Exp@_ T0 D -> 0 •• Thread;
coef2a = Solve@algEqs2, RTsymbolListD@@1DD;
coef2 = Join@coef2a, coef2a •. conjugateRuleD;
sol2 = sol2Form •. Function@8T0 , T1 , T2<, b_D :>
Function@8T0, T1 , T2 < •• Evaluate, b •. coef1 •. coef2 •• Expand •• EvaluateD;

H* Third-Order Problem *L
order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;
STerms2 =
Coefficient@#@@2DD •. expRule1 •. HOmgRule •. e -> 0L, Exp@I w T0 DD & •ž order3Eq;
SCond2 = SolveAccleftVec.STerms2 == 0, AH0,1L @T1, T2 DE@@1DD •• ExpandAll;

H* Reconstitution *L
moduEq = 2 I w A¢ == H2 I w dt@1D@ A@T1, T2 DD •. SCond1 •. SCond2 •• Collect@#, eD &L;
Print@"The second-order approximate solution:"D;
Print@
u@tD == Hu žž timeScales •. solRule •. e^ 3 -> 0 •. sol1 •. sol2 •. displayRuleLD;
IfAHF • f •. scalingL === e, PrintA"where\n L==f•H2Hw2 -W2LL"E
E;
Print@"\nThe modulation equations:"D;
Print@moduEq •. displayRuleD
E

Ÿ Primary Resonance: W » w

scaling1 = 9F -> e3 f, m -> e2 m=;

`

ResonanceCond1 = 9W == w + e2 s=;
108 Chapter4.nb

MMS1@scaling1, ResonanceCond1D •• Timing

The second-order approximate solution:

i
j A E E -2 I T0 w d A y z
u@tD == e IA E I T0 w + E -I T0 w A M + e 2 j € €€ z
• •2
j z
2 2 I T0 w d 2AdA
•

k {
€€€€€€€€€€€€€€€€
€€€€€€€€
€ - €€€€€€€€
€ €€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€
3 w2 w2 3 w2

The modulation equations:

i
j - A m - €€€€€€€€€€€€€€€€ 3 I A2 a A y z
2 I w A ¢ == 2 I e 2 w j z
• •
j z
I E I T2 s f 5 I A 2 d2 A

k {
€€€€
€ - €€€€€€€€€€€€€€€€
€€€€
€ €€ + €€€€€€€€€€€€€€€€
€€€€€
4w 3 w3 2w

81.513 Second, Null<

Ÿ Subharmonic Resonance of Order 1/2: W » 2w

scaling2 = 8F -> e f, m -> e m<;

`

ResonanceCond2 = 8W == 2 w + e s<;

MMS1@scaling2, ResonanceCond2D •• Timing

The second-order approximate solution:

u@tD ==
i
j A E
e IA E I T0 w + E-I T0 W L + E I T0 W L + E -I T0 w A M + e 2 j
j €€€€€€€€€€€€€€€€
2 2 I T0 w d
• 2 d L2 I A E I T0 w m A E I T0 w-I T0 W d L

k
€€€€€€€€€ - €€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€ +
3w 2 w 2 2w wW

2 A E I T0 w+I T0 W d L E -2 I T0 W d L 2 E 2 I T0 W d L 2 2 I E -I T0 W L m W 2 I E I T0 W L m W
W H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€ - €€€€€€€€€€€€€€€€
€€€€€ €€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€ -
w2 - 4 W2 w2 - 4 W2 w2 - W2 w2 - W2

E -2 I T0 w d A y z
€€€€€€€€€ €€ z
• • • • •2
z
2 A dA I E -I T0 w m A E -I T0 w+I T0 W d L A 2 E -I T0 w-I T0 W d L A

{
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€€ + €€€€€€€€€€€€€€€€
w2 2w wW 2 w W + W2 3 w2

L==f•H2Hw 2 -W2 LL
where

The modulation equations:

2 I w A ¢ == -2 I e IA m w - I E I T1 s d L A M +
•

i
j - €€€€€€€€€€€€€€€€ 3 I A2 a A y
2 I e2 w j €zz
• • •
j z
9 I A d2 L2 3 I A a L2 I A m2 5 I A2 d2 A 7 E I T1 s d L m A

k {
€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€
€ - €€€€€€€€
€ €€€
€ - €€€€€€€€€€€€€€€€
€€€€
€ €€ - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ €€€ + €€€€€€€€€€€€€€€€
€€€€
4 w3 w 2w 3 w3 3 w2 2w

88.562 Second, Null<

Ÿ Subharmonic Resonance of Order 1/3: W » 3w

scaling3 = 9F -> e f, m -> e2 m=;

`

ResonanceCond3 = 9W == 3 w + e2 s=;
Chapter4.nb 109

MMS1@scaling3, ResonanceCond3D •• Timing

The second-order approximate solution:

u@tD == e IA E I T0 w + E -I T0 W L + E I T0 W L + E -I T0 w A M +
•

i
j
j €€€€€€€€€€€€€€€€
e2 j
A 2 E 2 I T0 w d 2 d L2 2 A E I T0 w-I T0 W d L 2 A E I T0 w+I T0 W d L E -2 I T0 W d L 2

k H2 w - WL W W H2 w + WL
€€€€€ €€€€€ - €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€ -
3w 2 w 2 w2 - 4 W2

E -2 I T0 w d A y z
€€€€€€€€€ €€ z
• • • •2
z
E 2 I T0 W d L 2 2AdA 2 E -I T0 w+I T0 W d L A 2 E -I T0 w-I T0 W d L A
H2 w - WL W W H2 w + WL {
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€
w2 - 4 W 2 w2 3 w2

L==f•H2Hw 2 -W2 LL
where

The modulation equations:

2 I w A ¢ ==
i
j 3 I E I T2 s a L A y z
2 I e2 w j €€€€€€€€€ €€€€€€€€ € z
• • •2 •2
j - A m - €€€€€€€€€€€€€€€€ z
6 I A d 2 L2 3 I A a L2 5 I A2 d2 A 3 I A2 a A I E I T2 s d 2 L A

k {
€€€€€ €€ + €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€ + €€€€€€€€€€€€€€€€
5 w3 w 3 w3 2w w3 2w

87.03 Second, Null<

Ÿ Superharmonic Resonance of Order 2: 2W » w

scaling4 = 8F -> e f, m -> e m<;

`

ResonanceCond4 = 82 W == w + e s<;

MMS1@scaling4, ResonanceCond4D •• Timing

The second-order approximate solution:

i A 2 E 2 I T0 w d
u@tD ==

j
e IA E I T0 w + E-I T0 W L + E I T0 W L + E -I T0 w A M + e 2 j j €€€€€€€€€€€€€€€€
2 d L2 I A E I T0 w m 2 A E I T0 w+I T0 W d L
W H2 w + WL
•

k
€€€€€€€€€ - €€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ €€ -
3 w2 w2 2w

E -2 I T0 W d L 2 E 2 I T0 W d L 2 2 I E -I T0 W L m W 2 I E I T0 W L m W 2 A E I T0 w-I T0 W d L
2 w Hw + 2 WL 2 w Hw + 2 WL
€€€€€€€€€€€€€€€€€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ €€ -
w -W
2 2 w -W
2 2 2 w W - W2

E -2 I T0 w d A z y
€€€€€€€€€ €€ z
• • • • •2
z
I E -I T0 w m A 2 E -I T0 w-I T0 W d L A 2 E -I T0 w+I T0 W d L A
W H2 w + WL
2 A dA

{
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
w2 2w 2 w W - W2 3 w2

L==f•H2Hw 2 -W2 LL
where

The modulation equations:

2 I w A ¢ == -I e H- I E I T1 s d L 2 + 2 A m wL +
i
j 3 I A2 a A y
2 I e wj €€€€€ z
z
• •
j - €€€€€€€€€€€€€€€€ z
46 I A d2 L 2 13 E I T1 s d L 2 m 3 I A a L2 I A m2 5 I A2 d2 A

k {
2
€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€
€€€€€ €€ + €€€€€€€€€€€€€€€€
15 w 3 12 w 2 w 2w 3 w3 2w

88.993 Second, Null<

110 Chapter4.nb

Ÿ Superharmonic Resonance of Order 3: 3W » w

scaling5 = 9F -> e f, m -> e2 m=;

`

ResonanceCond5 = 93 W == w + e2 s=;

MMS1@scaling5, ResonanceCond5D •• Timing

The second-order approximate solution:

u@tD == e IA E I T0 w + E -I T0 W L + E I T0 W L + E -I T0 w A M +
•

i
jA E
e2 j
j €€€€€€€€€€€€€€€€
2 2 I T0 w d 2 d L2 2 A E I T0 w-I T0 W d L 2 A E I T0 w+I T0 W d L E -2 I T0 W d L 2
H2 w - WL W W H2 w + WL
k
€€€€€ €€€€€ - €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€ -
3 w2 w2 w2 - 4 W2

E -2 I T0 w d A y z
€€€€€€€€€ €€ z
• • • •2
z
E 2 I T0 W d L 2 2AdA 2 E -I T0 w+I T0 W d L A 2 E -I T0 w-I T0 W d L A
H2 w - WL W W H2 w + WL
{
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€
w2 - 4 W 2 w2 3 w2

L==f•H2Hw 2 -W2 LL
where

The modulation equations:

2 I w A ¢ ==
i 3 I A2 a A y
2 I e2 w j
j z
z
• •
j z
106 I A d 2 L 2 9 I E I T2 s d2 L 3 3 I A a L2 I E I T2 s a L 3 5 I A2 d2 A

k {
- A m - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€€€€
€ + €€€€€€€€€€€€€€€€
€€€€
€ + €€€€€€€€€€€€€€€€
€€€€€€€€
€€ - €€€€€€€€€€€€€€€€
€€€€
€ €€ + €€€€€€€€€€€€€€€€
€€€€€
35 w 3 5 w 3 w 2 w 3 w 3 2 w

87.08 Second, Null<

Ÿ Slowly-Modulated Load: W » 0

scaling6 = 8F -> e f, m -> e m<;

`

ResonanceCond6 = 8W == e s<;
Chapter4.nb 111

MMS1@scaling6, ResonanceCond6D •• Timing

The second-order approximate solution:

u@tD ==
i
j A E
e IA E I T0 w + E-I T0 W L + E I T0 W L + E -I T0 w A M + e 2 j
j €€€€€€€€€€€€€€€€
2 2 I T0 w d 2 d L2 I A E I T0 w m A E I T0 w-I T0 W d L
w H2 w - WL
•

k
€€€€€€€€€ - €€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€ -
3 w2 w2 2w

A E I T0 w+I T0 W d L E -2 I T0 W d L 2 E 2 I T0 W d L 2 2 I E -I T0 W L m W 2 I E I T0 W L m W
w H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ €€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€ - €€€€€€€€€€€€€€€€ €€€€€ €€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ -
w -4W
2 2 w -4W
2 2 w -W
2 2 w2 - W2

E -2 I T0 w d A y z
€€€€€€€€€ €€ z
• • • • •2
z
z
I E -I T0 w m A E -I T0 w+I T0 W d L A E -I T0 w-I T0 W d L A
w H2 w - WL w H2 w + WL
2 A dA

{
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€
w2 2w 3 w2

L==f•H2Hw 2 -W2 LL
where

The modulation equations:

2 I w A ¢ == -2 I A E -I T1 s e H-I d L - I E 2 I T1 s d L + E I T1 s m wL +
i 3 I A d2 L2
2 I e2 w j j
j - €€€€€€€€€€€€€€€€
3 I A E -2 I T1 s d 2 L 2 3 I A E 2 I T1 s d 2 L 2 3 I A a L2

k
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€ +
w3 2 w3 2 w3 w

3 I A2 a A z y
€€€€€ z
• •
z
3 I A E -2 I T1 s a L 2 3 I A E 2 I T1 s a L 2 I A m2 5 I A2 d2 A

{
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€ €€ + €€€€€€€€€€€€€€€€
2w 2w 2w 3 w3 2w

88.002 Second, Null<

à 4.3 The Generalized Method of Averaging

To apply either the method of averaging or the generalized method of averaging to FDuffingEq, we need first to use the
method of variation of parameters to transform it into a system of two first-order equations governing the amplitude and
phase. To this end, we introduce the following transformation:

states = 8a@tD, f@tD<;

transformRule = 8u -> a@tD Cos@f@tDD + L Cos@Y@tDD, u¢ -> - w a@tD Sin@f@tDD - W L Sin@Y@tDD<;

gRule = g -> -2 e m u¢ - e d u2 - e2 a u3 ;

After some algebraic manipulations (Nayfeh, 1973, 1981), we transform FDuffingEq into two first-order equations as

eq43a =

D@states, tD == i
j9- €€€€ Sin@f@tDD, w - €€€€€€€€€€€€€€€€ Cos@f@tDD= •. gRule •. transformRule •• Expand ••
k w
g g
a@tD w
y •• Thread;
TrigReduce •• Expandz
{

We seek approximate solutions to eq43a in the form

basicTerms = 8h@tD, j@tD, Y@tD<;

112 Chapter4.nb

solRule = 9a -> IEvaluateAh@tD + SumAej aj žž basicTerms, 8j, 2<E •. t -> #E &M,

f -> IEvaluateAj@tD + SumAej fj žž basicTerms, 8j, 2<E •. t -> #E &M=

8a ® Hh@#1D + e a1 @h@#1D, j@#1D, Y@#1DD + e2 a2@h@#1D, j@#1D, Y@#1DD &L,

f ® Hj@#1D + e f1 @h@#1D, j@#1D, Y@#1DD + e2 f2@h@#1D, j@#1D, Y@#1DD &L<

where h@tD and j@tD are expanded in power series in e as

basicDRule = D@basicTerms@@81, 2<DD, tD ->

9SumAei Ai žž basicTerms, 8i, 2<E, w + SumAei Fi žž basicTerms, 8i, 2<E= •• Thread

8h¢ @tD ® e A1@h@tD, j@tD, Y@tDD + e2 A2 @h@tD, j@tD, Y@tDD,

j¢ @tD ® w + e F1 @h@tD, j@tD, Y@tDD + e2 F2@h@tD, j@tD, Y@tDD<

The functions a 1 , a 2, f1 , and f2 are fast varying functions of j, while it follows from basicDRule that h, and hence the An
and Fn are slowly varying functions of t.

To the second approximation, we differentiate a@tD and f@tD once with respect to t, use solRule and basicDRule, expand the
result for small e, discard terms of order higher than e2, and obtain

eq43bLHS =
CoefficientListAExpand@#@@1DD •. solRule •. basicDRuleD •. e n_•;n>2 -> 0, eE & •ž eq43a;

Next, we substitute solRule into the right-hand sides of eq43a, expand the result for small e, keep terms up to OHe2L, and
rewrite their right-hand sides as

eq43bRHS =
CoefficientList@Series@#@@2DD •. solRule, 8e, 0, 2<D •• Normal •• Expand, eD & •ž eq43a;

Equating coefficients of like powers of e in eq43bLHS and eq43bRHS, we obtain

eqEps = MapThread@Equal, 8eq43bLHS, eq43bRHS<, 2D •• Transpose •• Rest;

Next, we use the method of separation of variables to separate fast and slowly varying terms in the first-order equations,
eqEps[[1]]. We first introduce the rules

SVT@j_D := 8Aj žž basicTerms, Fj žž basicTerms<

solVar@j_D := 8aj, fj<

psiRule = 8j@tD -> w t + b@tD, Y@tD -> W t<;

betaRule = Solve@psiRule •. Rule -> Equal, 8b@tD, W<D@@1DD

9b@tD ® -t w + j@tD, W ® €€€€€€€€€€€€ =

Y@tD
t

For the case of subharmonic resonance of OI €€2€€ M, we define

1

OmgRule = 8W -> 2 w + e s<;

Chapter4.nb 113

sigRule = Solve@OmgRule •. Rule -> Equal, sD@@1DD

9s ® - €€€€€€€€€€€€€€€ =
2w-W
e

expRule1 = f_@a_D :> f@Expand@a •. psiRule •. OmgRuleD •. e t -> t1 D;

expRule2 = f_@a_D :> f@Expand@a •. t1 -> e t •. sigRule •. betaRuleDD;

Using these rules, we find that the slowly varying parts of eqEps[[1]] are given by

SVT1Rule =
Table@Solve@eqEps@@1, iDD •. Thread@solVar@1D -> H0 &LD •. expRule1 •. f_@_ t + _.D -> 0 •.
expRule2, SVT@1D@@iDDD @@1DD, 8i, 2<D •• Flatten

9A1 @h@tD, j@tD, Y@tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

2 m w h@tD - d L Sin@2 j@tD - Y@tDD h@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2w

F1 @h@tD, j@tD, Y@tDD ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

€€€€€€€€€€€€€€€€€€€€ =
d L Cos@2 j@tD - Y@tDD
2w

whereas the fast varying parts are given by

FVT1 =
Table@Subtract žž eqEps@@1, iDD •. Thread@solVar@1D -> H0 &LD, 8i, 2<D •. SVT1Rule •• Expand

: €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
L m W Cos@j@tD - Y@tDD L m W Cos@j@tD + Y@tDD d L2 Sin@j@tDD
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w w 2w
2 2
d L Sin@j@tD - 2 Y@tDD d L Sin@j@tD + 2 Y@tDD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - m Cos@2 j@tDD h@tD -
4w 4w
d L Sin@2 j@tD + Y@tDD h@tD d Sin@j@tDD h@tD2 d Sin@3 j@tDD h@tD2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ ,
2w 4w 4w
d L Cos@Y@tDD d L Cos@2 j@tD + Y@tDD d L2 Cos@j@tDD
- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + m Sin@2 j@tDD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w 2w 2 w h@tD
d L2 Cos@j@tD - 2 Y@tDD d L2 Cos@j@tD + 2 Y@tDD L m W Sin@j@tD - Y@tDD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ +
4 w h@tD 4 w h@tD w h@tD

€€€€€€€€€€€€€ >
L m W Sin@j@tD + Y@tDD 3 d Cos@j@tDD h@tD d Cos@3 j@tDD h@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w h@tD 4w 4w

To determine a particular solution corresponding to these fast varying terms, we use the method of undetermined coeffi-
cients. To accomplish this, we first determine the possible forms of the terms in FVT1 as follows:
114 Chapter4.nb

FVT1Forms =
HCases@#, HCos È SinL@a_D -> 8Cos@aD, Sin@aD<, InfinityD •• Flatten •• UnionL & •ž FVT1

88Cos@j@tDD, Cos@2 j@tDD, Cos@3 j@tDD, Cos@j@tD - 2 Y@tDD,

Cos@j@tD - Y@tDD, Cos@j@tD + Y@tDD, Cos@2 j@tD + Y@tDD, Cos@j@tD + 2 Y@tDD,
Sin@j@tDD, Sin@2 j@tDD, Sin@3 j@tDD, Sin@j@tD - 2 Y@tDD, Sin@j@tD - Y@tDD,

8Cos@j@tDD, Cos@2 j@tDD, Cos@3 j@tDD, Cos@j@tD - 2 Y@tDD, Cos@j@tD - Y@tDD,

Sin@j@tD + Y@tDD, Sin@2 j@tD + Y@tDD, Sin@j@tD + 2 Y@tDD<,

Cos@Y@tDD, Cos@j@tD + Y@tDD, Cos@2 j@tD + Y@tDD, Cos@j@tD + 2 Y@tDD,

Sin@j@tDD, Sin@2 j@tDD, Sin@3 j@tDD, Sin@j@tD - 2 Y@tDD, Sin@j@tD - Y@tDD,
Sin@Y@tDD, Sin@j@tD + Y@tDD, Sin@2 j@tD + Y@tDD, Sin@j@tD + 2 Y@tDD<<

Using the principle of superposition, we seek a particular solution corresponding to the fast-varying terms as a linear combina-
tion of these possible forms:

sol1Form = MapIndexed@Hcoeffs1@#2@@1DDD = Array@c, Length@#1DDL.#1 &, FVT1FormsD

8c@1D Cos@j@tDD + c@2D Cos@2 j@tDD + c@3D Cos@3 j@tDD +

c@4D Cos@j@tD - 2 Y@tDD + c@5D Cos@j@tD - Y@tDD + c@6D Cos@j@tD + Y@tDD +
c@7D Cos@2 j@tD + Y@tDD + c@8D Cos@j@tD + 2 Y@tDD + c@9D Sin@j@tDD + c@10D Sin@2 j@tDD +
c@11D Sin@3 j@tDD + c@12D Sin@j@tD - 2 Y@tDD + c@13D Sin@j@tD - Y@tDD +
c@14D Sin@j@tD + Y@tDD + c@15D Sin@2 j@tD + Y@tDD + c@16D Sin@j@tD + 2 Y@tDD,
c@1D Cos@j@tDD + c@2D Cos@2 j@tDD + c@3D Cos@3 j@tDD + c@4D Cos@j@tD - 2 Y@tDD +
c@5D Cos@j@tD - Y@tDD + c@6D Cos@Y@tDD + c@7D Cos@j@tD + Y@tDD + c@8D Cos@2 j@tD + Y@tDD +
c@9D Cos@j@tD + 2 Y@tDD + c@10D Sin@j@tDD + c@11D Sin@2 j@tDD + c@12D Sin@3 j@tDD +
c@13D Sin@j@tD - 2 Y@tDD + c@14D Sin@j@tD - Y@tDD + c@15D Sin@Y@tDD +
c@16D Sin@j@tD + Y@tDD + c@17D Sin@2 j@tD + Y@tDD + c@18D Sin@j@tD + 2 Y@tDD<

Substituting sol1Form into eqEps[[1]], using SVT1Rule, collecting the coefficients of FVT1Forms, solving the resulting
algebraic equations for the undetermined coefficients, and then substituting the result back into sol1Form, we obtain the
solution
Chapter4.nb 115

sol1rhs = Table@sol1Form@@iDD •.
Solve@Coefficient@Subtract žž eqEps@@1, iDD •. SVT1Rule •. solVar@1D@@iDD -> HEvaluate@
sol1Form@@iDD •. Thread@basicTerms -> 8#1, #2, #3<DD &L, FVT1Forms@@iDDD ==
0 •• Thread, coeffs1@iDD@@1DD •. D@psiRule, tD •• Expand, 8i, 2<D

:- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
d L2 Cos@j@tDD d L2 Cos@j@tD - 2 Y@tDD d L2 Cos@j@tD + 2 Y@tDD
4 w Hw - 2 WL 4 w Hw + 2 WL
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ -
2w 2

w Hw - WL w Hw + WL 2 w H2 w + WL
L m W Sin@j@tD - Y@tDD L m W Sin@j@tD + Y@tDD d L Cos@2 j@tD + Y@tDD h@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

m Sin@2 j@tDD h@tD d Cos@j@tDD h@tD2 d Cos@3 j@tDD h@tD2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ ,
2w 4 w2 12 w2

2 w H2 w + WL w Hw - WL h@tD
m Cos@2 j@tDD d L Sin@Y@tDD d L Sin@2 j@tD + Y@tDD L m W Cos@j@tD - Y@tDD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ +
2w wW
d L2 Sin@j@tDD d L2 Sin@j@tD - 2 Y@tDD
w Hw + WL h@tD 4 w Hw - 2 WL h@tD
L m W Cos@j@tD + Y@tDD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ +
2 w2 h@tD

€€€€€€€€€€€€€ >
d L2 Sin@j@tD + 2 Y@tDD
4 w Hw + 2 WL h@tD
3 d Sin@j@tDD h@tD d Sin@3 j@tDD h@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4w 2 12 w2

which can be expressed in pure function form as

sol1Rule = Table@solVar@1D@@iDD ->

HEvaluate@sol1rhs@@iDD •. Thread@basicTerms -> 8#1, #2, #3<DD &L, 8i, 2<D

i d L2 Cos@#2D
:a1 ® j
j- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
d L2 Cos@#2 - 2 #3D
k 4 w Hw - 2 WL
€€€€€€€€€€€€ -
2 w2

4 w Hw + 2 WL w Hw - WL w Hw + WL
d L2 Cos@#2 + 2 #3D L m W Sin@#2 - #3D L m W Sin@#2 + #3D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ -

d Cos@3 #2D #12 y

€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ &zz,
d Cos@#2D #12
2 w H2 w + WL
d L Cos@2 #2 + #3D #1 m Sin@2 #2D #1
{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

i
2w 4 w2 12 w2

f1 ® j
j €€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w H2 w + WL w Hw - WL #1 w Hw + WL #1
m d L d L + L m W - L m W Cos@#2 + #3D
k
Cos@2 #2D Sin@#3D Sin@2 #2 #3D Cos@#2 #3D
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ +
2w wW
d Sin@3 #2D #1 y
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ &z
z>
d L2 Sin@#2D d L2 Sin@#2 - 2 #3D d L2 Sin@#2 + 2 #3D
4 w Hw - 2 WL #1 4 w Hw + 2 WL #1 {
3 d Sin@#2D #1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w #12 4w 2 12 w2

Substituting the slow- and fast-varying components of the first-order solution, SVT1Rule and sol1Rule, into the second-
order equations, we have

Horder2Expr = Table@TrigReduce •ž
HSubtract žž eqEps@@2, iDD •. SVT1Rule •. sol1Rule •• ExpandL, 8i, 2<D;L •• Timing

812.36 Second, Null<

Since we are seeking an expansion valid up to OHe2L, we do not need to solve for a 2 and f2 . All we need to do is to investi-
gate the above expressions to determine the slowly varying parts and determine A2 and F2. The result is
116 Chapter4.nb

SVT2Rule =
Table@Solve@order2Expr@@iDD == 0 •. Thread@solVar@2D -> H0 &LD •. expRule1 •. f_@_ t + _.D ->
0 •. expRule2 •. HOmgRule •. e -> 0L, SVT@2D@@iDDD, 8i, 2<D •• Flatten •• ExpandAll

:A2 @h@tD, j@tD, Y@tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , F2 @h@tD, j@tD, Y@tDD ®

7 d L m Cos@2 j@tD - Y@tDD h@tD
6 w2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€ >

9 d2 L2 3 a L2 m2 7 d L m Sin@2 j@tD - Y@tDD 5 d2 h@tD2 3 a h@tD2
- €€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€ - €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
16 w3 4w 2w 6 w2 12 w3 8w

Hence, to the second approximation, we find that

u@tD ==
HTrigReduce •ž HNormal@Series@u •. transformRule •. solRule, 8e, 0, 1<DD •. sol1Rule ••
ExpandL •• Collect@#, eD &L

i d L2
u@tD ==

L Cos@Y@tDD + Cos@j@tDD h@tD + e j j- €€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

d L2 Cos@2 Y@tDD d L2 Cos@2 Y@tDD
4 w Hw - 2 WL 4 w Hw + 2 WL w Hw - WL
L m W Sin@Y@tDD
k 2w 2

w Hw + WL
L m W Sin@Y@tDD d L Cos@j@tD - Y@tDD h@tD d L Cos@j@tD + Y@tDD h@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2w W 2wW
d Cos@2 j@tDD h@tD2 z y
€€€€€€€€€€€€€€€ z
d h@tD2
2 w H2 w + WL
d L Cos@j@tD + Y@tDD h@tD m Sin@j@tDD h@tD
{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2w 2 w2 6 w2

where the equations governing the amplitude and phase are

8ampEqs, phaseEqs< = HbasicDRule •. SVT1Rule •. SVT2Rule •• ExpandAllL •.

f_@a_D :> f@Collect@a •. psiRule, tDD •. Rule -> Equal

7 d e2 L m Cos@t H2 w - WL + 2 b@tDD h@tD d e L Sin@t H2 w - WL + 2 b@tDD h@tD

:h¢ @tD == -e m h@tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ ,

d e L Cos@t H2 w - WL + 2 b@tDD
6 w2 2w

j¢ @tD == - €€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€ + w + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

9 d2 e2 L2 3 a e2 L2 e2 m2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

7 d e2 L m Sin@t H2 w - WL + 2 b@tDD
16 w3 4w 2w 2w

€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

5 d2 e2 h@tD2 3 a e2 h@tD2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
6 w2 12 w3 8w

à 4.4 The Method of Normal Forms

In this section, we use the method of normal forms to determine a second-order uniform expansion of the solution of FDuff-
ingEq. To this end, we start with the corresponding first-order complex-valued equation, which was derived in Section 4.2.6
and a summary of the derivation is given below:

transformRule = 9u@tD -> Iz@tD + z@tDM, u¢ @tD -> I w Iz@tD - z@tDM, f Cos@W tD -> €€€€ Hz@tD + •
z@tDL=;
• • 1
2

gRule = g -> -2 e m u¢@tD - e d u@tD2 - e2 a u@tD3 + f Cos@W tD;

Chapter4.nb 117

eq44a = 8z¢ @tD == I w z@tD - I g • H2 wL •. gRule •. transformRule •• ExpandAll, z¢ @tD == I W z@tD<

:z¢ @tD ==

I z@tD I d e z@tD z@tD

- €€€€€€€€€€€€€€€€ - e m z@tD + I w z@tD + €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ + e m z@tD + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
• •
I z@tD I d e z@tD2 I a e2 z@tD3 •
4w 2w 2w 4w w
3 I a e z@tD z@tD I d e z@tD 3 I a e z@tD z@tD I a e z@tD
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ , z¢ @tD == I W z@tD>
2 2• • 2 2 • 2 2• 3
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2w 2w 2w 2w

According to the method of normal forms, we introduce the near-identity transformation

basicTerms = 8h@tD, •
h@tD, z@tD, •
z@tD<;

zetaRule = 9z -> FunctionAt, h@tD + SumAej hj@Sequence žž basicTermsD, 8j, 0, 2<E •• EvaluateE,
h@tD + SumAej hj@Sequence žž basicTermsD, 8j, 0, 2<E •• EvaluateE=
• •
z -> FunctionAt, •

8z ® Function@t, h@tD + h0 @h@tD, h@tD, z@tD, z@tDD +

• •

e h1 @h@tD, h@tD, z@tD, z@tDD + e2 h2@h@tD, h@tD, z@tD, z@tDDD,

• • • •

z ® Function@t, h@tD + h0 @h@tD, h@tD, z@tD, z@tDD +

• • • • •

e h1 @h@tD, h@tD, z@tD, z@tDD + e2 h2@h@tD, h@tD, z@tD, z@tDDD<

• • • • • •

that results in the simplest possible equation

etaRule = 9h¢ @tD -> I w h@tD + SumAej gj @tD, 8j, 2<E, z¢ @tD -> I W z@tD=

8h¢ @tD ® I w h@tD + e g1@tD + e2 g2 @tD, z¢ @tD ® I W z@tD<

where the overbar denotes the complex conjugate

conjugateRule = 8h -> •
h, •
h -> h, g -> •
g, •
g -> g, z -> •
z, •
z -> z, Complex@0, n_D -> Complex@0, -nD<;

Substituting the expansion for z, the zetaRule, into eq44a, using the etaRule, expanding the result for small e, and discarding
terms of order higher than e2, we have

eq44b = Heq44a@@1DD •. zetaRule •. etaRule •. HetaRule •. conjugateRuleL •• ExpandAllL •.

e n_•;n>2 -> 0;

Equating coefficients of like powers of e in eq44b yields

eqEps = CoefficientList@Subtract žž eq44b, eD == 0 •• Thread;

Ÿ First-Order Solution

Using the method of undetermined coefficients, we assume

coeffsL = Table@Di, 8i, Length@basicTermsD<D

8D1 , D2, D3, D4 <

118 Chapter4.nb

hFormL = 9h0 -> [email protected] •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
h0 -> [email protected] •. conjugateRule •.
•

Thread@basicTerms -> 8#1, #2, #3, #4<DD &L=;

Substituting the hFormL into eqEps[[1]], collecting the coefficients of the basicTerms, and solving for the coeffsL, we obtain

coeffsLRule =
Solve@Coefficient@eqEps@@1, 1DD •. hFormL, basicTermsD == 0 •• Thread, coeffsLD@@1DD

Solve::svars : Equations may not give solutions for all "solve" variables.

:D2 ® 0, D3 ® €€€€€€€€€€€€€€€€€€€€€€€€ , D4 ® €€€€€€€€€€€€€€€€€€€€€€€€ >

4 w Hw - WL 4 w Hw + WL
1 1

Substituting the coeffsLRule into the hFormL yields

hSolL = h0 žž basicTerms •. hFormL •. coeffsLRule •. Di_ -> 0

z@tD
•

4 w Hw - WL 4 w Hw + WL
z@tD
€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€

Then, we write the first-order solution in pure function form as

hRuleL = 9h0 -> HEvaluate@hSolL •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
h0 -> HEvaluate@hSolL •. conjugateRule •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L=
•

i y • i y
:h0 ® j €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ &z, h0 ® j €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ &z>
k 4 w Hw - WL 4 w Hw + WL { k 4 w Hw + WL 4 w Hw - WL {
#3 #4 #3 #4

Ÿ Second-Order Solution

Substituting the hRuleL into the left-hand side of the second-order equation, eqEps[[2]], yields

order2expr = eqEps@@2, 1DD •. hRuleL •• Expand

I d z@tD2 I d z@tD2 I d z@tD2

4 w Hw - WL 4 w Hw + WL 16 w3 Hw - WL Hw + WL
m z@tD m z@tD
32 w3 Hw - WL2 32 w3 Hw + WL2
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + m h@tD -

m z@tD m z@tD I d z@tD z@tD

• • •

4 w2 Hw - WL 4 w2 Hw + WL 4 w Hw - WL 4 w Hw + WL
I d z@tD h@tD I d z@tD h@tD I d h@tD2
16 w3 Hw - WL2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2w

I d z@tD z@tD I d z@tD z@tD I d h@tD z@tD I d h@tD z@tD I d z@tD2

• • • • •

16 w3 Hw + WL2 8 w3 Hw - WL Hw + WL 4 w2 Hw - WL 4 w2 Hw + WL 32 w3 Hw - WL2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

I d z@tD2 I d z@tD2 I d z@tD h@tD I d z@tD h@tD I d h@tD h@tD

€€€€€€€€€€€€€€ - m h@tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
• • • • •

32 w3 Hw + WL2 16 w3 Hw - WL Hw + WL 4 w2 Hw - WL 4 w2 Hw + WL
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w

I d z@tD h@tD I d z@tD h@tD I d h@tD2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€ + g1 @tD - I w h1@h@tD, h@tD, z@tD, z@tDD -
• • • • •

4 w2 Hw - WL 4 w2 Hw + WL
• •
2w
IW• z@tD hH0,0,0,1L @h@tD, • h@tD, z@tD, • z@tDD + I W z@tD hH0,0,1,0L @h@tD, • h@tD, z@tD, • z@tDD -
I w h@tD h1 @h@tD, h@tD, z@tD, z@tDD + I w h@tD h1 @h@tD, h@tD, z@tD, z@tDD
H0,1,0,0L H1,0,0,0L
1 1
• • • • •
Chapter4.nb 119

We choose h 1 to eliminate as many terms from order2expr, thereby reducing it to its simplest possible form. It turns out that
we can eliminate all nonresonance terms. To determine the resonance terms in order2expr, we first determine all of the
possible forms of its nonhomogeneous terms:

possibleQTerms =
basicTerms~ Join~HOuter@Times, basicTerms, basicTermsD •• Flatten •• UnionL

9h@tD, h@tD, z@tD, z@tD, z@tD2 , z@tD h@tD, h@tD2,

• •

z@tD z@tD, h@tD z@tD, z@tD2, z@tD h@tD, h@tD h@tD, z@tD h@tD, h@tD2 =
• • • • • • • •

It follows from the linear parts of the etaRule that

form = 9h@tD -> EI w t, •

h@tD -> E-I w t , z@tD -> EIW t , •
z@tD -> E-I W t=;

Hence, the possible resonance terms are given by

ResonantQTerm =
IE-I w t possibleQTerms •. form •. W -> 2 w •. E _ t -> 0M possibleQTerms •• Union •• Rest

8h@tD, z@tD h@tD<

•

We choose g 1 to eliminate the resonance terms in order2expr according to

gRuleQ =
g1 @tD -> -HCoefficient@order2expr •. h1 -> H0 &L, ResonantQTermD •. Thread@basicTerms -> 0DL.
ResonantQTerm

i y
g1 @tD ® -m h@tD - j
j- €€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€ z
z z@tD h@tD
k 4 w2 Hw - WL 4 w2 Hw + WL {
Id Id •

The nonresonance terms are the complement of the resonance terms; that is,

NRQT = Complement@possibleQTerms, ResonantQTermD;

We associate with each of them an undetermined coefficient according to

coeffsQ = Table@Gj, 8j, Length@NRQTD<D

8G1 , G2, G3, G4 , G5 , G6, G7 , G8 , G9, G10, G11 , G12<

Therefore, h 1 has the form

hFormQ = 8h1 -> [email protected] •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L<

8h1 ® H#3 G1 + #32 G2 + #1 #3 G3 + #12 G4 + #4 G5 +

#3 #4 G6 + #1 #4 G7 + #42 G8 + #2 G9 + #1 #2 G10 + #2 #4 G11 + #22 G12 &L<

Substituting for h 1 in order2expr, equating the coefficients of the possible nonresonance terms to zero, and solving the
resulting equations for the Gi, we obtain
120 Chapter4.nb

coeffsQRule = Solve@Coefficient@order2expr •. gRuleQ •. hFormQ •• Expand, NRQTD == 0 •.

Thread@basicTerms -> 0D •• Thread, coeffsQD@@1DD

9G3 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w W H-w + WL Hw + WL 2 w W H-w + WL Hw + WL
d d d d
4 w2 Hw - WL2 Hw + WL2
€€€€€€€€€€€€€€€€ , G4 ® €€€€€€€€€€ , G6 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ , G7 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ ,
2 w2
Im d d ImW d
2 w Hw - WL2 Hw + WL 8 w Hw - 2 WL Hw - WL2 Hw + WL2
G9 ® €€€€€€€€ , G10 ® - €€€€€€ , G12 ® - €€€€€€€€€€ , G1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€ , G2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2w w2 6 w2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ =
2 w Hw - WL Hw + WL H2 w + WL
Im W d d
2 w Hw - WL Hw + WL2 8 w Hw - WL2 Hw + WL2 Hw + 2 WL
G5 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , G11 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , G8 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

Substituting the coeffsQRule into the hFormQ yields

hSolQ = h1 žž basicTerms •. hFormQ •. coeffsQRule

d z@tD2
2 w W H-w + WL Hw + WL
I m W z@tD d z@tD h@tD
2 w Hw - WL Hw + WL 8 w Hw - 2 WL Hw - WL Hw + WL
- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ +
2 2 2

I m W z@tD d z@tD z@tD d h@tD z@tD

• • •
d h@tD2
2 w Hw - WL Hw + WL2 4 w2 Hw - WL2 Hw + WL2 2 w W H-w + WL Hw + WL
€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ -
2 w2

d z@tD2 I m h@tD d h@tD h@tD d z@tD h@tD d h@tD2

• • • • • •

8 w Hw - WL2 Hw + WL2 Hw + 2 WL 2 w Hw - WL Hw + WL H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€
2w w2 6 w2

Hence, we can write the second-order solution in pure function form as

hRuleQ = 9h1 -> HEvaluate@hSolQ •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
h1 -> HEvaluate@hSolQ •. conjugateRule •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L=;
•

Ÿ Third-Order Equations

Substituting the hRuleL, hRuleQ, and gRuleQ into the left-hand side of the third-order equation, eqEps[[3]], yields

order3expr =
eqEps@@3, 1DD •. hRuleL •. hRuleQ •. gRuleQ •. HgRuleQ •. conjugateRuleL •• Expand;

The nonhomogeneous terms in order3expr are proportional to

possibleCTerms = Outer@Times, possibleQTerms, basicTermsD •• Flatten •• Union

9z@tD2 , z@tD3, z@tD h@tD, z@tD2 h@tD, h@tD2, z@tD h@tD2 , h@tD3, z@tD z@tD, z@tD2 z@tD,
• •

h@tD z@tD, z@tD h@tD z@tD, h@tD2 z@tD, z@tD2 , z@tD z@tD2, h@tD z@tD2 , z@tD3, z@tD h@tD,
• • • • • • • •

z@tD2 h@tD, h@tD h@tD, z@tD h@tD h@tD, h@tD2 h@tD, z@tD h@tD, z@tD z@tD h@tD,
• • • • • • • •

h@tD z@tD h@tD, z@tD2 h@tD, h@tD2 , z@tD h@tD2, h@tD h@tD2 , z@tD h@tD2, h@tD3 =
• • • • • • • • • •
Chapter4.nb 121

possibleTerms = Join@possibleQTerms, possibleCTermsD •• Union

9z@tD, z@tD2 , z@tD3, h@tD, z@tD h@tD, z@tD2 h@tD, h@tD2, z@tD h@tD2 , h@tD3, z@tD, z@tD z@tD,
• •

z@tD2 z@tD, h@tD z@tD, z@tD h@tD z@tD, h@tD2 z@tD, z@tD2, z@tD z@tD2 , h@tD z@tD2,
• • • • • • •

z@tD3 , h@tD, z@tD h@tD, z@tD2 h@tD, h@tD h@tD, z@tD h@tD h@tD, h@tD2 h@tD, z@tD h@tD,
• • • • • • • • •

z@tD z@tD h@tD, h@tD z@tD h@tD, z@tD h@tD, h@tD , z@tD h@tD , h@tD h@tD , z@tD h@tD2 , h@tD3 =
• • • • • 2• • 2 • 2 • 2 • • •

Next, we determine the resonance terms according to

ResonantCTerm =
IE-I w t possibleTerms •. form •. W -> 2 w •. E _ t -> 0M possibleTerms •• Union •• Rest

8h@tD, z@tD h@tD z@tD, z@tD h@tD, h@tD2 h@tD<

• • •

Choosing g 2 to eliminate the resonance terms from order3expr yields

gRuleC = g2 @tD ->

- Factor@Coefficient@order3expr •. h2 -> H0 &L, ResonantCTermD •. Thread@basicTerms -> 0DD.
ResonantCTerm

I H-5 d2 w + 6 a w3 - 2 d2 W + 3 a w2 WL z@tD h@tD z@tD

g2 @tD ® - €€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
•
I m2 h@tD
4 w3 Hw - WL2 Hw + WL2 H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ -
2w
d m H2 w3 + w2 W - 6 w W2 - W3L z@tD h@tD I H10 d2 - 9 a w2 L h@tD2 h@tD
• •

4 w2 Hw - WL2 W Hw + WL2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
6 w3

The nonresonance terms are the complement of the resonance terms; that is,

NRCT = Complement@possibleTerms, ResonantCTermD;

Again, we associate an undetermined coefficient with each of these terms as follows:

coeffsC = Table@Lj, 8j, Length@NRCTD<D

8L1 , L2, L3, L4 , L5 , L6, L7 , L8 , L9, L10, L11 , L12, L13 , L14 , L15,
L16 , L17, L18 , L19 , L20, L21 , L22, L23, L24 , L25, L26 , L27 , L28, L29 , L30<

Consequently, h 2 has the form

hFormC = 8h2 -> [email protected] •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L<;

Substituting for h 2 into order3expr, equating the coefficient of each possible nonresonance term to zero, and solving the
resulting algebraic equations for the Li, we obtain

coeffsCRule = Solve@Coefficient@order3expr •. gRuleC •. hFormC •• Expand, NRCTD == 0 •.

Thread@basicTerms -> 0D •• Thread, coeffsCD@@1DD •• ExpandAll;

Substituting the coeffsCRule into the hFormC yields

hSolC = h2 žž basicTerms •. hFormC •. coeffsCRule;

122 Chapter4.nb

Combining the etaRule, gRuleQ, and gRuleC and letting

solForm = 9h -> IA@#D EI w # &M, •

h -> IA@#D E-I w # &M, z -> If EI W # &M, •
z -> If E-I W # &M=;
•

we obtain the modulation equation

moduEq = I2 I w E-I wt Hh¢ @tD - Hh¢ @tD •. etaRule •. gRuleQ •. gRuleCLL •. solForm •• Expand ••
Collect@#, eD &M == 0

i E-2 I tw+I t W f d A@tD E-2 I t w+It W f d A@tD z y

ej
j €€€€ z
• •
j z+
k Hw Hw {
2 I m w A@tD + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
2 w - WL 2 w + WL
i 2
e2 j
j
j-m A@tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
5 f2 d2 A@tD 3 f2 a w A@tD

k 2 w Hw - WL2 Hw + WL2 H2 w + WL Hw - WL2 Hw + WL2 H2 w + WL

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ +

3 f2 a W A@tD f2 d2 W A@tD
2 Hw - WL2 Hw + WL2 H2 w + WL w2 Hw - WL2 Hw + WL2 H2 w + WL
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

I E-2I t w+I t W f d m w A@tD I E-2I t w+I t W f d m w2 A@tD 3 I E-2 I t w+It W f d m W A@tD

• • •

2 Hw - WL2 Hw + WL2 Hw - WL2 W Hw + WL2 Hw - WL2 Hw + WL2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

I E-2I t w+I t W f d m W2 A@tD 10 d2 A@tD2 A@tD y

€€€€€€€€€€€€€€€€€€€€€€€€€ + 3 a A@tD2 A@tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z z
z + 2 I w A @tD == 0
• •
•
2 w Hw - WL Hw + WL {
¢
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 2 3 w2

or

moduEq •. f -> 2 L IW2 - w2M •. W -> 2 w

e H2 I m w A@tD - 2 d L A@tDL +
•

i 14 I d L m A@tD 10 d2 A@tD2 A@tD z y

e2 j
j @tD €€€€€€ z
• •
j z+
9 d2 L2 A@tD 2•

k {
2 2
6 a L A@tD - m A@tD - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€
€ - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€€€€
€ + 3 a A@tD A - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
2 w 2 3 w 3 w 2

2 I w A @tD == 0
¢

which is in agreement with those obtained by using the method of multiple scales and the generalized method of averaging.
Chapter 5
Higher-Order Approximations for Systems
with Internal Resonances
Off@General::spell1D

à 5.1 Euler-Lagrange Equations

In this chapter, we use different methods to determine approximate solutions of nonlinear systems possessing internal
resonances to orders higher than the order at which the influence of the internal resonance first appears. To describe the
methods with minimum algebra, we consider the free oscillations of a two-degree-of-freedom conservative system possess -
ing a two-to-one internal resonance and having simple quadratic nonlinearities. In particular, we consider a system governed
by the Lagrangian

Lagrangian = €€€€ u1 ¢@tD2 + €€€€ u2 ¢ @tD2 - €€€€ w21 u1 @tD2 - €€€€ w22 u2@tD2 + d u1 @tD2 u2 @tD;
1 1 1 1
2 2 2 2

where w2 » 2 w1 . Writing down the Euler-Lagrange equations corresponding to the Lagrangian, we obtain the following
second-order equations of motion:

EOM1 = Table@D@D@Lagrangian, ui¢ @tDD, tD - D@Lagrangian, ui @tDD == 0, 8i, 2<D

8w21 u1@tD - 2 d u1@tD u2 @tD + u²1 @tD == 0, -d u1 @tD2 + w22 u2 @tD + u²2 @tD == 0<

Approximate solutions of general two-degree-of-freedom system with quadratic and cubic nonlinearities having a two-to-one
internal resonance can be obtained by replacing the Lagrangian with the Lagrangian corresponding to the general system.

As discussed in the next section, treating this second-order form, EOM1, of the governing equations by using the method of
multiple scales may lead to results that violate the conservative nature of the system under consideration. To determine an
approximate solution that preserves the conservative nature of the system, we treat a first-order form of the governing

more states v1@tD and v2 @tD such that

equations. To transform EOM1 into a system of first-order equations, we form a modified Lagrangian by introducing two

velRule = 8u1 ¢@tD -> v1 @tD, u2¢ @tD -> v2@tD<;

Then, the Lagrangian becomes

Lag1 = Lagrangian •. velRule

- €€€€ w21 u1@tD2 + d u1 @tD2 u2 @tD - €€€€ w22 u2@tD2 + €€€€ v1@tD2 + €€€€ v2 @tD2
1 1 1 1
2 2 2 2
120 Chapter5.nb

Substituting Lag1 into Euler-Lagrange equations, we obtain the following first-order equations

eq51a = Table@D@D@Lag1, vi@tDD, tD - D@Lag1, ui@tDD == 0, 8i, 2<D

8w21 u1@tD - 2 d u1@tD u2 @tD + v¢1 @tD == 0, -d u1 @tD2 + w22 u2 @tD + v¢2 @tD == 0<

Using the velRule and eq51a, we obtain the following first-order equations of motion:

EOM2 = 8velRule •. Rule -> Equal, eq51a< •• Transpose •• Flatten

8u¢1 @tD == v1@tD, w21 u1@tD - 2 d u1@tD u2 @tD + v¢1 @tD == 0,

u¢2 @tD == v2@tD, -d u1@tD2 + w22 u2 @tD + v¢2 @tD == 0<

The corresponding linear system of EOM2 can be defined as x = A x, where x is the state vector; that is,

states = Table@8ui@tD, vi@tD<, 8i, 2<D •• Flatten

8u1 @tD, v1@tD, u2 @tD, v2 @tD<

Using these states, we find that the coefficient matrix A of the linear part of EOM2 is given by

matrixA = Outer@D, #@@2DD & •ž Solve@EOM2, D@states, tDD@@1DD •. d -> 0, statesD

880, 1, 0, 0<, 8-w21 , 0, 0, 0<, 80, 0, 0, 1<, 80, 0, -w22, 0<<

whose adjoint is defined by

conjugateRule = 9A -> A, A -> A, Complex@0, n_D -> Complex@0, -nD=;

• •

hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose

The eigenvalues and eigenvectors of matrixA can be obtained as

matrixA •• Eigensystem

:8-I w1, I w1 , -I w2, I w2 <, :: €€€€€€ , 1, 0, 0>, :- €€€€€€ , 1, 0, 0>, :0, 0, €€€€€€ , 1>, :0, 0, - €€€€€€ , 1>>>
I I I I
w1 w1 w2 w2

Hence, the right eigenvectors of matrixA corresponding to the eigenvalues I w1 and I w2, respectively, are

rightVec = Eigenvectors@matrixAD@@82, 4<DD

::- €€€€€€ , 1, 0, 0>, :0, 0, - €€€€€€ , 1>>

I I
w1 w2

Moreover, the left eigenvectors of matrixA corresponding to the eigenvalues I w1 and I w2 , respectively, are

leftVec = Eigenvectors@hermitian@matrixADD@@81, 3<DD

88-I w1, 1, 0, 0<, 80, 0, -I w2, 1<<

whose complex conjugate is

Chapter5.nb 121

ccleftVec = leftVec •. conjugateRule

88I w1 , 1, 0, 0<, 80, 0, I w2 , 1<<

à 5.2 Method of Multiple Scales

We use the method of multiple scales to obtain a second-order uniform expansion of the solution of EOM1 in Section 5.2.1,
EOM2 in Section 5.2.2, and the corresponding first-order complex-valued equations of EOM1 in Section 5.2.3. Again, we
need the three time scales T0 = t, T1 = e t, and T2 = e2 t, which we symbolize by

Needs@"Utilities`Notation`"D

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

timeScales = 8T0 , T1, T2 <;

In terms of the new time scales, we can express the time derivatives as

dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, 2<E;

dt@2D@expr_D := Hdt@1D@dt@1D@exprDD •• ExpandL •. ei_•;i>3 -> 0;

To represent some of the expressions in a more concise way, we introduce the following display rule:

displayRule =
9Derivative@a__DAu_i_,j_ E@__D :> SequenceFormATimes žž MapIndexedAD #1
#2@@1DD-1 &, 8a<E, ui,jE,

Derivative@a__DAA_i_E@__D :> SequenceFormATimes žž MapIndexedAD #1

#2@@1DD &, 8a<E, Ai E,

u_i_,j_@__D -> ui,j , A_i_@__D -> Ai =;

Ÿ 5.2.1 Second-Order Real-Valued System

Using method of multiple scales, we assume that the solution of EOM1 can be expressed in the form

solRule = ui_ -> ISumAej ui,j@#1, #2, #3D, 8j, 3<E &M;

multiScales = 8ui_@tD -> ui @T0, T1, T2 D, Derivative@n_D@ui_ D@tD :> dt@nD@ui@T0 , T1, T2DD<;

Substituting multiScales and solRule into EOM1, expanding the result for small e, and discarding terms of order higher than
e3 , we obtain
122 Chapter5.nb

eq521a = HEOM1 •. multiScales •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

eq521a •. displayRule

8e HD20 u1,1 L + e2 HD20 u1,2 L + e3 HD20 u1,3 L + 2 e2 HD0 D1 u1,1L + 2 e3 HD0 D1u1,2 L + e3 HD21u1,1 L + 2 e3 HD0 D2 u1,1 L +
e w21 u1,1 + e2 w21 u1,2 + e3 w21 u1,3 - 2 d e2 u1,1 u2,1 - 2 d e3 u1,2 u2,1 - 2 d e3 u1,1 u2,2 == 0,
e HD20 u2,1 L + e2 HD20 u2,2 L + e3 HD20 u2,3 L + 2 e2 HD0 D1 u2,1L + 2 e3 HD0 D1u2,2 L + e3 HD21u2,1 L +
2 e3 HD0 D2 u2,1 L - d e2 u21,1 - 2 d e3 u1,1 u1,2 + e w22 u2,1 + e2 w22 u2,2 + e3 w22 u2,3 == 0<

Equating coefficients of like powers of e, we obtain

eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eq521a •• Transpose;

To place the linear operators on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := I#@@1DD & •ž eqEps@@1DD •. u_ k_,1 -> u k,i M ==

I#@@1DD & •ž eqEps@@1DD •. u_ k_,1 -> u k,iM - H#@@1DD & •ž eqEps@@iDDL •• Thread

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

eqOrder@1D •. displayRule •• TableForm

eqOrder@2D •. displayRule •• TableForm
eqOrder@3D •. displayRule •• TableForm

D20 u1,1 + w21 u1,1 == 0

D20 u2,1 + w22 u2,1 == 0

D20 u1,2 + w21 u1,2 == -2 HD0 D1u1,1 L + 2 d u1,1 u2,1

D20 u2,2 + w22 u2,2 == -2 HD0 D1u2,1 L + d u21,1

D20 u1,3 + w21 u1,3 == -2 HD0 D1u1,2 L - D21 u1,1 - 2 HD0 D2 u1,1 L + 2 d u1,2 u2,1 + 2 d u1,1 u2,2
D20 u2,3 + w22 u2,3 == -2 HD0 D1u2,2 L - D21 u2,1 - 2 HD0 D2 u2,1 L + 2 d u1,1 u1,2

Ÿ First-Order Equations: Linear System

To obtain the solution of eqOrder[1] by using DSolve, we transform eqOrder[1] into a set of ordinary-differential equations
as

order1Eq = eqOrder@1D •. ui_,j_ -> Hui,j@#1D &L

8w21 u1,1@T0 D + u²1,1 @T0 D == 0, w22 u2,1 @T0 D + u²2,1 @T0 D == 0<

The particular solution of order1Eq can be expressed as

sol1p = DSolve@order1Eq, 8u1,1@T0 D, u2,1@T0 D<, T0D@@1DD •. C@_D -> 0

8u1,1 @T0D ® 0, u2,1 @T0D ® 0<

Hence, we write the first-order solution in function form as

Chapter5.nb 123

sol1 = TableAui,1 -> FunctionA8T0 , T1, T2 <,

Ai @T1, T2 D Exp@I wi T0 D + Ai @T1, T2 D Exp@-I wi T0D + ui,1@T0 D •. sol1p •• EvaluateE, 8i, 2<E
•

8u1,1 ® Function@8T0 , T1 , T2<, EI T0 w1 A1@T1 , T2 D + E-I T0 w1 A1@T1 , T2DD,

•

u2,1 ® Function@8T0 , T1 , T2<, EI T0 w2 A2@T1 , T2 D + E-I T0 w2 A2@T1 , T2DD<

•

Ÿ Second-Order Equations

Substituting sol1 into eqOrder[2] yields

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

:D20 u1,2 + w21 u1,2 == 2 EIT0 w1+I T0 w2 d A1 A2 - 2 I EI T0 w1 HD1A1 L w1 +

2 I E-I T0 w1 HD1 A1L w1 + 2 E-I T0 w1+I T0 w2 d A2 A1 + 2 EI T0 w1 -I T0 w2 d A1 A2 + 2 E-I T0 w1 -IT0 w2 d A1 A2 ,
• • • • •

D20 u2,2 + w22 u2,2 == E2I T0 w1 d A21 - 2 I EI T0 w2 HD1 A2L w2 + 2 I E-I T0 w2 HD1A2 L w2 + 2 d A1 A1 + E-2 I T0 w1 d A1 >
• • •2

To describe quantitatively the nearness of the two-to-one internal resonance w2 » 2 w1 , we introduce the detuning parameter
s defined by

ResonanceCond = 8w2 == 2 w1 + e s<;

In eliminating secular terms, we need to express w2 in terms of w1 in some of the equations and w1 in terms of w2 in the other
equations. To accomplish this, we let

omgList = Table@wi, 8i, 2<D;

omgRule = Solve@ResonanceCond, Drop@omgList, 8#<DD@@1DD & •ž 81, 2<

98w2 ® e s + 2 w1 <, 9w1 ® €€€€ H-e s + w2L==

1
2

To convert small-divisor terms into secular terms, we define the rule

expRule1@j_D := Exp@arg_D :> Exp@Expand@arg •. omgRule@@jDDD •. e T0 -> T1 D

Substituting expRule1[i] into the right-hand sides of order2Eq and collecting the coefficients of EI wi T 0 , we obtain the
solvability conditions as

SCond1 = Coefficient@order2Eq@@#, 2DD •. expRule1@#D, Exp@I w # T0DD == 0 & •ž 81, 2<

92 EI T1 s d A2@T1 , T2 D A1@T1 , T2D - 2 I w1 A1 @T1 , T2 D == 0,

• H1,0L

E-I T1 s d A1@T1 , T2 D2 - 2 I w2 A2 @T1 , T2 D == 0=

H1,0L

Because the system is conservative, the modulation equations given by SCond1 must be derivable from a Lagrangian. These
equations are the Euler-Lagrange equations corresponding to the Lagrangian
124 Chapter5.nb

Lag = EI T1 s d A2 @T1, T2 D A1@T1 , T2 D + E-I T1 s d A1 @T1, T2D2 A2@T1 , T2D -

• 2 •
H1,0L H1,0L
I w1 A1@T1 , T2D A1 @T1, T2 D - I w2 A2@T1 , T2 D A2 @T1, T2 D +
• •
•H1,0L •H1,0L
I w1 A1@T1 , T2D A1 @T1 , T2 D + I w2 A2@T1 , T2D A2 @T1, T2 D;

because

•H1,0L
@T1, T2 DE, T1 E + DALag, A k @T1, T2 DE == 0, 8k, 2<E
•
alSCond1 = TableA- DADALag, Ak

92 EI T1 s d A2@T1 , T2 D A1@T1 , T2D - 2 I w1 A1 @T1 , T2 D == 0,

• H1,0L

E-I T1 s d A1@T1 , T2 D2 - 2 I w2 A2 @T1 , T2 D == 0=

H1,0L

Equations alSCond1 are identical to equations SCond1 as shown below

alSCond1 === SCond1

True

Alternatively, we can determine the modulation equations by using the time-averaged Lagrangian. To this end, we first
define the rule

expRule2 = Exp@a_D :> Exp@Expand@a •. HResonanceCond •. Equal -> RuleLD •. e T0 -> T1 D;

Substituting multiScales , solRule, sol1, and expRule2 into the Lagrangian and selecting the slow-varying terms, we obtain
the following time-averaged Lagrangian:

TAL1 =
SelectAIHLagrangian •. multiScales •. solRule •• ExpandL •. e n_•;n>3 -> 0 •. sol1 •• ExpandM •.
expRule2, FreeQ@#, T0 D &E •. e -> 1;
TAL1 •. displayRule

I HD1 A1L A1 w1 + I HD1A2 L A2 w2 - I HD1 A1L w1 A1 + EI T1 s d A2 A1 + E-I T1 s d A21 A2 - I HD1 A2 L w2 A2

• • • •2 • •

One can easily verify that TAL1 is identical to Lag. In fact,

TAL1 == Lag

True

Next, we rewrite SCond1 as

H1,0L H1,0L
SCond1Rule1 = SolveASCond1, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD

I EI T1 s d A2@T1 , T2D A1 @T1, T2D I E-IT1 s d A1 @T1, T2D2

:A1 @T1 , T2 D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ @T1 , T2D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ >
H1,0L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , AH1,0L
•
2
w1 2 w2

whose complex conjugate is

ccSCond1Rule1 = SCond1Rule1 •. conjugateRule;

Chapter5.nb 125

To solve for the particular solution of order2Eq, we express the detuning parameter s in terms of the wi ; that is,

sigRule = Solve@ResonanceCond, sD@@1DD

9s ® - €€€€€€€€1€€€€€€€€€2€€€ =
2w -w
e

Moreover, we express the scale T1 in terms of the scale T0 by using the rule

expRule3 = Exp@a_D :> Exp@a •. T1 -> e T0 •. sigRule •• ExpandD;

With the solvability conditions and the expRule3, order2Eq becomes

order2Eqm = order2Eq •. SCond1Rule1 •. ccSCond1Rule1 •. expRule3;

order2Eqm •. displayRule

8D20 u1,2 + w21 u1,2 == 2 EIT0 w1+I T0 w2 d A1 A2 + 2 E-I T0 w1 -I T0 w2 d A1 A2, D20u2,2 + w22 u2,2 == 2 d A1 A1 <
• • •

Again, to use DSolve, we first transform order2Eqm into a set of ordinary-differential equations and then obtain the particular
solution of order2Eqm as

sol2p = DSolve@order2Eqm •. ui_,j_ -> Hui,j@#1D &L, 8u1,2@T0 D, u2,2@T0 D<, T0D@@1DD •. C@_D -> 0 ••
Simplify

:u1,2 @T0D ®

- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ H2 E-I T0 Hw1 +w2L d HE2I T0 Hw1+w2 L A1 @T1, T2 D A2@T1 , T2 D + A1 @T1, T2 D A2@T1 , T2 DLL,
w2 H2 w1 + w2 L
1 • •

2 d A1@T1 , T2D A1 @T1, T2 D

u2,2 @T0D ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
•

w22

The second-order solution can be expressed in pure function form as

sol2 = TableAui,2 -> FunctionA8T0 , T1, T2 <, Bi @T1, T2 D Exp@I wi T0 D +

Bi @T1 , T2 D Exp@-I wi T0D + sol2p@@i, 2DD •• Expand •• EvaluateE, 8i, 2<E;
•

sol2 •. displayRule

2 EI T0 Hw1+w2 L d A1 A2 2 E-IT0 Hw1+w2 L d A1 A2

:u1,2 ® FunctionB8T0 , T1 , T2<, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€ + E-I T0 w1 B1F,
• •
•
w2 H2 w1 + w2 L w2 H2 w1 + w2 L
€€€€€€€€€€€€ + EIT0 w1 B1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

u2,2 ® FunctionB8T0 , T1 , T2<, EI T0 w2 B2 + €€€€€€€€€€€€€€€€€€€€ + E-IT0 w2 B2 F>

•
2 d A1 A1 •
w22

As shown below, the obtained modulation equations will violate the conservative nature of system unless we include an

neous solutions with the undetermined functions Bi @tD. They will be determined so as to preserve the conservative nature of
appropriate part of the homogeneous solutions of the second-order problem. This is the reason why we included the homoge-

the system.
126 Chapter5.nb

Ÿ Third-Order Equations

Substituting sol1 and sol2 into eqOrder[3] yields

order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

Substituting expRule1[i] into the right-hand sides of order3Eq and collecting the coefficients of EI wi T 0 , we obtain the
solvability conditions as

SCond2 = Coefficient@order3Eq@@#, 2DD •. expRule1@#D, Exp@I w # T0DD == 0 & •ž 81, 2<;

SCond2 •. displayRule

:-HD21A1 L - 2 I HD1 B1L w1 - 2 I HD2 A1L w1 + 2 EI T1 s d B2 A1 + €€€€€€€€€€€€€€€€

• •
• 4 d2 A21 A1 4 d2 A1 A2 A2 •
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 EI T1 s d A2 B1 == 0,
w2 2 2 w1 w2 + w2 2

-HD21A2 L + 2 E-I T1 s d A1 B1 - 2 I HD1B2 L w2 - 2 I HD2A2 L w2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ == 0>

2 •
4 d A1 A2 A1
2 w1 w2 + w22

Using the solvability conditions at second order, SCond1Rule1, we obtain

SCond1Rule2 = D@SCond1Rule1, T1D •. SCond1Rule1 •. ccSCond1Rule1;

SCond1Rule2 •. displayRule

:D21 A1 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€1€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ , D21A2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ - €€€€€€€€€€€€€€€€€€€€€€€€ >

• • • •
EI T1 s d s A2 A1 d2 A2 A1 d2 A1 A2 A2 E-I T1 s d s A2 d2 A1 A2 A1
w1 2 w1 w2 w21 2 w2 w1 w2

Hence, we rewrite the solvability conditions at third order as

H0,1L H0,1L
SCond2Rule1 = SolveASCond2, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD •. SCond1Rule2 •• ExpandAll;
SCond2Rule1 •. displayRule

:D2 A1 ® -HD1 B1L + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• •
I EI T1 s d s A2 A1 I EIT1 s d B2 A1
2 w21 w1
• • • • •
2 I d2 A21 A1 I d2 A21 A1 I d2 A1 A2 A2 2 I d2 A1 A2 A2 I EI T1 s d A B1
w1 H2 w1 w2 + w22 L
€€€€€€€€€€€€€€€€€€€€€€€€
€ - €€€€€€€€€€€€€€€€
€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€2€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €,
w1 w22 4 w21 w2 2 w31 w1

D2 A2 ® -HD1 B2L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

• •
I E-I T1 s d s A21 I E-I T1 s d A1 B1 I d2 A1 A2 A1 2 I d2 A1 A2 A1
4 w22 w2 2 w1 w22 w2 H2 w1 w2 + w22 L

Ÿ Reconstitution

Using SCond1Rule1 and SCond2Rule1, we reconstitute the modulation equations as

Chapter5.nb 127

moduEq = Table@2 I w k A k ¢, 8k, 2<D ==

HTable@2 I w k dt@1D@ A k @T1, T2 DD, 8k, 2<D •. SCond1Rule1 •. SCond2Rule1 •• Expand ••
Collect@#, eD &L •• Thread;
moduEq •. displayRule

:2 I w1 A¢1 ==

i
j
j HD L
• • •
j
• • EIT1 s d s A2 A1 4 d2 A21 A1 d2 A21 A1

k
2 EI T1 s d e A2 A1 + e2 I T1 s
-2 I 1 1B w1 + 2 E d B A
2 1 - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ €€€
€ + €€€€€€€€€€€€€€€€
€€€€
€ + €€€€€€€€
€ €€€
€ €€€€€ -
w1 w22 2 w1 w2

• y
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 EIT1 s d A2 B1z z
• •
z, 2 I w2 A2 ==
d2 A1 A2 A2 4 d2 A1 A2 A2

{
¢
w1 2 2 w1 w2 + w2 2

2i
j 4 d2 A1 A2 A1 y
j2 E d A1 B1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 I HD1 B2 L w2 + €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€2€€€€ z z
z>
• •
d e A1 + e j
-I T s 2
E 1
d s A1 d2 A1 A2 A1

k 2 w1 w2 + w2 {
-I T1 s 2 -I T1 s
E
2 w2 w1 w2

Without including the Bi in the second-order solution, the following two coefficients:

CoefficientAmoduEq@@1, 2DD, A1@T1 , T2 D A2@T1 , T2D A2 @T1, T2 DE •. w2 -> 2 w1 •• Simplify

•

3 d2 e2
- €€€€€€€€€€€€€€€€
2 w21

CoefficientAmoduEq@@2, 2DD, A1@T1 , T2 D A2@T1 , T2D A1 @T1, T2 DE •. w2 -> 2 w1 •• Simplify

•

are not the same, which violates the conservative nature of the system in that the modulation equations must be derivable

choose the functions Bi@tD so that these coefficients are the same. To this end, we assume that
from a Lagrangian. Using the flexibility given by the homogeneous parts of the solutions of the second-order problem, we

BForm = 9B1 -> FunctionA8T1 , T2 <, c1 EI T1 s A1 @T1, T2 D A2@T1 , T2 D E,

•

B1 -> FunctionA8T1 , T2 <, c1 E-I T1 s A1 @T1, T2 D A2@T1 , T2 D E,

• •

B2 -> FunctionA8T1 , T2 <, c2 E-I T1 s A1 @T1, T2 D2 E=;

where c1 and c2 are undetermined constants. Substituting BForm and the solvability conditions into the right-hand sides of
moduEq yields

moduEqMod = HExpand@# •. BForm •. SCond1Rule1 •. ccSCond1Rule1D & •ž moduEqL •• ExpandAll;

moduEqMod •. displayRule

:2 I w1 A¢1 == 2 EI T1 s d e A2 A1 + 2 d e2 A21 c2 A1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 EI T1 s e2 s A2 c1 w1 A1 + €€€€€€€€€€€€€€€€€€€€€€€€

• •
• • EI T1 s d e2 s A2 A1 • 4 d2 e2 A21 A1
€€€€ +
w1 w22
• • • •
d2 e2 A21 A1 d e2 A21 c1 w1 A1 • d2 e2 A1 A2 A2 4 d2 e2 A1 A2 A2
€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€ + 4 d e2 A1 A2 c1 A2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
€€€€€€€€ - €€€€€€€€€€€€€€€€
2 w1 w2 w2 w21 2 w1 w2 + w22
E-I T1 s d e2 s A2 •
2 I w2 A¢2 == E-I T1 s d e A21 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€€ - 2 E-I T1 s e2 s A21 c2 w2 + 2 d e2 A1 A2 c1 A1 +
2 w2

€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

• • •
d2 e2 A1 A2 A1 4 d e2 A1 A2 c2 w2 A1 4 d2 e2 A1 A2 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w1 w2 w1 2 w1 w2 + w22
128 Chapter5.nb

In order that the moduEqMod be derivable from a Lagrangian, the following condition must be satisfied:

CoefficientAmoduEqMod@@1, 2DD, A1@T1 , T2 D A2 @T1 , T2D A2 @T1, T2 DE -

•

CoefficientAmoduEqMod@@2, 2DD, A1@T1 , T2D A2 @T1, T2D A1 @T1, T2 DE == 0 •. e -> 1 •. w2 -> 2 w1

•

3 d2
2 d c1 + 8 d c2 - €€€€€€€€€€ == 0
2 w21

This provides a compatibility condition. However, one extra constant is floating !!!

To obtain a consistent second-order uniform expansion, we apply the method of multiple scales to the system of four first-
order real-valued equations EOM2 in Section 5.2.2 and to a system of two first-order complex-valued equations in Section
5.2.3.

Ÿ 5.2.2 First-Order Real-Valued System

As shown in the preceding section, treating the second-order form of the equations governing a system may lead to
inconsistent results unless appropriate parts of the homogeneous solutions of the perturbation equations are included.
Moreover, the consistent expansions might contain arbitrary constants that need to be chosen judiciously. In this section,
we show that treatment of an equivalent set of first-order real-valued equations, namely EOM2, leads to consistent results
without floating constants.

We seek a second-order uniform expansion of the solution of EOM2 in the form

solRule =
9ui_ -> ISumAej ui,j@#1, #2, #3D, 8j, 3<E &M, vi_ -> ISumAej vi,j@#1, #2, #3D, 8j, 3<E &M=;

Expressing the time derivative in EOM2 in terms of the time scales T0 , T1, and T2 , we have

multiScales = 9u_i_@tD -> ui @T0, T1 , T2 D, Derivative@n_DAu_i_ E@tD :> dt@nD@ui@T0 , T1, T2DD=;

Substituting multiScales and solRule into EOM2, expanding the result for small e, and discarding terms of order higher
than e3 , we obtain

eq522a = HEOM2 •. multiScales •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e, we obtain

eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eq522a •• Transpose;

To place the linear operators on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := I#@@1DD & •ž eqEps@@1DD •. u_ k_,1 -> u k,i M ==

I#@@1DD & •ž eqEps@@1DD •. u_ k_,1 -> u k,iM - H#@@1DD & •ž eqEps@@iDDL •• Thread

Using eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

Chapter5.nb 129

eqOrder@1D •. displayRule •• TableForm

eqOrder@2D •. displayRule •• TableForm
eqOrder@3D •. displayRule •• TableForm

D0 u1,1 - v1,1 == 0
D0 v1,1 + w21 u1,1 == 0
D0 u2,1 - v2,1 == 0
D0 v2,1 + w22 u2,1 == 0

D0 u1,2 - v1,2 == -HD1u1,1 L

D0 v1,2 + w21 u1,2 == -HD1 v1,1L + 2 d u1,1 u2,1
D0 u2,2 - v2,2 == -HD1u2,1 L
D0 v2,2 + w22 u2,2 == -HD1 v2,1L + d u21,1

D0 u1,3 - v1,3 == -HD1u1,2 L - D2 u1,1

D0 v1,3 + w21 u1,3 == -HD1 v1,2L - D2v1,1 + 2 d u1,2 u2,1 + 2 d u1,1 u2,2
D0 u2,3 - v2,3 == -HD1u2,2 L - D2 u2,1
D0 v2,3 + w22 u2,3 == -HD1 v2,2L - D2v2,1 + 2 d u1,1 u1,2

Ÿ First-Order Equations

The homogeneous solution of eqOrder[1] can be obtained by using

eigenForm = #@@2DD Exp@#@@1DD T0 D & •ž Transpose@Eigensystem@matrixADD

:: €€€€€€€€€€€€€€€€€€€€€ , E-I T0 w1 , 0, 0>, :- €€€€€€€€€€€€€€€€€€ , EI T0 w1 , 0, 0>,

I E-I T0 w1 I EI T0 w1
w1 w1

:0, 0, €€€€€€€€€€€€€€€€€€€€€ , E-I T0 w2 >, :0, 0, - €€€€€€€€€€€€€€€€€€ , EI T0 w2 >>

I E-I T0 w2 I EI T0 w2
w2 w2

sol1hForm = FlattenATableAI wi 9-Ai@T1 , T2D, Ai@T1 , T2D=, 8i, 2<EE.eigenForm

•

8EI T0 w1 A1@T1 , T2D + E-I T0 w1 A1 @T1, T2D, I EI T0 w1 w1 A1 @T1, T2 D - I E-I T0 w1 w1 A1 @T1, T2D,
• •

EI T0 w2 A2@T1 , T2D + E-I T0 w2 A2 @T1, T2D, I EI T0 w2 w2 A2 @T1, T2 D - I E-I T0 w2 w2 A2 @T1, T2D<

• •

Since the first-order equations are homogeneous, the solution of eqOrder[1] can be expressed as

states1 = 8u1,1, v1,1 , u2,1 , v2,1<;

sol1 = Table@states1@@iDD -> Function@8T0 , T1 , T2<, sol1hForm@@iDD •• EvaluateD, 8i, 4<D

8u1,1 ® Function@8T0 , T1 , T2<, EI T0 w1 A1@T1 , T2 D + E-I T0 w1 A1@T1 , T2DD,

•

v1,1 ® Function@8T0 , T1 , T2<, I EI T0 w1 w1 A1 @T1, T2D - I E-I T0 w1 w1 A1 @T1 , T2DD,

•

u2,1 ® Function@8T0 , T1 , T2<, EI T0 w2 A2@T1 , T2 D + E-I T0 w2 A2@T1 , T2DD,

•

v2,1 ® Function@8T0 , T1 , T2<, I EI T0 w2 w2 A2 @T1, T2D - I E-I T0 w2 w2 A2 @T1 , T2DD<

•
130 Chapter5.nb

Ÿ Second-Order Equations

Substituting the first-order solution into the second-order equations, eqOrder[2], yields

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

:D0 u1,2 - v1,2 == -EI T0 w1 HD1 A1L - E-I T0 w1 HD1 A1L,

•

D0 v1,2 + w21 u1,2 == 2 EIT0 w1+I T0 w2 d A1 A2 - I EI T0 w1 HD1A1 L w1 + I E-I T0 w1 HD1 A1L w1 + 2 E-I T0 w1 +I T0 w2 d A2 A1 +
• •

2 EI T0 w1 -I T0 w2 d A1 A2 + 2 E-I T0 w1 -IT0 w2 d A1 A2 , D0 u2,2 - v2,2 == -EI T0 w2 HD1 A2L - E-I T0 w2 HD1 A2L,
• • • •

D0 v2,2 + w22 u2,2 == E2I T0 w1 d A21 - I EI T0 w2 HD1 A2L w2 + I E-I T0 w2 HD1A2 L w2 + 2 d A1 A1 + E-2 I T0 w1 d A1>
• • •2

Next, we substitute the expRule1[i] into the right-hand sides of order2Eq to transform the small-divisor terms into secular
terms, collect the terms responsible for secular terms, and obtain

ST11 = Coefficient@#@@2DD •. expRule1@1D, Exp@I w1 T0DD & •ž order2Eq;

ST11 •. displayRule

8-HD1A1 L, -I HD1A1 L w1 + 2 EI T1 s d A2 A1 , 0, 0<

•

ST12 = Coefficient@#@@2DD •. expRule1@2D, Exp@I w2 T0DD & •ž order2Eq;

ST12 •. displayRule

80, 0, -HD1 A2L, E-I T1 s d A21 - I HD1 A2L w2 <

The solvability conditions of the second-order equations demands that ST11 and ST12 be orthogonal to every solution of
the adjoint homogeneous problems, namely the components of ccleftVec. Imposing these conditions, we obtain

SCond1 = 8ccleftVec@@1DD.ST11 == 0, ccleftVec@@2DD.ST12 == 0<;

SCond1 •. displayRule

8-2 I HD1A1 L w1 + 2 EI T1 s d A2 A1 == 0, E-I T1 s d A21 - 2 I HD1 A2 L w2 == 0<

•

or

H1,0L H1,0L
SCond1Rule1 = SolveASCond1, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD;
SCond1Rule1 •. displayRule

:D1 A1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , D1 A2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€

€>
•
I EIT1 s d A2 A1 I E-I T1 s d A21
w1 2 w2

whose complex conjugate is

ccSCond1Rule1 = SCond1Rule1 •. conjugateRule;

Using these solvability conditions and expRule3, we rewrite order2Eq as

Chapter5.nb 131

order2Eqm = order2Eq •. SCond1Rule1 •. ccSCond1Rule1 •. expRule3;

order2Eqm •. displayRule

9D0 u1,2 - v1,2 == €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• •
I E-IT0 w1+I T0 w2 d A2 A1 I EI T0 w1-I T0 w2 d A1 A2
€€€€€€€€€€€€€€€ , D0v1,2 + w21 u1,2 ==
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w1 w1
• • • •
2 EI T0 w1 +I T0 w2 d A1 A2 + E-I T0 w1 +I T0 w2 d A2 A1 + EI T0 w1 -I T0 w2 d A1 A2 + 2 E-I T0 w1-I T0 w2 d A1 A2 ,
•2
I E2I T0 w1 d A2 I E-2 I T0 w1 d A1
D0 u2,2 - v2,2 == €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2 w2 2 w2

D0 v2,2 + w22 u2,2 == €€€€ E2 I T0 w1 d A21 + 2 d A1 A1 + €€€€ E-2 IT0 w1 d A1 =

1 • 1 •2
2 2

Transforming order2Eqm into a set of ordinary-differential equations and using DSolve, we obtain the particular solutions as

Isol2p = DSolveAorder2Eqm •. u_i_,2 -> Hui,2@#1D &L,

8u1,2@T0 D, v1,2@T0 D, u2,2@T0 D, v2,2 @T0 D<, T0 E@@1DD •. C@_D -> 0 •• Simplify;M •• Timing

813.139 Second, Null<

Hence, the second-order solutions of order2Eqm can be expressed as

states2 = states1 •. u_i_,1 -> ui,2

8u1,2 , v1,2, u2,2 , v2,2<

sol2 = Table@states2@@iDD -> Function@8T0 , T1 , T2<,

Expand@sol2p@@i, 2DDD •. Exp@a_D :> Exp@a •• ExpandD •• EvaluateD, 8i, 4<D;
sol2 •. displayRule

9u1,2 ® FunctionB8T0 , T1 , T2<, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

•
2 EI T0 w1 +I T0 w2 d A1 A2 E-IT0 w1+I T0 w2 d A2 A1
w2 H2 w1 + w2 L w1 H2 w1 + w2L
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ +

€€€€€€€€€€€€€€€€€ F,
-IT0 w1+I T0 w2 • IT0 w1-I T0 w2 • IT0 w1-I T0 w2 • • •
2 E-I T0 w1-I T0 w2 d A1 A2
w2 H2 w1 + w2L w1 H2 w1 + w2 L w2 H2 w1 + w2 L w2 H2 w1 + w2L
2E d A2 A1 E d A1 A2 2E d A1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

v1,2 ® FunctionB8T0 , T1 , T2<, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

2 I EIT0 w1+I T0 w2 d A1 A2 2 I EI T0 w1+I T0 w2 d w1 A1 A2
w2 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ -
2 w1 + w2
• • •
I E-IT0 w1+I T0 w2 d A2 A1 2 I E-I T0 w1+I T0 w2 d w1 A2 A1 I EI T0 w1 -I T0 w2 d A1 A2
w2 H2 w1 + w2 L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ +
2 w1 + w2 2 w1 + w2

€€€€€€€€€€€€€€€€€€€€€€€€€€ F,
• • • • •
2 I EIT0 w1-I T0 w2 d w1 A1 A2 2 I E-I T0 w1 -I T0 w2 d A1 A2 2 I E-I T0 w1 -I T0 w2 d w1 A1 A2
w2 H2 w1 + w2 L w2 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w1 + w2

u2,2 ® FunctionA8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ E,

• • •2
E2I T0 w1 d A21 E-2 I T0 w1 d A1
2 w2 H2 w1 + w2 L w2 H2 w1 + w2 L 2 w2 H2 w1 + w2L
4 d w1 A1 A1 2 d A1 A1
w22 H2 w1 + w2 L

v2,2 ® FunctionA8T0 , T1 , T2<, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ E=

•2
I E2 IT0 w1 d A2 I E-2 I T0 w1 d A1
2 H2 w1 + w2L 2 H2 w1 + w2L
132 Chapter5.nb

Ÿ Third-Order Equations

Substituting the first- and second-order solutions into the right-hand sides of the third-order equations, eqOrder[3], yields

order3Eqrhs = Expand@#@@2DD •. sol1 •. sol2D & •ž eqOrder@3D;

To eliminate the terms that lead to secular terms from order3Eqrhs , we use expRule1[i] to transform the terms that convert
small-divisor terms into secular terms, calculate the vectors proportional to the EI wi T 0 , and obtain

ST21 = Coefficient@order3Eqrhs •. expRule1@1D, Exp@I w1 T0DD;

ST21 •. displayRule

EI T1 s d HD1A1 L A2 2 EI T1 s d HD1 A1 L A2 EI T1 s d HD1 A2L A1 2 EI T1 s d HD1 A2L A1

:-HD2A1 L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• • • •

w1 H2 w1 + w2 L w2 H2 w1 + w2 L w1 H2 w1 + w2 L w2 H2 w1 + w2 L
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ ,

d HD1 A1L A2 d HD1 A1L A2 w1 d HD1A2 L A1

-I HD2A1 L w1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
s
• s
• s
•

w2 H2 w1 + w2 L
IE I T 1 2 IE I T1 IE I T1
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ +
2 w1 + w2 2 w1 + w2
d HD1 A2L w1 A1
2 2 • 2 2• I T1 s • 2 •

w2 H2 w1 + w2 L w2 H2 w1 + w2 L w1 H2 w1 + w2 L
8 d A1 w1 A1 5 d A1 A1 2 d A1 A2 A2
w22 H2 w1 + w2 L
2IE
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , 0, 0>

ST22 = Coefficient@order3Eqrhs •. expRule1@2D, Exp@I w2 T0DD;

ST22 •. displayRule

E-I T1 s d HD1 A1L A1 I E-IT1 s d HD1 A1 L A1

:0, 0, -HD2 A2L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ , -I HD2 A2L w2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ >
•
2 d2 A1 A2 A1
w2 H2 w1 + w2 L w1 H2 w1 + w2 L
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€
2 w1 + w2

Hence, demanding that ST21 and ST22 be orthogonal to the components of ccleftVec yields the solvability conditions

SCond2 = 8Expand@ccleftVec@@1DD.ST21D == 0, Expand@ccleftVec@@2DD.ST22D == 0<;

SCond2 •. displayRule

:-2 I HD2A1 L w1 + €€€€€€€€€€€€€€€€

• • •
8 d2 A21 w1 A1 5 d2 A21 A1 2 d2 A1 A2 A2
w22 H2 w1 + w2 L w2 H2 w1 + w2L w1 H2 w1 + w2L
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ == 0,

-2 I HD2A2 L w2 + €€€€€€€€€€€€€€€€
•
2 d2 A1 A2 A1
w1 H2 w1 + w2 L
€€€€€€€€€€€€€€ == 0>

or

SCond2Rule1 = SolveASCond2, 9AH0,1L @T1 , T2D, AH0,1L @T1, T2 D=E@@1DD;

SCond2Rule1 •. displayRule
1 2

I H8 d2 A21 w21 A1 + 5 d2 A21 w1 w2 A1 + 2 d2 A1 A2 w22 A2 L

:D2 A1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ , D2 A2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
• • • •
I d2 A1 A2 A1
2 w21 w22 H2 w1 + w2L w1 w2 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

Ÿ Reconstitution

We reconstitute SCond1Rule1 and SCond2Rule1 and obtain

Chapter5.nb 133

moduEq = Table@2 I w k A k ¢, 8k, 2<D ==

HTable@2 I w k dt@1D@ A k @T1, T2 DD, 8k, 2<D •. SCond1Rule1 •. SCond2Rule1 ••
ExpandL •• Thread;
moduEq •. displayRule

:2 I w1 A¢1 == 2 EI T1 s d e A2 A1 + €€€€€€€€€€€€€€€€€€€€€€€€
• • •
8 d2 e2 A21 w1 A1 5 d2 e2 A21 A1
w2 H2 w1 + w2L w1 H2 w1 + w2 L
• 2 d2 e2 A1 A2 A2
w22 H2 w1 + w2 L
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

2 I w2 A¢2 == E-I T1 s d e A21 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

•
2 d2 e2 A1 A2 A1
w1 H2 w1 + w2L

These equations are derivable from a Lagrangian because

CoefficientAmoduEq@@1, 2DD, A1@T1 , T2 D A2@T1 , T2D A2 @T1, T2 DE -

•

CoefficientAmoduEq@@2, 2DD, A1@T1 , T2 D A2 @T1 , T2D A1 @T1, T2 DE == 0 •. e -> 1 •. w2 -> 2 w1

•

True

Therefore, treatment of the first-order version of the governing equations yields consistent results without floating
constants.

Ÿ 5.2.3 First-Order Complex-Valued System

In this section, we show that treatment of a first-order complex -valued version of the governing equations yields consistent
results without floating constants. We start by transforming the second-order form EOM1 of the governing equations into a
system of first-order complex-valued equations by introducing the transformation

transfRule = 9u k_@tD -> z k @tD + z k @tD, u k_¢ @tD -> I w k Iz k @tD - z k @tDM=;
• •

Then, we define the dependent variables

depVar = 8u1 @tD, u1¢ @tD, u2@tD, u2¢ @tD<;

nmodes = 9z1 @tD, z1@tD, z2 @tD, z2@tD=;

• •

Substituting transfRule into depVar and solving for the nmodes, we obtain

zetaRule = Solve@depVar == HdepVar •. transfRuleL •• Thread, nmodesD@@1DD

-w1 u1 @tD + I u¢1 @tD -w1 u1 @tD - I u¢1 @tD

:z1 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z1 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
•
€€€€€€€€€€€€ , €€€€€€€€€€€€ ,

-w2 u2 @tD + I u¢2 @tD -w2 u2 @tD - I u¢2 @tD

2 w1 2 w1

z2 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z2 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

€€€€€€€€€€€€ >
•
€€€€€€€€€€€€ ,
2 w2 2 w2

It follows from EOM1 that the acceleration is given by

acceleration = Solve@EOM1, 8u1² @tD, u2 ²@tD<D@@1DD

8u²1 @tD ® -w21 u1@tD + 2 d u1@tD u2 @tD, u²2@tD ® d u1 @tD2 - w22 u2 @tD<
134 Chapter5.nb

Then, differentiating the dependent variables in zetaRule and using transfRule and acceleration, we transform the system of
two second-order equations EOM1 into the following two first-order complex-valued equations:

EOM3 = z # ¢@tD == HD@z # @tD •. zetaRule, tD •. acceleration •. transfRule •• ExpandL & •ž 81, 2<

I d z1@tD z2 @tD I d z2 @tD z @tD I d z1@tD z @tD I d z1 @tD z2@tD

:z¢1 @tD == I w1 z1 @tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€2€€€€€€€€€€€ - €€€€€€€€€€€€€€€€
• • • •
€€€€€€€€€€€€€€€€€€€€ ,
w1 w1 w1 w1

I d z1@tD2 I d z1 @tD z @tD I d z1 @tD

z¢2 @tD == - €€€€€€€€€€€€€€€€€€€€€€€€ + I w2 z2 @tD - €€€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ >
• • 2

2 w2 w2 2 w2

To determine a second-order uniform expansion of the solution of EOM3 using the method of multiple scales, we expand z
••
and z in the form

solRule =
9zi_ -> ISumAej zi,j@#1, #2, #3D, 8j, 3<E &M, zi_ -> ISumAej zi,j@#1, #2, #3D, 8j, 3<E &M=;
• •

eq523a = HEOM3 •. multiScales •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e, we obtain

eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eq523a •• Transpose;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := I#@@1DD & •ž eqEps@@1DD •. u_ k_,1 -> u k,i M ==

I#@@1DD & •ž eqEps@@1DD •. u_ k_,1 -> u k,iM - H#@@1DD & •ž eqEps@@iDDL •• Thread

Using eqOrder[i] and the displayRule, we rewrite eqEps in a concise way as

eqOrder@1D •. displayRule •• TableForm

eqOrder@2D •. displayRule •• TableForm
eqOrder@3D •. displayRule •• TableForm

D0 z1,1 - I w1 z1,1 == 0
D0 z2,1 - I w2 z2,1 == 0

D0 z1,2 - I w1 z1,2 == -HD1 z1,1L - €€€€€€€€€€€€€€€€

• • • •
I d z1,1 z2,1 I dz2,1 z1,1 I d z1,1 z2,1 I d z1,1 z2,1
€€€€€ - €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€
€€€€€
w w w w

D0 z2,2 - I w2 z2,2 == -HD1 z2,1L - €€€€€€€€

1 1 1 1
• •2
I d z21,1 I d z1,1 z1,1 Id z1,1
2w
€€€€€ - €€€€€€€€€€€€€€€€
w2
€€€€€ - €€€€€€€€
2 w2
€€€€€
2

D0 z1,3 - I w1 z1,3 == -HD1 z1,2L - D2z1,1 - €€€€€€€€€€€€€€€€

• • • •
Idz z Idz z Idz z I dz z Idz z Idz
1,2 2,1
w1
€€€€€ - €€€€€€€€€€€€€€€€
1,1 2,2
w1
€€€€€ - €€€€€€€€€€€€€€€€
2,2 1,1
w1
€€€€€ - €€€€€€€€€€€€€€€€
2,1 1,2
w1
€€€€€ - €€€€€€€€€€€€€€€€
1,2 2,1
w1
€€€€€ - €€€€€€€€1,2
€w€€€1€

D0 z2,3 - I w2 z2,3 == -HD1 z2,2L - D2z2,1 - €€€€€€€€€€€€€€€€

• • • •
I d z1,1 z1,2 I d z1,2 z1,1 I d z1,1 z1,2 I dz1,1 z1,2
€€€€€ - €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€
€€€€€
w w 2 w 2 w 2 2

Ÿ First-Order Equations

Since eqOrder[1] are homogeneous, we can write their solutions as

Chapter5.nb 135

sol1 = TableA9zi,1 -> Function@8T0 , T1, T2 <, Ai @T1, T2 D Exp@I wi T0 D •• EvaluateD,

zi,1 -> FunctionA8T0 , T1, T2<, Ai @T1, T2D Exp@-I wi T0 D •• EvaluateE=, 8i, 2<E •• Flatten
• •

8z1,1 ® Function@8T0 , T1 , T2<, EI T0 w1 A1@T1 , T2 DD,

z1,1 ® Function@8T0 , T1 , T2<, E-I T0 w1 A1@T1 , T2 DD,
• •

z2,1 ® Function@8T0 , T1 , T2<, EI T0 w2 A2@T1 , T2 DD,

z2,1 ® Function@8T0 , T1 , T2<, E-I T0 w2 A2@T1 , T2 DD<
• •

Ÿ Second-Order Equations

Substituting the first-order solution into the second-order equations, eqOrder[2], we have

order2Eq = eqOrder@2D •. sol1;

order2Eq •. displayRule

9D0 z1,2 - I w1 z1,2 == -EI T0 w1 HD1A1 L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

I EI T0 w1 +I T0 w2 d A1 A2
€€€€€€€€€€€€€€€ -
w1
• • • •
I E-IT0 w1+I T0 w2 d A2 A1 I EIT0 w1-I T0 w2 d A1 A2 I E-I T0 w1-I T0 w2 d A1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ ,
w1 w1 w1

D0 z2,2 - I w2 z2,2 == -EI T0 w2 HD1A2 L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ - €€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ =

• •2
I E2 I T0 w1 d A2 I d A1 A1 I E-2 I T0 w1 d A
2 w2 w2 2 w2

Converting the terms that produce small-divisor terms by using the expRule1 and then eliminating the terms that lead to
secular terms from order2Eq yields

SCond1 = Coefficient@order2Eq@@#, 2DD •. expRule1@#D, Exp@I w # T0DD == 0 & •ž 81, 2<;

SCond1 •. displayRule

:-HD1A1 L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ == 0, -HD1 A2L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€

•
I EI T1 s d A2 A1 I E-I T1 s d A21
€ == 0>
w1 2 w2

or

H1,0L H1,0L
SCond1Rule1 = SolveASCond1, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD;
SCond1Rule1 •. displayRule

:D1 A1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , D1 A2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€

€>
•
I EIT1 s d A2 A1 I E-I T1 s d A21
w1 2 w2

Substituting SCond1Rule1 and expRule3 into order2Eq yields

136 Chapter5.nb

order2Eqm = order2Eq •. SCond1Rule1 •. expRule3;

order2Eqm •. displayRule

9D0 z1,2 - I w1 z1,2

• • •
I EI T0 w1 +IT0 w2 d A1 A2 I EI T0 w1 -IT0 w2 d A1 A2 I E-IT0 w1-I T0 w2 d A1 A2
== - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ ,
w1 w1 w1

== - €€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ =
• •2
I d A1 A1 I E-2 I T0 w1 d A
D0 z2,2 - I w2 z2,2
w2 2 w2

We transform order2Eqm into a set of ordinary-differential equations and obtain their particular solutions as

sol2p = DSolve@order2Eqm •. zi_,2 -> Hzi,2@#1D &L, 8z1,2@T0 D, z2,2@T0 D<, T0D@@1DD •. C@_D -> 0;

expRule4 = Exp@a_D :> Exp@Expand@aDD;

We write the second-order solution in function form as

sol2 = TableA9zi,2 -> Function@8T0 , T1, T2 <, Hsol2p@@i, 2DD •• ExpandL •. expRule4 •• EvaluateD,
zi,2 -> Function@8T0 , T1, T2<, Hsol2p@@i, 2DD •• ExpandL •. expRule4 •. conjugateRule ••
•

EvaluateD=, 8i, 2<E •• Flatten;

sol2 •. displayRule

9z1,2 ® FunctionB8T0 , T1 , T2<, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

EI T0 w1 +IT0 w2 d A1 A2 2 EI T0 w1 +I T0 w2 d A1 A2
w1 H2 w1 + w2 L w2 H2 w1 + w2 L
€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ +

€€€€€€€€€€€€€ F,
• • • •
EI T0 w1 -IT0 w2 d A1 A2 2 EI T0 w1 -I T0 w2 d A1 A2 E-I T0 w1 -I T0 w2 d A1 A2
w1 H2 w1 + w2 L w2 H2 w1 + w2 L w1 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

z1,2 ® FunctionB8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

I T0 w1 +I T0 w2 •
E-I T0 w1 +I T0 w2 d A2 A1
w1 H2 w1 + w2L w1 H2 w1 + w2 L
• E d A1 A2
€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ +

€€€€€€€€€€€€€€€€€ F,
• • • • •
2 E-IT0 w1+I T0 w2 d A2 A1 E-IT0 w1-I T0 w2 d A1 A2 2 E-I T0 w1-I T0 w2 d A1 A2
w2 H2 w1 + w2L w1 H2 w1 + w2L w2 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

z2,2 ® FunctionA8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ E,

• • •2
E-2I T0 w1 d A1
w2 H2 w1 + w2 L 2 w2 H2 w1 + w2 L
2 d w1 A1 A1 d A1 A1
w22 H2 w1 + w2L

z2,2 ® FunctionB8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ F=

• •
• E2I T0 w1 d A21 2 d w1 A1 A1 d A1 A1
2 w2 H2 w1 + w2 L w22 H2 w1 + w2 L w2 H2 w1 + w2 L

Ÿ Third-Order Equations

Substituting the first- and second-order solutions into the right-hand sides of the third-order equations, eqOrder[3], we
obtain

order3Eqrhs = Expand@#@@2DD •. sol1 •. sol2D & •ž eqOrder@3D;

Converting the terms that produce small-divisor terms and eliminating the terms that produce secular terms in z i,3 from
order3Eqrhs demands the following conditions:
Chapter5.nb 137

SCond2 = Coefficient@order3Eqrhs@@#DD •. expRule1@#D, Exp@I w # T0 DD == 0 & •ž 81, 2<;

SCond2 •. displayRule

:-HD2A1 L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ == 0, -HD2A2 L - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• • • •
4 I d2 A21 A1 5 I d2 A21 A1 I d2 A1 A2 A2 I d2 A1 A2 A1
w22 H2 w1 + w2L 2 w1 w2 H2 w1 + w2L w21 H2 w1 + w2L w1 w2 H2 w1 + w2L
€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ == 0>

or

SCond2Rule1 = SolveASCond2, 9AH0,1L @T1 , T2D, AH0,1L @T1, T2 D=E@@1DD;

SCond2Rule1 •. displayRule
1 2

I H8 d2 A21 w21 A1 + 5 d2 A21 w1 w2 A1 + 2 d2 A1 A2 w22 A2 L

:D2 A1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ , D2 A2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
• • • •
I d2 A1 A2 A1
2 w21 w22 H2 w1 + w2L w1 w2 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

Ÿ Reconstitution

We reconstitute SCond1Rule1 and SCond2Rule1 and obtain

moduEq = Table@2 I w k A k ¢, 8k, 2<D ==

HTable@2 I w k dt@1D@ A k @T1, T2 DD, 8k, 2<D •. SCond1Rule1 •. SCond2Rule1 ••
ExpandL •• Thread;
moduEq •. displayRule

:2 I w1 A¢1 == 2 EI T1 s d e A2 A1 + €€€€€€€€€€€€€€€€
• • •
8 d2 e2 A21 w1 A1 5 d2 e2 A21 A1 2 d2 e2 A1 A2 A2
w2 H2 w1 + w2L w1 H2 w1 + w2 L
•
w22 H2 w1 + w2 L
€€€€€€€€ € €€€€€€€
€ + €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ € + €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€€€€
€,

2 I w2 A¢2 == E-I T1 s d e A21 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

•
2 d2 e2 A1 A2 A1
w1 H2 w1 + w2L

à 5.3 Method of Normal Forms

To apply the method of normal forms, we find it convenient to start with the first-order complex-valued form EOM3

I d z1 @tD z2@tD I d z2 @tD z @tD I d z1 @tD z @tD I d z1 @tD z2 @tD

• • • •
EOM3 = 9z¢1 @tD == I w1 z1 @tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€2€€€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ ,
w1 w1 w1 w1

I d z1@tD2 I d z1@tD z @tD I d z1 @tD

• •
z¢2 @tD == - €€€€€€€€€€€€€€€€€€€€€€€€ + I w2 z2 @tD - €€€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€ =;
2

2 w2 w2 2 w2

of the governing equations. Then, according to the method of normal forms, we introduce the near-identity transformation

basicTerms = Table@8h k@tD, •

h k @tD<, 8k, 2<D •• Flatten

8h1 @tD, h1@tD, h2 @tD, h2 @tD<

• •
138 Chapter5.nb

zetaRule =
9zi_ -> FunctionAt, e hi @tD + SumAej+1 hi,j@Sequence žž basicTermsD, 8j, 2<E •• EvaluateE,
hi @tD + SumAej+1 hi,j@Sequence žž basicTermsD, 8j, 2<E •• EvaluateE=
• •
zi_ -> FunctionAt, e •

8zi_ ® Function@t,
e hi @tD + e2 hi,1 @h1 @tD, h1@tD, h2 @tD, h2@tDD + e3 hi,2 @h1@tD, h1 @tD, h2 @tD, h2@tDDD, zi_ ®
• • • • •

Function@t, e hi @tD + e2 hi,1 @h1@tD, h1 @tD, h2 @tD, h2@tDD + e3 hi,2 @h1@tD, h1 @tD, h2@tD, h2 @tDDD<
• • • • • • •

that results in the simplest possible equations

etaRule = hi_ ¢@tD -> I wi hi @tD + SumAej gi,j@tD, 8j, 2<E

h¢i_ @tD ® I wi hi @tD + e gi,1@tD + e2 gi,2 @tD

where the overbar denotes the complex conjugate

conjugateRule = 8h -> •
h, •
h -> h, g -> •
g, •
g -> g, Complex@0, n_D -> Complex@0, -nD<;

Substituting the expansion for the z i , the zetaRule, into EOM3 , using the etaRule, expanding the result for small e, and
discarding terms of order higher than e3, we have

eq53a =
HEOM3 •. zetaRule •. etaRule •. HetaRule •. conjugateRuleL •• ExpandAllL •. e n_•;n>3 -> 0;

Ÿ Second-Order Equations

Collecting the coefficients of e2 in eq53a yields

order2expr = CoefficientASubtract žž #, e2 E & •ž eq53a;

To determine the resonance terms in order2expr, we note that all of the possible forms of its nonhomogeneous terms are

possibleQTerms = Outer@Times, basicTerms, basicTermsD •• Flatten •• Union

9h1 @tD2, h1 @tD h2@tD, h2 @tD2 , h1@tD h1 @tD,

•

h2 @tD h1@tD, h1 @tD2, h1 @tD h2@tD, h2 @tD h2@tD, h1@tD h2 @tD, h2@tD2 =
• • • • • • •

It follows from the linear parts of the etaRule that to the first approximation

form = 9hi_ @tD -> EI wi t, •

hi_@tD -> E-I wi t =;

Hence, the possible resonance terms are given by

ResonantQTerm@1D =
IE-I w1 t possibleQTerms •. form •. w2 -> 2 w1 •. E _ t -> 0M possibleQTerms •• Union •• Rest

8h2 @tD h1@tD<

•
Chapter5.nb 139

ResonantQTerm@2D =
i
j E-I w2 t possibleQTerms •. form •. w1 -> €€€€ w2 •. E _ t -> 0y
z possibleQTerms •• Union •• Rest
k
1
2 {

8h1 @tD2<

Then, the nonresonance terms and their associated coefficients can be defined as

Do@NRQT@iD = Complement@possibleQTerms, ResonantQTerm@iDD;

coeffsQ@iD = Table@Gi,j, 8j, Length@NRQT@iDD<D, 8i, 2<D

Thus, the h i,1 and their complex conjugates have the form

hFormQ =
TableA9hi,1 -> HEvaluate@[email protected]@iD •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
hi,1 -> HEvaluate@[email protected]@iD •. conjugateRule •.
•

Thread@basicTerms -> 8#1, #2, #3, #4<DD &L=, 8i, 2<E •• Flatten;

Substituting for the h i,1 in order2expr, equating the coefficients of the possible nonresonance terms to zero, and solving the
resulting algebraic equations for the Gi, j, we obtain

coeffsQRule = Table@
Solve@Coefficient@order2expr@@iDD •. hFormQ, NRQT@iDD == 0 •• Thread, coeffsQ@iDD@@1DD,
8i, 2<D •• Flatten

:G1,1 ® 0, G1,2 ® - €€€€€€€€€€€€ , G1,4 ® 0, G1,5 ® 0, G1,6 ® €€€€€€€€€€€€ , G1,7 ® 0,

d d
w1 w2 w1 w2

w1 H2 w1 + w2 L
d d
G1,3 ® 0, G1,8 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , G1,9 ® 0, G2,1 ® 0, G2,2 ® 0, G2,3 ® €€€€€€ ,
w22

2 w2 H2 w1 + w2 L
d
G2,4 ® 0, G2,7 ® 0, G2,9 ® 0, G2,5 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , G2,6 ® 0, G2,8 ® 0>

Using these coefficients, we rewrite hFormQ as

hSolQ = Table@hi,1@Sequence žž basicTermsD ->

Hhi,1 @Sequence žž basicTermsD •. hFormQ •. coeffsQRuleL, 8i, 2<D

d h1 @tD h2@tD d h1@tD h2 @tD d h1 @tD h2@tD

:h1,1 @h1@tD, h1 @tD, h2@tD, h2 @tDD ® - €€€€€€€€€€€€€€€€
• • •

w1 H2 w1 + w2 L
• •
€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ ,
w1 w2 w1 w2
d h1 @tD h1@tD d h1 @tD
h2,1 @h1@tD, h1 @tD, h2@tD, h2 @tDD ® €€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ >
• • 2

2 H2 w1 + w2 L
• •
€€€€€€€€ + €€€€€€€€€€€€€€€€
w22 2 w

or in pure function form as

hRuleQ = TableA9hi,1 -> HEvaluate@hSolQ@@i, 2DD •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
hi,1 -> HEvaluate@hSolQ@@i, 2DD •. conjugateRule •.
•

Thread@basicTerms -> 8#1, #2, #3, #4<DD &L=, 8i, 2<E •• Flatten;
140 Chapter5.nb

We choose the g i,1 to eliminate the resonance terms in order2expr according to

gRuleQ = Table@
gi,1 @tD -> - Coefficient@order2expr@@iDD, [email protected]@iD, 8i, 2<D

I d h2@tD • h @tD I d h1@tD2

:g1,1 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€€€€ , g2,1 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€ >
w1 2 w2

Ÿ Third-Order Equations

Collecting the coefficients of e3 in eq53a and using hRuleQ and gRuleQ, we have

order3expr = MapACoefficientASubtract žž #, e3E &, eq53aE •. hRuleQ •. gRuleQ •.

HgRuleQ •. conjugateRuleL •• Expand;

The nonhomogeneous terms in order3expr are proportional to

possibleCTerms = Outer@Times, possibleQTerms, basicTermsD •• Flatten •• Union

9h1 @tD3, h1 @tD2 h2 @tD, h1 @tD h2 @tD2 , h2@tD3 , h1@tD2 h1@tD,

•

h1 @tD h2@tD h1 @tD, h2@tD2 h1@tD, h1 @tD h1@tD2 , h2@tD h1 @tD2, h1@tD3 ,
• • • • •

h1 @tD2 h2 @tD, h1@tD h2 @tD h2@tD, h2 @tD2 h2 @tD, h1@tD h1 @tD h2@tD,
• • • • •

h2 @tD h1@tD h2 @tD, h1@tD h2@tD, h1 @tD h2@tD , h2@tD h2 @tD , h1@tD h2 @tD2 , h2 @tD3=
• • • 2• • 2 • 2 • • •

Next, we determine the resonance terms according to

ResonantCTerm@1D =
IE-I w1 t possibleCTerms •. form •. w2 -> 2 w1 •. E _ t -> 0M possibleCTerms •• Union •• Rest

8h1 @tD2 h1 @tD, h1@tD h2 @tD h2@tD<

• •

ResonantCTerm@2D =
i
j E-I w2 t possibleCTerms •. form •. w1 -> €€€€ w2 •. E _ t -> 0y
z possibleCTerms •• Union •• Rest
k
1
2 {

8h1 @tD h2@tD h1 @tD, h2@tD2 h2@tD<

• •

Then, the nonresonance terms and their associated coefficients are given by

Do@NRCT@iD = Complement@possibleCTerms, ResonantCTerm@iDD;

coeffsC@iD = Table@Li,j, 8j, Length@NRCT@iDD<D, 8i, 2<D

Consequently, the h i,2 and their complex conjugates have the forms

hFormC =
TableA9hi,2 -> HEvaluate@[email protected]@iD •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
hi,2 -> HEvaluate@[email protected]@iD •. conjugateRule •.
•

Thread@basicTerms -> 8#1, #2, #3, #4<DD &L=, 8i, 2<E •• Flatten;
Chapter5.nb 141

Substituting for the h i,2 in order3expr, equating the coefficient of each possible nonresonance term to zero, and solving the
resulting algebraic equations for the Li, j , we obtain

coeffsCRule = Table@
Solve@Coefficient@order3expr@@iDD •. hFormC, NRCT@iDD == 0 •• Thread, coeffsC@iDD@@1DD,
8i, 2<D •• Flatten

:L1,1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
d2 w1 + d2 w2 d2 -5 d2 w1 - 3 d2 w2
2 w21 w22 H2 w1 + w2L w1 w22 H2 w1 + w2 L 2 w21 w22 H2 w1 + w2 L
€€€€€€€€ , L1,3 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L1,7 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ ,

d2 d2 -d2 w1 - d2 w2
4 w21 w2 H2 w1 + w2 L w21 w2 H2 w1 + w2 L w21 w22 H2 w1 + w2 L
L1,9 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ , L1,13 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L1,15 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

d2
w1 w2 Hw1 + w2 L H2 w1 + w2L
L1,2 ® 0, L1,16 ® 0, L1,4 ® 0, L1,17 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ , L1,5 ® 0,

d2
w21 Hw1 - w2 L w2
L1,18 ® 0, L1,10 ® 0, L1,8 ® 0, L1,6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L1,11 ® 0, L1,14 ® 0, L1,12 ® 0,

d2 d2
w1 w22 H2 w1 + w2 L w1 w22 H2 w1 + w2 L
L2,2 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L2,4 ® 0, L2,8 ® 0, L2,12 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L2,16 ® 0, L2,18 ® 0,

d2
w1 Hw1 - w2L w22
L2,1 ® 0, L2,3 ® 0, L2,5 ® 0, L2,6 ® 0, L2,7 ® 0, L2,9 ® 0, L2,10 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

d2
w22 Hw1 + w2L H2 w1 + w2 L
L2,11 ® 0, L2,17 ® 0, L2,13 ® 0, L2,14 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ , L2,15 ® 0>

Using these coefficients, we rewrite hFormC as

hSolC = Table@hi,2@Sequence žž basicTermsD ->

Hhi,2 @Sequence žž basicTermsD •. hFormC •. coeffsCRuleL, 8i, 2<D

Hd2 w1 + d2 w2 L h1@tD3 d2 h1@tD h2 @tD2 d2 h2 @tD2 h1 @tD

:h1,2 @h1@tD, h1 @tD, h2@tD, h2 @tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
•

2 w21 w22 H2 w1 + w2 L w1 w22 H2 w1 + w2L w21 Hw1 - w2L w2

• •
€€€€€€€€€ €€€
€ €€€
€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€
€€€€€€€€€€€€ -

H-5 d2 w1 - 3 d2 w2 L h1@tD h1 @tD2 d2 h1@tD3 d2 h2@tD h1 @tD h2@tD

• • • •

2 w21 w22 H2 w1 + w2L 4 w21 w2 H2 w1 + w2 L w21 w2 H2 w1 + w2 L

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ -

H-d2 w1 - d2 w2L h1 @tD h2 @tD2 d2 h1@tD h2 @tD2

€€€€€€€€€€€€€€€€€€€€€€€€€ , h2,2 @h1 @tD, h1 @tD, h2@tD, h2 @tDD ®
• • •

w21 w22 H2 w1 + w2L w1 w2 Hw1 + w2 L H2 w1 + w2L

• •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

d2 h1 @tD2 h2 @tD d2 h1@tD2 h2@tD d2 h1 @tD h1@tD h2 @tD d2 h1 @tD2 h2 @tD

€€€€€€€€€€€€€€€€€€€€ >
• • • • •

w1 w22 H2 w1 + w2L w1 Hw1 - w2L w22 w1 w22 H2 w1 + w2 L w22 Hw1 + w2 L H2 w1 + w2 L

€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

or in pure function form as

hRuleC = TableA9hi,2 -> HEvaluate@hSolC@@i, 2DD •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
hi,2 -> HEvaluate@hSolC@@i, 2DD •. conjugateRule •.
•

Thread@basicTerms -> 8#1, #2, #3, #4<DD &L=, 8i, 2<E •• Flatten;

Choosing the g i,2 to eliminate the resonance terms from order3expr yields
142 Chapter5.nb

gRuleC = Table@
gi,2 @tD -> - Coefficient@order3expr@@iDD, [email protected]@iD, 8i, 2<D

i 2 I d2 y I d2 h1 @tD h2 @tD h2@tD

:g1,2 @tD ® -j
j €€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z h1 @tD2 •
h1 @tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
•
I d2
k w1 w22 2 w1 w2 H2 w1 + w2L { w21 H2 w1 + w2 L
€€€€€€€€€€€€€€€€€€€€ ,

I d2 h1 @tD h2@tD h1 @tD

g2,2 @tD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ >
•

w1 w2 H2 w1 + w2 L

Combining etaRule, gRuleQ, and gRuleC and letting hi = Ai @tD EI wi t , we obtain the modulation equation

moduEq = TableA2 I w k E-I wk t Hh k ¢@tD - Hh k¢ @tD •. etaRule •. gRuleQ •. gRuleCLL == 0 •.

9h -> IA @#D EI wi # &M, • h -> IA @#D E-I wi # &M= •.
•

Exp@a_D :> ExpAa •. w2 -> 2 w1 + e s •• ExpandE •• ExpandAll, 8k, 2<E

i_ i i_ i
2

4 d2 e2 A1 @tD2 A1 @tD 2 d2 e2 w1 A1 @tD2 A1 @tD

:- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• •
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ -
w22 4 w21 w2 + 2 w1 w22
2 d2 e2 w1 A1 @tD A2@tD A2 @tD
d e A2@tD A1 @tD - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 I w1 A¢1 @tD == 0,
•
•
2 EI t e
2s

2 w31 + w21 w2
2 d2 e2 w2 A1 @tD A2@tD A1 @tD
d e A1 @tD2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 I w2 A¢2@tD == 0>
•
-E-I t e
2s

2 w21 w2 + w1 w22

Using zetaRule, hRuleQ, and hRuleC, we obtain the following second-order uniform expansion of the solution of EOM3:
Chapter5.nb 143

solution = TableAu k@tD == CollectAz k @tD + z k @tD •. zetaRule •. hRuleQ •. hRuleC, eE, 8k, 2<E
•

i d h1 @tD h2@tD d h1@tD h2 @tD

:u1 @tD == e Hh1@tD + h1@tDL + e2 j j- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
k w1 H2 w1 + w2L
•

d h2@tD h1 @tD d h1 @tD h2@tD d h1@tD h2 @tD d h1 @tD h2@tD y

w1 w2

€€€€€€€€€€€€€€€€ z
z+
• • • • • •

w1 H2 w1 + w2 L {
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€

i
w1 w2 w1 w2 w1 w2
d2 h1@tD3 Hd2 w1 + d2 w2L h1 @tD3 d2 h1 @tD h2 @tD2 d2 h1@tD h2 @tD2
e3 j
j
j €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
k 4 w1 w2 H2 w1 + w2 L 2 w1 w2 H2 w1 + w2 L w1 w2 H2 w1 + w2L w1 w2 Hw1 + w2 L H2 w1 + w2L
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ -
2 2 2 2

H-5 d2 w1 - 3 d2 w2 L h1@tD2 h1@tD d2 h2@tD2 h1@tD H-d2 w1 - d2 w2L h2 @tD2 h1 @tD

• • •

2 w21 w22 H2 w1 + w2L w21 Hw1 - w2L w2 w21 w22 H2 w1 + w2L

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

H-5 d2 w1 - 3 d2 w2 L h1@tD h1 @tD2 d2 h1 @tD3 Hd2 w1 + d2 w2 L h1@tD3

• • •

2 w21 w22 H2 w1 + w2L 4 w21 w2 H2 w1 + w2 L 2 w21 w22 H2 w1 + w2 L

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€ -

d2 h1 @tD h2@tD h2 @tD d2 h2@tD h1 @tD h2@tD d2 h1 @tD h @tD2

• • • •

w21 w2 H2 w1 + w2 L w21 w2 H2 w1 + w2L w21 Hw1 - w2L w2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€2€€€€€€€€€€€€ -

H-d2 w1 - d2 w2 L h1 @tD h2 @tD2 d2 h1 @tD h2@tD2 d2 h1 @tD h2 @tD2 y

€€€€€€€€€€€€€€€€€€€€€€€€€ z
z,
z
• • • • •

w1 w2 H2 w1 + w2L w1 w2 H2 w1 + w2 L w1 w2 Hw1 + w2L H2 w1 + w2 L {

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€
€ €€€€€€€
€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€
€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 2 2

i d h1@tD2 2 d h1 @tD h1@tD d h1@tD2 y

u2 @tD == e2 j j
j €€€€€€€€€€€€€€€€€€ zz
z + e Hh2 @tD + h2 @tDL +
• •

H2 L H2 L
•
k 2 {
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€
€ € + €€€€€€€€€€€€€€€€€€€€€€€€
€ €€€€€€€€ €€
€ + €€€€€€€€€€€€€€€€
2 w w + w w 2 2 w w + w

i d h1 @tD h2 @tD d h1 @tD h2 @tD d h1 @tD h2@tD h1 @tD d2 h2 @tD h @tD2

2 1 2 2 2 1

e3 j
j
j €€€€€€€€€€€€€€€€
2 2 2 2 2 • •

k w1 w2 H2 w1 + w2 L w22 Hw1 + w2 L H2 w1 + w2L w1 w22 H2 w1 + w2 L w1 Hw1 - w2 L w22

€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€€€€€ -
2

d2 h1 @tD2 h @tD d2 h1@tD h1 @tD h2@tD d2 h1 @tD2 h2 @tD d2 h1 @tD2 h2 @tD y

€€€€€€€€€€€€€€€€€€€€ z
z
z>
• • • • • • •

w1 Hw1 - w2 L w2 w1 w2 H2 w1 + w2 L w1 w2 H2 w1 + w2 L w2 Hw1 + w2L H2 w1 + w2 L {

€€€€€€€€€€€€€€€€€€€€€€€€€2€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 2 2 2

à 5.4 Generalized Method of Averaging

We define the nonlinear part of EOM1, a transformation rule, and the states vector as

NLTerms = 9f1 -> 2 e d u1 @tD u2@tD, f2 -> e d u1@tD2 =;

transformRule = ui_ -> Hai @#D Sin@fi@#DD &L;
states = 8a1 @tD, a2@tD, f1 @tD, f2@tD<;

Substituting the above rules into the four first-order equations obtained by using the method of variation of parameters, we
have
144 Chapter5.nb

eq54a = D@states, tD ==
i
j9 €€€€€€ Cos@f1@tDD, €€€€€€ Cos@f2@tDD, w1 - €€€€€€€€€€€€€€€€
€€€€ Sin@f1@tDD, w2 - €€€€€€€€€€€€€€€€
€€€€ Sin@f2 @tDD= •.
k w1
f1 f2
a1 @tD w1
f1
a2 @tD w2
f2
w2
y •• Thread
NLTerms •. transformRule •• TrigReduce •• Expandz
{
d e Cos@2 f1 @tD - f2 @tDD a1 @tD a2@tD d e Cos@2 f1@tD + f2@tDD a1 @tD a2@tD
:a¢1 @tD == €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€ , a¢2@tD ==
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w1 2 w1
d e Cos@2 f1 @tD - f2 @tDD a1@tD2 d e Cos@f2 @tDD a1@tD2 d e Cos@2 f1 @tD + f2 @tDD a1 @tD2
- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ ,

d e Sin@2 f1 @tD - f2 @tDD a2 @tD d e Sin@f2 @tDD a2@tD

4 w2 2 w2 4 w2

f¢1 @tD == w1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

€€€€€€€€€€€€€€€€€ +

d e Sin@2 f1 @tD + f2 @tDD a2@tD

2 w1 w1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , f¢2 @tD ==

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w1
d e Sin@2 f1@tD - f2@tDD a1 @tD2 d e Sin@f2@tDD a1@tD2 d e Sin@2 f1 @tD + f2 @tDD a1@tD2
€€€€€€€€ >
4 w2 a2@tD 2 w2 a2@tD 4 w2 a2@tD
w2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

We seek a second-order approximate solution of eq54a in the form

basicTerms = 8h1@tD, h2@tD, j1 @tD, j2@tD<;

solRule = 9 ai_ -> IEvaluateAhi @tD + SumAej ai,j @Sequence žž basicTermsD, 8j, 2<E •. t -> #E &M,
fi_ -> IEvaluateAji @tD + SumAej fi,j@Sequence žž basicTermsD, 8j, 2<E •. t -> #E &M=;

where the hi @tD and ji @tD are expanded in power series in e as

basicDRule =
D@basicTerms, tD -> 9SumAei A1,i @h1@tD, h2 @tDD, 8i, 2<E, SumAei A2,i @h1@tD, h2 @tDD, 8i, 2<E,
w1 + SumAei F1,i @h1 @tD, h2@tDD, 8i, 2<E, w2 + SumAei F2,i@h1 @tD, h2@tDD, 8i, 2<E= •• Thread

8h¢1 @tD ® e A1,1 @h1 @tD, h2@tDD + e2 A1,2@h1 @tD, h2 @tDD,

h¢2 @tD ® e A2,1 @h1 @tD, h2@tDD + e2 A2,2@h1 @tD, h2 @tDD,
j¢1 @tD ® w1 + e F1,1@h1 @tD, h2@tDD + e2 F1,2 @h1@tD, h2 @tDD,
j¢2 @tD ® w2 + e F2,1@h1 @tD, h2@tDD + e2 F2,2 @h1@tD, h2 @tDD<

The functions a i, j and fi, j are fast varying functions of the ji, while it follows from basicDRule that the hi , and hence the
Ai, j and Fi, j , are slowly varying functions of t.

To the second approximation, we differentiate the a i @tD and fi@tD once with respect to t, use solRule and basicDRule, expand
the result for small e, discard terms of order higher than e2, and obtain

eq54bLHS =
CoefficientListAExpand@#@@1DD •. solRule •. basicDRuleD •. e n_•;n>2 -> 0, eE & •ž eq54a;

Next, we substitute solRule into the right-hand sides of eq54a, expand the result for small e, keep terms up to OHe2L, and
rewrite their right-hand sides as

eq54bRHS = CoefficientList@Series@#@@2DD •. solRule, 8e, 0, 2<D •• Normal, eD & •ž eq54a;

Chapter5.nb 145

Equating coefficients of like powers of e in eq54bLHS and eq54bRHS, we obtain

eqEps = MapThread@Equal, 8eq54bLHS, eq54bRHS<, 2D •• Transpose •• Rest;

To express the nearness of the two-to-one internal resonance, we introduce the detuning parameter s defined according to

omgRule = 8w2 -> 2 w1 + e s<;

Hence, s can be related to the wi by

sigRule = Solve@omgRule •. Rule -> Equal, sD@@1DD

9s ® - €€€€€€€€€€€€€€€€€€€€ =
2 w1 - w2
e

To use the method of averaging, we define the following rules:

psiRule = 8j1 @tD -> w1 t + b1 , j2 @tD -> w2 t + b2<;

betaRule = Solve@psiRule •. Rule -> Equal, 8b1 , b2 <D@@1DD

8b1 ® -t w1 + j1 @tD, b2 ® -t w2 + j2 @tD<

expRule1 = f_@a_D :> f@Expand@a •. psiRule •. omgRuleD •. e t -> t1 D;

expRule2 = f_@a_D :> f@Expand@a •. t1 -> e t •. sigRule •. betaRuleDD;

Next, we separate the fast and slowly varying terms in the first-order equations, eqEps[[1]]. To this end, we define a slow
state vector and a fast state vector according to

SVT@j_D := 8A1,j@h1 @tD, h2 @tDD, A2,j @h1 @tD, h2@tDD, F1,j @h1@tD, h2 @tDD, F2,j@h1 @tD, h2@tDD<
solVar@j_D := 8a1,j, a2,j , f1,j, f2,j <

Then, the slowly varying parts of eqEps[[1]] are given by

SVT1Rule =
Table@Solve@eqEps@@1, iDD •. Thread@solVar@1D -> H0 &LD •. expRule1 •. f_@_ t + _.D -> 0 •.
expRule2 •. f_@a_D :> f@Expand@aDD, SVT@1D@@iDDD @@1DD, 8i, 4<D •• Flatten

d Cos@2 j1 @tD - j2 @tDD h1 @tD h2@tD

:A1,1 @h1@tD, h2 @tDD ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ ,
2 w1
d Cos@2 j1@tD - j2@tDD h1 @tD2
A2,1 @h1@tD, h2 @tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

d Sin@2 j1@tD - j2@tDD h2 @tD

4 w2

F1,1 @h1@tD, h2 @tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2 w1
d Sin@2 j1@tD - j2@tDD h1 @tD2
F2,1 @h1@tD, h2 @tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
4 w2 h2 @tD

whereas the fast varying parts are given by

146 Chapter5.nb

FVT1 = Table@Subtract žž eqEps@@1, iDD •. Thread@solVar@1D -> H0 &LD, 8i, 4<D •. SVT1Rule

d Cos@2 j1@tD + j2@tDD h1 @tD h2 @tD d Cos@j2@tDD h1 @tD2 d Cos@2 j1@tD + j2@tDD h1 @tD2
: €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ , - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

d Sin@j2@tDD h2 @tD d Sin@2 j1 @tD + j2 @tDD h2 @tD

2 w1 2 w2 4 w2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
w1 2 w1
d Sin@j2@tDD h1 @tD2 d Sin@2 j1@tD + j2@tDD h1 @tD2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
2 w2 h2@tD 4 w2 h2@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

To determine the solution corresponding to these fast varying terms, we use the method of undetermined coefficients. To
accomplish this, we first identify the possible forms of the nonhomogeneous terms. The result is

FVT1Forms = Flatten@Cases@#, HCos È SinL@a_D -> 8Cos@aD, Sin@aD<, InfinityDD & •ž FVT1

88Cos@2 j1 @tD + j2 @tDD, Sin@2 j1 @tD + j2 @tDD<,

8Cos@j2 @tDD, Sin@j2@tDD, Cos@2 j1 @tD + j2 @tDD, Sin@2 j1 @tD + j2 @tDD<,
8Cos@j2 @tDD, Sin@j2@tDD, Cos@2 j1 @tD + j2 @tDD, Sin@2 j1 @tD + j2 @tDD<,
8Cos@j2 @tDD, Sin@j2@tDD, Cos@2 j1 @tD + j2 @tDD, Sin@2 j1 @tD + j2 @tDD<<

The fast-varying component of the first-order solution can be expressed as a linear combination of these forms; that is,

sol1Form = MapIndexed@Hcoeffs1@#2@@1DDD = Array@c, Length@#1DDL.#1 &, FVT1FormsD

8c@1D Cos@2 j1@tD + j2@tDD + c@2D Sin@2 j1@tD + j2@tDD,

c@1D Cos@j2@tDD + c@3D Cos@2 j1@tD + j2@tDD + c@2D Sin@j2 @tDD + c@4D Sin@2 j1 @tD + j2 @tDD,
c@1D Cos@j2@tDD + c@3D Cos@2 j1@tD + j2@tDD + c@2D Sin@j2 @tDD + c@4D Sin@2 j1 @tD + j2 @tDD,
c@1D Cos@j2@tDD + c@3D Cos@2 j1@tD + j2@tDD + c@2D Sin@j2 @tDD + c@4D Sin@2 j1 @tD + j2 @tDD<

Substituting sol1Form into eqEps[[1]], using SVT1Rule, collecting the coefficients of FVT1Forms, solving the resulting
algebraic equation for the undetermined coefficients, and then substituting the result back into sol1Form, we obtain

sol1rhs = Table@sol1Form@@iDD •.
Solve@Coefficient@Subtract žž eqEps@@1, iDD •. SVT1Rule •. solVar@1D@@iDD ->
HEvaluate@sol1Form@@iDD •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L,
FVT1Forms@@iDDD == 0 •• Thread, coeffs1@iDD@@1DD, 8i, 4<D

d Sin@2 j1 @tD + j2 @tDD h1@tD h2 @tD d Sin@j2@tDD h1 @tD2 d Sin@2 j1@tD + j2@tDD h1 @tD2
:- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w1 H2 w1 + w2 L 4 w2 H2 w1 + w2 L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ , €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2

d Cos@j2@tDD h2 @tD d Cos@2 j1 @tD + j2 @tDD h2 @tD

2 w2

2 w1 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
w1 w2
d Cos@j2@tDD h1 @tD2 d Cos@2 j1@tD + j2@tDD h1 @tD2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
2 w22 h2@tD 4 w2 H2 w1 + w2L h2 @tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

or in a pure function form as

Chapter5.nb 147

sol1Rule = Table@solVar@1D@@iDD ->

HEvaluate@sol1rhs@@iDD •. Thread@basicTerms -> 8#1, #2, #3, #4<DD &L, 8i, 4<D

i d Sin@#4D #12 d Sin@2 #3 + #4D #12 y

:a1,1 ® i
d Sin@2 #3 + #4D #1 #2 y
€€€€€€€€€€€€€€€€€€ &z, a2,1 ® j
j- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ &z
j €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z,
k 2 w1 H2 w1 + w2 L { k 2 w22 4 w2 H2 w1 + w2L {

i d Cos@#4D #2 d Cos@2 #3 + #4D #2 y i d Cos@#4D #12 d Cos@2 #3 + #4D #12 y

€€€€€€€€€€€€ &z, f2,1 ® j
f1,1 ® j €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ &z
j €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z>
k w1 w2 2 w1 H2 w1 + w2L { k 2 #2 w22 4 #2 w2 H2 w1 + w2 L {

Substituting the slow- and fast-varying components of the first-order solution, SVT1Rule and sol1Rule, into the second-
order equations, we have

Horder2Expr = Table@Subtract žž eqEps@@2, iDD •. SVT1Rule •. sol1Rule •• Expand •• TrigReduce,

8i, 4<D;L •• Timing

85.748 Second, Null<

Since we are seeking an expansion valid up to OHe2 L, we do not need to solve for the a i, 2 and fi, 2 . All we need to do is to
investigate the above expressions to determine the slowly varying parts and hence determine the Ai, 2 and Fi, 2 . The result is

SVT2Rule = Table@
Solve@order2Expr@@iDD == 0 •. Thread@solVar@2D -> H0 &LD •. expRule1 •. f_@_ t + _.D -> 0 •.
expRule2 •. f_@a_D :> f@Expand@aDD, SVT@2D@@iDDD, 8i, 4<D •• Flatten

:A1,2 @h1@tD, h2 @tDD ® 0, A2,2@h1 @tD, h2@tDD ® 0,

8 d2 w21 h1@tD2 + 5 d2 w1 w2 h1@tD2 + 2 d2 w22 h2 @tD2

F1,2 @h1@tD, h2 @tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
8 w21 w22 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ ,

d2 h1 @tD2
F2,2 @h1@tD, h2 @tDD ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ >
4 w1 w2 H2 w1 + w2L

Hence, we find that, to the second approximation, the solution is given by

eq54c =
8u1 @tD, u2@tD< == HNormal@Series@8u1@tD, u2 @tD< •. transformRule •. solRule, 8e, 0, 1<DD •.
sol1Rule •• TrigReduce •• ExpandL •• Thread

2 Sin@j1 @tDD w1 h1 @tD Sin@j1 @tDD w2 h1 @tD d e Cos@j1 @tD - j2 @tDD h1 @tD h2@tD
:u1 @tD == €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w1 H2 w1 + w2L
€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ +

d e Cos@j1 @tD - j2 @tDD h1@tD h2 @tD d e Cos@j1 @tD + j2 @tDD h1@tD h2 @tD
2 w1 + w2 2 w1 + w2

w2 H2 w1 + w2 L w2 H2 w1 + w2 L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ ,

d e w1 h1 @tD2 d e h @tD2 d e Cos@2 j1 @tDD h1@tD2

u2 @tD == €€€€€€€€€€€€€€€€
w2 H2 w1 + w2L 2 w2 H2 w1 + w2 L 4 w2 H2 w1 + w2L
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€1€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ +
2

2 Sin@j2@tDD w1 h2 @tD Sin@j2@tDD w2 h2 @tD

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ >
€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w1 + w2 2 w1 + w2

where the amplitudes and phases are given by

148 Chapter5.nb

ampEqs = basicDRule@@81, 2<DD •. SVT1Rule •. SVT2Rule •. psiRule •.

f_@a_D :> f@Collect@a, tDD •. Rule -> Equal

d e Cos@2 b1 - b2 - t H-2 w1 + w2LD h1 @tD h2@tD

:h¢1 @tD == €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2 w1
d e Cos@2 b1 - b2 - t H-2 w1 + w2 LD h1 @tD2
h¢2 @tD == - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€ >
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 w2

phaseEqs = basicDRule@@83, 4<DD •. SVT1Rule •. SVT2Rule •. psiRule •.

f_@a_D :> f@Collect@a, tDD •. Rule -> Equal

:j¢1 @tD ==

d e Sin@2 b1 - b2 - t H-2 w1 + w2 LD h2@tD e2 H8 d2 w21 h1 @tD2 + 5 d2 w1 w2 h1@tD2 + 2 d2 w22 h2 @tD2L

8 w21 w22 H2 w1 + w2L
w1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ ,
2 w1
d2 e2 h1@tD2 d e Sin@2 b1 - b2 - t H-2 w1 + w2 LD h1@tD2
j¢2 @tD == w2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€ >
4 w1 w2 H2 w1 + w2 L 4 w2 h2@tD
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
Chapter 6
Forced Oscillators of Systems with Finite
Degrees of Freedom
In this chapter, we discuss nonlinear systems having finite degrees of freedom. The discussion is limited to weakly nonlinear
systems, and approximate solutions are obtained by using the method of multiple scales. In the case of strongly nonlinear
systems, perturbation methods can be used in cases for which a basic exact nonlinear solution exists. For the other cases,
recourse is often made to numerical methods and / or geometrical methods to obtain a qualitative description of the behavior
of the system, including its stability.

In contrast with a single-degree-of-freedom system, which has only a single linear natural frequency and a single mode of
oscillation, an n-degree-of-freedom system has n linear natural frequencies and n corresponding modes. Let us denote these
frequencies by w1, w2 , …, wn and assume that all of them are real and different from zero. An important case occurs when-
ever two or more of these frequencies are commensurate or nearly commensurate. Examples of near-commensurability are

8w2 » 2 w1 , w3 » w2 ± w1, w2 » 3 w1 , w3 » 2 w2 ± w1 , w4 » w3 ± w2 ± w1<

Depending on the order of the nonlinearity in the system, these commensurate relationships of frequencies can cause the
corresponding modes to be strongly coupled, and an internal or autoparametric resonance is said to exist. For example, if the
system has quadratic nonlinearities, then to first order an internal resonance may exist if wm » 2 wk or wq » w p ± wm. For a
system with cubic nonlinearities, to first order an internal resonance may exist if wm » 3 wk or wq » w p ± wm ± wk or wm » wk
or wq » 2 w p ± wm. When an internal resonance exists in a free undamped system, energy imparted initially to one of the
modes involved in the internal resonance will be continuously exchanged among the modes involved in that internal reso -
nance. If damping is present in the system, then the energy will decay with time as it is being continuously exchanged.

In a conservative nongyroscopic system, if the linear motion is oscillatory, then the nonlinear motion is bounded. For a
conservative gyroscopic multidegree-of-freedom system, the nonlinear motion may be unbounded and hence unstable if an
internal resonance exists.

If an external harmonic excitation of frequency W acts on a multidegree-of-freedom system, then in addition to all of the
primary and secondary resonances ( p W » q wm , with p and q being integers) of a single-degree-of-freedom system, there
might exist other resonance combinations of the frequencies in the form

p W == a1 w1 + a2 w2 + º + a N w N

where p and the a n are positive or negative integers such that

É p É + Ú Nn=1 É a n É == M

where M is the order of the nonlinearity plus one and N is the number of degrees of freedom. The type of combination
resonance which might exist in a system depends on the order of its nonlinearity. For a system having quadratic nonlineari-
ties, to first order the combination resonances that might exist involve two frequencies in addition to W; that is, W » wm + wk
148 Chapter6.nb

or W » wm - wk . The first of these is called a summed combination resonance or a combination resonance of the additive
type, whereas the second is called a difference combination resonance or a combination resonance of the difference type.
These types of combination resonances were predicted theoretically by Malkin (1956) and found experimentally by Yama-
moto (1957, 1960). For a system having cubic nonlinearities, to first order the combination resonances that might exist involve
either two or three of the natural frequencies in addition to W; that is,

8W » w p ± w m ± w k , W » 2 w p ± w m , W » w p ± 2 w m , 2 W » w m ± w k<

à Preliminaries
Off@General::spell1D

Needs@"Utilities`Notation`"D

To use the method of multiple scales, we introduce different time scales, symbolize them as

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

and form a list of them as follows:

timeScales = 8T0 , T1, T2 <;

In terms of the time scales Tn , the time derivatives become

dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, maxOrder<E;

dt@2D@expr_D := Hdt@1D@dt@1D@exprDD •• ExpandL •. ei_•;i>maxOrder -> 0;

In the course of the analysis, we need the complex conjugates of A and G. We define them using the following rule:

conjugateRule = 9A -> A, A -> A, G -> G, G -> G, Complex@0, n_D -> Complex@0, -nD=;
• • • •

To represent some of the expressions in a more concise way, we introduce the following display rule:

displayRule =
9Derivative@a__DAu_i_,j_ E@__D :> SequenceFormATimes žž MapIndexedAD #1
#2@@1DD-1 &, 8a<E, ui,jE,

Derivative@a__DAA_i_E@__D :> SequenceFormATimes žž MapIndexedAD #1

#2@@1DD &, 8a<E, Ai E,

u_i_,j_@__D :> ui,j , A_i_@__D :> Ai =;

à 6.1 Externally Excited Linearly Uncoupled Systems

Ÿ System of Equations

We consider the response of the following two-degree-of-freedom system with quadratic and cubic geometric nonlineari-
ties to an external (additive) excitation:
Chapter6.nb 149

eq61a = 9
u1 ²@tD + w21 u1 @tD + 2 m1 u1¢ @tD + D@V@u1 @tD, u2 @tDD, u1@tDD == F1 Cos@W t + t1D,
`

u2 ²@tD + w22 u2 @tD + 2 m2 u2¢ @tD + D@V@u1 @tD, u2 @tDD, u2@tDD == F2 Cos@W t + t2D=
`

8w21 u1@tD + 2 m1 u¢1 @tD + u²1 @tD + VH1,0L@u1 @tD, u2 @tDD == Cos@t W + t1D F1 ,
`

w22 u2@tD + 2 m2 u¢2 @tD + u²2 @tD + VH0,1L@u1 @tD, u2 @tDD == Cos@t W + t2D F2 <
`

where the potential function V possesses cubic and quartic nonlinearities. The most general such potential function can be
expressed as follows:

basicModes = 8u1@tD, u2@tD<;

cubicTerms = Nest@Outer@Times, basicModes, #D &, basicModes, 2D •• Flatten •• Union

8u1 @tD3, u1 @tD2 u2 @tD, u1 @tD u2 @tD2 , u2@tD3 <

quarticTerms = Outer@Times, basicModes, cubicTermsD •• Flatten •• Union

8u1 @tD4, u1 @tD3 u2 @tD, u1 @tD2 u2 @tD2, u1 @tD u2@tD3 , u2@tD4 <

potential =
V -> HEvaluate@Sum@di cubicTerms@@iDD, 8i, Length@cubicTermsD<D + Sum@ai quarticTerms@@iDD,
8i, Length@quarticTermsD<D •. Thread@basicModes -> 8#1, #2<DD &L

V ® H#14 a1 + #13 #2 a2 + #12 #22 a3 + #1 #23 a4 + #24 a5 + #13 d1 + #12 #2 d2 + #1 #22 d3 + #23 d4 &L

Approximate solutions of such a system can be obtained using a number of techniques, including the method of normal
forms, the method of multiple scales, the Krylov-Bogoliubov-Mitropolsky technique, and the method of averaging. With
these methods, one determines a set of ordinary-differential equations (modulation equations) that govern the time evolution
of the amplitudes and phases of the modes participating in the response. If these equations are time-dependent, one usually
transforms them into a set of autonomous equations. In this chapter, we describe how the method of multiple scales can be
used to determine uniform expansions of this system, including the modulation equations.

Secondary resonances of nonlinear systems include subharmonic, superharmonic, and combination resonances. One or more
of these resonances might be activated in the presence or absence of internal resonances. When a combination resonance is
activated in the presence of an internal resonance, fractional harmonics might be excited.

two states v1@tD and v2 @tD defined by

For a consistent expansion, we first transform eq61a into a system of four first-order equations. To this end, we introduce the

vel = 8u1 ¢@tD -> v1 @tD, u2¢ @tD -> v2@tD<;

Substituting the velocity and acceleration terms, using vel and potential, into eq61a and combining the result with vel, we
transform eq61a into the following set of four first-order equations:
150 Chapter6.nb

eq61b = 8vel •. Rule -> Equal, eq61a •. D@vel, tD •. vel •. potential< •• Transpose •• Flatten

8u¢1 @tD == v1@tD, w21 u1@tD + 3 d1 u1 @tD2 + 4 a1 u1 @tD3 + 2 d2 u1 @tD u2@tD + 3 a2 u1 @tD2 u2 @tD +
d3 u2@tD2 + 2 a3 u1@tD u2 @tD2 + a4 u2 @tD3 + 2 m1 v1 @tD + v¢1 @tD == Cos@t W + t1 D F1,
`

u¢2 @tD == v2@tD, d2 u1@tD2 + a2 u1@tD3 + w22 u2 @tD + 2 d3 u1 @tD u2@tD + 2 a3 u1 @tD2 u2 @tD +
3 d4 u2@tD2 + 3 a4 u1@tD u2 @tD2 + 4 a5 u2 @tD3 + 2 m2 v2 @tD + v¢2 @tD == Cos@t W + t2D F2 <
`

We seek a uniform second-order expansion of the solution of eq61b in the form

solRule =
9ui_ -> ISumAej ui,j@#1, #2, #3D, 8j, 3<E &M, vi_ -> ISumAej vi,j@#1, #2, #3D, 8j, 3<E &M=;

and set

maxOrder = 2;

The damping and forcing terms in eq61b need to be scaled so that they balance the influence of the nonlinearity. The scaling
of the forcing depends on the type of resonance. Next, we treat different resonance cases in the following sections.

Ÿ 6.1.1 W » w2 and w2 » 2 w1

In this case, we have a combination of a primary resonance of the second mode, a subharmonic resonance of order one-half
of the first mode, and a two-to-one internal resonance. In order to bring the effects of damping, forcing, and nonlinearity at
the same order, we let

scaling = 9m̀ n_ -> e m n , F1 -> e f1, F2 -> e2 f2 =;

To describe quantitatively the nearness of the resonances, we introduce the two detunings s1 and s2 defined by

ResonanceCond = 8w2 == 2 w1 + e s1 , W == w2 + e s2 <;

Then, we define the following lists:

omgList = 8w1 , w2<;

omgRule = Solve@ResonanceCond, 8W, #< •• FlattenD@@1DD & •ž omgList •• Reverse

98W ® e s1 + e s2 + 2 w1 , w2 ® e s1 + 2 w1 <, 9W ® e s2 + w2 , w1 ® €€€€ H-e s1 + w2 L==

1
2

Using the time scales T0, T1 , and T2, we express the time derivative as

multiScales = 8ui_@tD -> ui žž timeScales, vi_ @tD -> vi žž timeScales,

Derivative@1D@u_D@tD -> dt@1D@u žž timeScalesD, t -> T0<;

Substituting the scaling, multiScales, and solRule into eq61b, expanding the result for small e, and discarding terms of
order higher than e3 , we obtain
Chapter6.nb 151

eq611a = Heq61b •. scaling •. multiScales •. solRule •• TrigToExp •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e, we obtain

eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eq611a •• Transpose;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •. u_ k_,1 -> u k,iM ==

I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •. u_ k_,1 -> u k,i M - H#@@1DD & •ž eqEps@@iDDL •• Thread

Using the eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

eqOrder@1D •. displayRule
eqOrder@2D •. displayRule
eqOrder@3D •. displayRule

9D0 u1,1 - v1,1 == 0, D0 v1,1 + w21 u1,1 == €€€€ E-I T0 W-I t1 f1 + €€€€ EI T0 W+I t1 f1 ,
1 1
2 2
2
D0 u2,1 - v2,1 == 0, D0 v2,1 + w2 u2,1 == 0=

9D0 u1,2 - v1,2 == -HD1u1,1 L, D0v1,2 + w21 u1,2 == -HD1v1,1 L - 3 d1 u21,1 - 2 d2 u1,1 u2,1 - d3 u22,1 - 2 m1 v1,1 ,
D0 u2,2 - v2,2 == -HD1u2,1 L, D0v2,2 + w22 u2,2 ==

-HD1v2,1 L + €€€€ E-I T0 W-It2 f2 + €€€€ EI T0 W+I t2 f2 - d2 u21,1 - 2 d3 u1,1 u2,1 - 3 d4 u22,1 - 2 m2 v2,1 =
1 1
2 2

8D0 u1,3 - v1,3 == -HD1u1,2 L - D2 u1,1,

D0 v1,3 + w21 u1,3 == -HD1 v1,2L - D2v1,1 - 4 a1 u31,1 - 6 d1 u1,1 u1,2 - 3 a2 u21,1 u2,1 -
2 d2 u1,2 u2,1 - 2 a3 u1,1 u22,1 - a4 u32,1 - 2 d2 u1,1 u2,2 - 2 d3 u2,1 u2,2 - 2 m1 v1,2,
D0 u2,3 - v2,3 == -HD1u2,2 L - D2 u2,1, D0 v2,3 + w22 u2,3 == -HD1 v2,2L - D2v2,1 - a2 u31,1 - 2 d2 u1,1 u1,2 -
2 a3 u21,1 u2,1 - 2 d3 u1,2 u2,1 - 3 a4 u1,1 u22,1 - 4 a5 u32,1 - 2 d3 u1,1 u2,2 - 6 d4 u2,1 u2,2 - 2 m2 v2,2 <

Ÿ First-Order Problem: Linear System

The first-order problem, eqOrder[1], consists of two sets of uncoupled linear nonhomogeneous differential equations.
Hence, the general solution of each set can be obtained by using the principle of superposition as the sum of any particular
solution and the solution of the homogeneous equations. To determine the general solution of the homogeneous sets, we
rewrite the homogeneous part of eqOrder[1] as

linearSys = #@@1DD & •ž eqOrder@1D;

linearSys •. displayRule

8D0 u1,1 - v1,1, D0 v1,1 + w21 u1,1 , D0u2,1 - v2,1 , D0v2,1 + w22 u2,1<

Next, we seek a solution of the linearSys in the form u i,1 = PEI wi T 0 and vi,1 = Q EI wi T 0 and obtain
152 Chapter6.nb

expr1 =
linearSys •. 9ui_,1 -> IP EI wi #1 &M, vi_,1 -> IQ EI wi #1 &M= •. Exp@a_D -> 1 •• Partition@#, 2D &

88-Q + I P w1 , I Q w1 + P w21<, 8-Q + I P w2, I Q w2 + P w22 <<

The coefficient matrices of expr1 are

coefMat = Outer@D, #, 8P, Q<D & •ž expr1

888I w1 , -1<, 8w21, I w1 <<, 88I w2, -1<, 8w22 , I w2<<<

and their adjoints are defined by

hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose

Hence, the right and left eigenvectors of coefMat are

rightVec = # • #@@1DD & •ž HNullSpace@#D@@1DD & •ž coefMatL

881, I w1 <, 81, I w2<<

leftVec = NullSpace@hermitian@#DD@@1DD & •ž coefMat

88-I w1, 1<, 8-I w2 , 1<<

whose complex conjugates are

ccleftVec = leftVec •. conjugateRule

88I w1 , 1<, 8I w2, 1<<

To determine particular solutions of the partial-differential equations eqOrder[1] using the Mathematica function DSolve, we
first transform them into a set of ordinary-differential equations as

order1Eq = eqOrder@1D •. u_i_,1 -> Hui,1@#1D &L

9-v1,1@T0 D + u¢1,1 @T0D == 0, w21 u1,1@T0 D + v¢1,1 @T0D == €€€€ E-I T0 W-I t1 f1 + €€€€ EI T0 W+I t1 f1,
1 1

-v2,1@T0 D + u2,1 @T0D == 0, w2 u2,1@T0 D + v2,1 @T0D == 0=

2 2
¢ 2 ¢

Then, the particular solutions of order1Eq are

sol1p =
DSolve@order1Eq, 8u1,1@T0 D, v1,1@T0 D, u2,1@T0 D, v2,1@T0 D<, T0 D@@1DD •. C@_D -> 0 •• Simplify

E-I HT0 W+t1 L H1 + E2 I HT0 W+t1 LL f1

:u1,1 @T0D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 HW2 - w21L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

I E-I HT0 W+t1 L H-1 + E2I HT0 W+t1L L W f1

v1,1 @T0D ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ , u2,1@T0 D ® 0, v2,1@T0 D ® 0>
2 HW2 - w21 L
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
Chapter6.nb 153

To simplify some of the expressions, we let

fRule = 9fi_ -> 2 Li Iw2i - W2 M=;

and express u i,1 and vi,1 in pure function form as

sol1u = TableAui,1 -> FunctionA8T0 , T1, T2 <, Ai @T1, T2 D Exp@I wi T0 D + Ai @T1, T2 D Exp@-I wi T0 D +
•

Hui,1@T0 D •. sol1p •. fRule •• Simplify •• ExpandL •• EvaluateE, 8i, 2<E

8u1,1 ® Function@8T0 , T1 , T2<, E-I T0 W-It1 L1 + EI T0 W+I t1 L1 + EI T0 w1 A1 @T1 , T2D + E-I T0 w1 A1 @T1, T2 DD,
•

u2,1 ® Function@8T0 , T1 , T2<, EI T0 w2 A2@T1 , T2 D + E-I T0 w2 A2@T1 , T2DD<

•

sol1v =
Table@vi,1 -> Function@8T0 , T1, T2<, D@ui,1 žž timeScales, T0D •. sol1u •• EvaluateD, 8i, 2<D

8v1,1 ® Function@8T0 , T1 , T2<,

-I E-I T0 W-I t1 W L1 + I EI T0 W+I t1 W L1 + I EI T0 w1 w1 A1 @T1, T2 D - I E-I T0 w1 w1 A1 @T1, T2 DD,
•

v2,1 ® Function@8T0 , T1 , T2<, I EI T0 w2 w2 A2 @T1, T2D - I E-I T0 w2 w2 A2 @T1 , T2DD<

•

Then, we combine them using the Mathematica function Join as

sol1 = Join@sol1u, sol1vD;

Ÿ Second-Order Problem

Substituting the first-order solution sol1 into the second-order problem, eqOrder[2], yields

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

9D0 u1,2 - v1,2 == -EI T0 w1 HD1 A1L - E-I T0 w1 HD1 A1L,

•

D0 v1,2 + w21 u1,2 == -3 E2 I T0 w1 A21 d1 - 2 EI T0 w1 +I T0 w2 A1 A2 d2 - E2I T0 w2 A22 d3 -

6 E-I T0 W-I t1 +IT0 w1 A1 d1 L1 - 6 EI T0 W+I t1 +IT0 w1 A1 d1 L1 - 2 E-I T0 W-I t1 +IT0 w2 A2 d2 L1 -
2 EI T0 W+I t1 +IT0 w2 A2 d2 L1 - 6 d1 L21 - 3 E-2I T0 W-2 It1 d1 L21 - 3 E2 I T0 W+2 I t1 d1 L21 +
2 I E-I T0 W-I t1 W L1 m1 - 2 I EI T0 W+I t1 W L1 m1 - I EI T0 w1 HD1A1 L w1 + I E-IT0 w1 HD1 A1L w1 -
•
• • •
2 I EI T0 w1 A1 m1 w1 - 6 A1 d1 A1 - 2 E-I T0 w1+I T0 w2 A2 d2 A1 - 6 E-I T0 W-I t1-I T0 w1 d1 L1 A1 -
• • •2 • •
6 EI T0 W+I t1 -IT0 w1 d1 L1 A1 + 2 I E-I T0 w1 m1 w1 A1 - 3 E-2 I T0 w1 d1 A1 - 2 EI T0 w1-I T0 w2 A1 d2 A2 - 2 A2 d3 A2 -
• • • • •2
2 E-I T0 W-I t1 -IT0 w2 d2 L1 A2 - 2 EI T0 W+I t1 -IT0 w2 d2 L1 A2 - 2 E-I T0 w1 -I T0 w2 d2 A1 A2 - E-2 I T0 w2 d3 A2 ,
D0 u2,2 - v2,2 == -EI T0 w2 HD1 A2L - E-I T0 w2 HD1 A2L,
•

1 1
D0 v2,2 + w22 u2,2 == €€€€ E-I T0 W-I t2 f2 + €€€€ EI T0 W+I t2 f2 - E2 I T0 w1 A21 d2 - 2 EI T0 w1+I T0 w2 A1 A2 d3 -
2 2
2 I T0 w2 2 -I T0 W-I t1+I T0 w1
3E A2 d4 - 2 E A1 d2 L1 - 2 EI T0 W+I t1 +I T0 w1 A1 d2 L1 -
-I T0 W-I t1 +IT0 w2 I T0 W+I t1 +IT0 w2
A2 d3 L1 - 2 d2 L21 - E-2 IT0 W-2 I t1 d2 L21 -
E2 I T0 W+2 I t1 d2 L21 - I EI T0 w2 HD1 A2L w2 + I E-I T0 w2 HD1A2 L w2 - 2 I EI T0 w2 A2 m2 w2 -
2E A2 d3 L1 - 2 E
•
• • • •
2 A1 d2 A1 - 2 E-I T0 w1 +I T0 w2 A2 d3 A1 - 2 E-I T0 W-I t1 -I T0 w1 d2 L1 A1 - 2 EI T0 W+It1 -I T0 w1 d2 L1 A1 -
•2 • • •
E-2 I T0 w1 d2 A1 - 2 EI T0 w1 -I T0 w2 A1 d3 A2 - 6 A2 d4 A2 - 2 E-I T0 W-I t1 -I T0 w2 d3 L1 A2 -
2 EI T0 W+I t1 -IT0 w2 d3 L1 A2 + 2 I E-I T0 w2 m2 w2 A2 - 2 E-I T0 w1-I T0 w2 d3 A1 A2 - 3 E-2 IT0 w2 d4 A2 =
• • • • •2
154 Chapter6.nb

Any particular solution of order2Eq contains secular terms and small-divisor terms, making the expansion nonuniform. For a
uniform expansion, we choose D1 A1 and D1 A2 to eliminate these terms. To accomplish this, we first convert the small-
divisor terms into secular terms using the rule

expRule1@i_D := Exp@a_D :> Exp@Expand@a •. omgRule@@iDDD •. e T0 -> T1 D

To eliminate the terms that produce secular terms (i.e., determine the solvability conditions) from order2Eq, we collect the
terms proportional to EI w1 T 0 and EI w2 T 0 and obtain

ST11 = Coefficient@order2Eq@@#, 2DD •. expRule1@1D, Exp@I w1 T0DD & •ž 81, 2<;

ST11 •. displayRule

8-HD1A1 L, -I HD1A1 L w1 - 2 I A1 m1 w1 - 2 EI T1 s1 A2 d2 A1 - 6 EI T1 s1+I T1 s2 +I t1 d1 L1 A1<

• •

ST12 = Coefficient@order2Eq@@#, 2DD •. expRule1@2D, Exp@I w2 T0DD & •ž 83, 4<;

ST12 •. displayRule

9-HD1A2 L, €€€€ EI T1 s2+I t2 f2 - E-I T1 s1 A21 d2 - I HD1 A2L w2 - 2 I A2 m2 w2 =

1
2

Then, the solvability conditions, conditions for the elimination of the terms that produce secular terms, demand that ST11
and ST12 be orthogonal to every solution of the corresponding adjoint problem, namely, the components of ccleftVec. The
result is

SCond1 = 8ccleftVec@@1DD.ST11 == 0, ccleftVec@@2DD.ST12 == 0<;

SCond1 •. displayRule

9-2 I HD1A1 L w1 - 2 I A1 m1 w1 - 2 EI T1 s1 A2 d2 A1 - 6 EI T1 s1+I T1 s2 +I t1 d1 L1 A1 == 0,

• •

€€€€ EI T1 s2+I t2 f2 - E-I T1 s1 A21 d2 - 2 I HD1 A2L w2 - 2 I A2 m2 w2 == 0=

1
2

Solving SCond1 for D1 A1 and D1 A2 yields

H1,0L H1,0L
SCond1Rule1 = SolveASCond1, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD •• ExpandAll;
SCond1Rule1 •. displayRule

:D1 A1 ® -A1 m1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• •
I EI T1 s1 A2 d2 A1 3 I EI T1 s1 +IT1 s2+I t1 d1 L1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
w1 w1

D1 A2 ® -A2 m2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€ >

I EI T1 s2 +I t2 f2 I E-I T1 s1 A2 d2
4 w2 2 w2

that they be orthogonal to solutions of the corresponding adjoint problems; that is, we damand that 8u 1,2 , v1,2 < be orthogo-
With these conditions, order2Eq are solvable. However, their solutions are not unique. To render them unique, we demand

nal to ccleftVec[[1]] and 8u 2,2 , v2,2 < be orthogonal to ccleftVec[[2]].

Next, we use the solvability conditions to eliminate D1 A1 and D1 A2 from order2Eq. To simplify the resulting equations using
Mathematica, we replace the scale T1 in SCond1Rule1 with e T0. To this end, we express the detuning parameters si in terms
of the wi and W as
Chapter6.nb 155

sigRule = Solve@ResonanceCond, 8s1, s2 <D@@1DD

9s1 ® - €€€€€€€€€€€€€€€€€€€€ , s2 ® - €€€€€€€€€€€€€€€€€ =

2 w1 - w2 -W + w2
e e

and replace T1 with e T0 using the rule

expRule2 = Exp@a_D :> Exp@a •. T1 -> e T0 •. sigRule •• ExpandD;

Using the expRule2, we rewrite SCond1Rule1 as

SCond1Rule2 = SCond1Rule1 •. expRule2;

SCond1Rule2 •. displayRule

:D1 A1 ® -A1 m1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• •
I E-2 I T0 w1 +I T0 w2 A2 d2 A1 3 I EI T0 W+I t1 -2 I T0 w1 d1 L1 A1
€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
w1 w1

€€€€€€€€€€€€€ >
I EI T0 W+I t2 -I T0 w2 f2 I E2 I T0 w1-I T0 w2 A21 d2
D1 A2 ® -A2 m2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 w2 2 w2

whose complex conjugate is

ccSCond1Rule2 = SCond1Rule2 •. conjugateRule;

Substituting this form for the solvability conditions into order2Eq yields
156 Chapter6.nb

order2Eqm = order2Eq •. SCond1Rule2 •. ccSCond1Rule2 •• ExpandAll;

order2Eqm •. displayRule

9D0 u1,2 - v1,2 == EI T0 w1 A1 m1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

3 I E-IT0 W-I t1+I T0 w1 A1 d1 L1 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + E-I T0 w1 m1 A1 -
w1
-IT0 w1+I T0 w2 • I T0 W+I t1 -IT0 w1 • •
IE A2 d2 A1 3IE d1 L1 A1 I EI T0 w1-I T0 w2 A1 d2 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ ,
w1 w1 w1
D0 v1,2 + w21 u1,2 == -3 E2 I T0 w1 A21 d1 - 2 EI T0 w1 +I T0 w2 A1 A2 d2 - E2I T0 w2 A22 d3 - 3 E-I T0 W-It1 +I T0 w1 A1 d1 L1 -
6 EI T0 W+I t1 +IT0 w1 A1 d1 L1 - 2 E-I T0 W-I t1 +IT0 w2 A2 d2 L1 - 2 EI T0 W+I t1 +I T0 w2 A2 d2 L1 - 6 d1 L21 -
3 E-2 I T0 W-2 I t1 d1 L21 - 3 E2 I T0 W+2I t1 d1 L21 + 2 I E-I T0 W-I t1 W L1 m1 - 2 I EI T0 W+I t1 W L1 m1 -
• • •
I EI T0 w1 A1 m1 w1 - 6 A1 d1 A1 - E-IT0 w1+I T0 w2 A2 d2 A1 - 6 E-I T0 W-I t1-I T0 w1 d1 L1 A1 -
• • •2 • •
3 EI T0 W+I t1 -IT0 w1 d1 L1 A1 + I E-I T0 w1 m1 w1 A1 - 3 E-2 I T0 w1 d1 A1 - EI T0 w1-I T0 w2 A1 d2 A2 - 2 A2 d3 A2 -
• • • • •2
2 E-I T0 W-I t1 -IT0 w2 d2 L1 A2 - 2 EI T0 W+I t1 -IT0 w2 d2 L1 A2 - 2 E-I T0 w1 -I T0 w2 d2 A1 A2 - E-2 I T0 w2 d3 A2 ,
I E-IT0 W-I t2 f2 I EI T0 W+I t2 f2
D0 u2,2 - v2,2 == EI T0 w2 A2 m2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
4 w2 4 w2
2 IT0 w1 2 -2 IT0 w1 •2
IE A d2 IE d2 A •
€€€€€€€€€€€€€€€€€€€€€€€€€1€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€€ + E-I T0 w2 m2 A2 ,
2 w2 2 w2
1 1 1
D0 v2,2 + w22 u2,2 == €€€€ E-I T0 W-I t2 f2 + €€€€ EI T0 W+I t2 f2 - €€€€ E2I T0 w1 A21 d2 - 2 EI T0 w1+I T0 w2 A1 A2 d3 -
4 4 2
3 E2 I T0 w2 A22 d4 - 2 E-I T0 W-I t1+I T0 w1 A1 d2 L1 - 2 EI T0 W+I t1 +I T0 w1 A1 d2 L1 - 2 E-I T0 W-I t1 +I T0 w2 A2 d3 L1 -
2 EI T0 W+I t1 +IT0 w2 A2 d3 L1 - 2 d2 L21 - E-2 IT0 W-2 I t1 d2 L21 - E2 I T0 W+2 I t1 d2 L21 - I EI T0 w2 A2 m2 w2 -
• • • •
2 A1 d2 A1 - 2 E-I T0 w1 +I T0 w2 A2 d3 A1 - 2 E-I T0 W-I t1 -I T0 w1 d2 L1 A1 - 2 EI T0 W+It1 -I T0 w1 d2 L1 A1 -
1 •2 • • •
€€€€ E-2 I T0 w1 d2 A1 - 2 EI T0 w1 -I T0 w2 A1 d3 A2 - 6 A2 d4 A2 - 2 E-I T0 W-It1 -I T0 w2 d3 L1 A2 -

2 EI T0 W+I t1 -IT0 w2 d3 L1 A2 + I E-I T0 w2 m2 w2 A2 - 2 E-I T0 w1-I T0 w2 d3 A1 A2 - 3 E-2 IT0 w2 d4 A2=

2
• • • • •2

Next, we use the method of undetermined coefficients to determine the particular solutions of order2Eqm. To accomplish
this, we first identify the form of the nonhomogeneous terms. To this end, we let

basicH = TableA9Ai@T1 , T2 D EIwi T0 , Ai @T1, T2 D E-I wi T0 =, 8i, 2<E •• Flatten

•

8EI T0 w1 A1@T1 , T2D, E-I T0 w1 A1 @T1 , T2 D, EI T0 w2 A2 @T1, T2D, E-I T0 w2 A2@T1 , T2D<
• •

collectForm = JoinAbasicH, 9L1 EI W T0 +I t1 , L1 E-I W T0-I t1 =E

8EI T0 w1 A1@T1 , T2D, E-I T0 w1 A1 @T1 , T2 D,

•

EI T0 w2 A2@T1 , T2D, E-I T0 w2 A2 @T1 , T2 D, EI T0 W+I t1 L1 , E-I T0 W-I t1 L1<

•

Then, the possible form of the nonhomogeneous terms in order2Eqm is

Chapter6.nb 157

possibleTerms = JoinAcollectForm, 9EIW T0 +I t2 , E-I W T0-I t2 =,

Outer@Times, collectForm, collectFormD •• Flatten •• UnionE

:EI T0 w1 A1@T1 , T2D, E-I T0 w1 A1 @T1 , T2 D, EI T0 w2 A2 @T1, T2D, E-I T0 w2 A2@T1 , T2D,
• •

EI T0 W+I t1 L1 , E-I T0 W-I t1 L1 , EI T0 W+I t2 , E-I T0 W-I t2 , L21, E-2 I T0 W-2 I t1 L21 , E2 I T0 W+2 I t1 L21 ,
E-I T0 W-I t1 +I T0 w1 L1 A1@T1 , T2D, EIT0 W+I t1+I T0 w1 L1 A1 @T1, T2 D, E2 I T0 w1 A1 @T1, T2D2 ,
E-I T0 W-I t1 +I T0 w2 L1 A2@T1 , T2D, EIT0 W+I t1+I T0 w2 L1 A2 @T1, T2 D, EI T0 w1 +I T0 w2 A1 @T1, T2 D A2 @T1 , T2 D,
E2 I T0 w2 A2@T1 , T2D2 , E-I T0 W-I t1 -IT0 w1 L1 A1 @T1 , T2D, EI T0 W+It1 -I T0 w1 L1 A1@T1 , T2 D,
• •

A1 @T1, T2 D A1@T1 , T2 D, E-I T0 w1 +IT0 w2 A2 @T1, T2 D A1@T1 , T2 D, E-2I T0 w1 A1@T1 , T2 D ,

• • • 2

E-I T0 W-I t1 -I T0 w2 L1 A2@T1 , T2D, EIT0 W+I t1-I T0 w2 L1 A2 @T1, T2 D, EI T0 w1 -I T0 w2 A1 @T1, T2 D A2 @T1 , T2 D,
• • •

A2 @T1, T2 D A2@T1 , T2 D, E-I T0 w1 -IT0 w2 A1 @T1, T2 D A2@T1 , T2 D, E-2I T0 w2 A2@T1 , T2 D >

• • • • 2

Using the method of undetermined coefficients, we assume that the ui,2 and vi,2 are linear combinations of these possible
terms, substitute the result into the governing equations, order2Eqm, equate the coefficients of each of the possible terms on
both sides of each equation, and obtain a system of nonhomogeneous algebraic equations governing the unknown coeffi-
cients. Associated with each possible term is a pair of two algebraic equations. The solutions of all pairs, except those
associated with resonance terms, are unique. The solutions of the pairs corresponding to the possible resonance terms are
rendered unique by requiring them to be orthogonal to the components of ccleftVec. The possible resonance terms are
proportional to EI w1 T 0 , EI w2 T 0 , and their complex conjugates; that is,

ResonantTerms@i_D :=
I# •. 8a_ •; a =!= 0 -> 1< & •ž IE-I wi T0 possibleTerms •. expRule1@iD •. Exp@_ T0 + _.D -> 0MM
possibleTerms •• Union •• Rest

RT = Array@ResonantTerms, 2D

88EI T0 w1 A1@T1 , T2D, EIT0 W+I t1-I T0 w1 L1 A1 @T1, T2 D, E-I T0 w1 +I T0 w2 A2 @T1, T2 D A1@T1 , T2 D<,
• •

8EI T0 W+I t2 , EIT0 W+I t1 L1, E2 I T0 w1 A1@T1 , T2 D2, EI T0 w2 A2@T1 , T2 D<<

Hence, the coefficients of RT in the nonhomogeneous terms in order2Eqm[[1]] are

r1Rule =
MapIndexed@r1,#2@@1DD -> Coefficient@order2Eqm@@1, 2DD, #1D &, RT@@1DDD •. Exp@_ T0 + _.D -> 0

:r1,1 ® m1, r1,2 ® - €€€€€€€€€€€€€1€ , r1,3 ® - €€€€€€€€2€€ >

3Id Id
w1 w1

and the coefficients of RT in the nonhomogeneous terms in order2Eqm[[2]] are

r2Rule =
MapIndexed@r2,#2@@1DD -> Coefficient@order2Eqm@@2, 2DD, #1D &, RT@@1DDD •. Exp@_ T0 + _.D -> 0

8r2,1 ® -I m1 w1, r2,2 ® -3 d1 , r2,3 ® -d2 <

It follows from
158 Chapter6.nb

Table@ccleftVec@@1DD.8r1,i , r2,i <, 8i, 3<D •. r1Rule •. r2Rule

80, 0, 0<

that the parts r1Rule and r2Rule are orthogonal to the solution of the adjoint, the first component of ccleftVec, as they
should.

Similarly, the coefficients of RT in the nonhomogeneous terms in order2Eqm[[3]] are

r3Rule = MapIndexed@r3,#2@@1DD -> Coefficient@order2Eqm@@3, 2DD, #1D &, RT@@2DDD

:r3,1 ® €€€€€€€€€2€ , r3,2 ® 0, r3,3 ® - €€€€€€€€2€€ , r3,4 ® m2 >

If Id
4 w2 2 w2

and the coefficients of RT in the nonhomogeneous terms in order2Eqm[[4]] are

r4Rule =
MapIndexed@r4,#2@@1DD -> Coefficient@order2Eqm@@4, 2DD, #1D &, RT@@2DDD •. Exp@_ T0 + _.D -> 0

9r4,1 ® €€€€2€€ , r4,2 ® 0, r4,3 ® - €€€€€€ , r4,4 ® -I m2 w2=

f d2
4 2

Again, it follows from

Table@ccleftVec@@2DD.8r3,i , r4,i <, 8i, 4<D •. r3Rule •. r4Rule

80, 0, 0, 0<

that the vector 8r3, i , r4, i< is orthogonal to the solution of the adjoint, the second component of ccleftVec, as it should.

The complement of the resonance terms RT in possibleTerms yields the nonresonance terms; that is,

NRT = Complement@possibleTerms, Join@#, # •. conjugateRuleDD & •ž RT

::E-I T0 W-I t2 , EI T0 W+I t2 , E-I T0 W-I t1 L1 , EI T0 W+I t1 L1 , L21, E-2 I T0 W-2 I t1 L21 ,

E2 I T0 W+2 I t1 L21 , EI T0 W+I t1+I T0 w1 L1 A1 @T1, T2D, E2 I T0 w1 A1@T1 , T2D2 , EI T0 w2 A2@T1 , T2D,
E-I T0 W-I t1 +I T0 w2 L1 A2@T1 , T2D, EIT0 W+I t1+I T0 w2 L1 A2 @T1, T2 D, EI T0 w1 +I T0 w2 A1 @T1, T2 D A2 @T1 , T2 D,
E2 I T0 w2 A2@T1 , T2D2 , E-I T0 W-I t1 -I T0 w1 L1 A1 @T1 , T2D, A1 @T1, T2D A1 @T1, T2 D, E-2 I T0 w1 A1 @T1, T2 D ,
• • • 2

E-I T0 w2 A2@T1 , T2D, E-I T0 W-It1 -I T0 w2 L1 A2@T1 , T2 D, EI T0 W+I t1-I T0 w2 L1 A2 @T1, T2D,
• • •

A2 @T1, T2 D A2@T1 , T2 D, E-I T0 w1 -IT0 w2 A1 @T1, T2 D A2@T1 , T2D, E-2I T0 w2 A2@T1 , T2 D >,
• • • • 2

:L21 , E-2 I T0 W-2 I t1 L21, E2 I T0 W+2I t1 L21 , EI T0 w1 A1@T1 , T2D, E-I T0 W-I t1+I T0 w1 L1 A1 @T1, T2 D,
EI T0 W+I t1 +I T0 w1 L1 A1@T1 , T2D, E-IT0 W-I t1+I T0 w2 L1 A2 @T1, T2 D, EI T0 W+I t1 +I T0 w2 L1 A2 @T1 , T2D,
EI T0 w1 +I T0 w2 A1 @T1 , T2D A2 @T1, T2 D, E2 I T0 w2 A2 @T1 , T2D2 , E-I T0 w1 A1 @T1, T2D,
•

E-I T0 W-I t1 -I T0 w1 L1 A1@T1 , T2D, EIT0 W+I t1-I T0 w1 L1 A1 @T1, T2 D, A1 @T1 , T2 D A1@T1 , T2D,
• • •

E-I T0 w1 +I T0 w2 A2 @T1, T2D A1 @T1, T2 D, E-I T0 W-I t1 -I T0 w2 L1 A2 @T1 , T2D,

• •

EI T0 W+I t1 -I T0 w2 L1 A2@T1 , T2D, EIT0 w1-I T0 w2 A1@T1 , T2 D A2@T1 , T2D,

• •

A2 @T1, T2 D A2@T1 , T2 D, E-I T0 w1 -IT0 w2 A1 @T1, T2 D A2@T1 , T2D, E-2I T0 w2 A2@T1 , T2 D >>
• • • • 2
Chapter6.nb 159

To implement the method of undetermined coefficients, we associate with each possible resonance term an unknown coeffi-
cient using the rules

RTsymbolList1 = Table@G1,j, 8j, Length@RT@@1DDD<D

8G1,1 , G1,2, G1,3 <

RTsymbolList2 = Table@G3,j, 8j, Length@RT@@2DDD<D

8G3,1 , G3,2, G3,3 , G3,4<

And, we associate with each possible nonresonance term two unknown coefficients using the rules

NRTsymbolList1@i_D = Table@Li,j , 8j, Length@NRT@@1DDD<D

8Li,1 , Li,2, Li,3 , Li,4, Li,5 , Li,6, Li,7 , Li,8, Li,9 , Li,10, Li,11 ,
Li,12 , Li,13 , Li,14, Li,15 , Li,16, Li,17 , Li,18, Li,19 , Li,20, Li,21 , Li,22, Li,23 <

NRTsymbolList2@i_D = Table@Li+2,j , 8j, Length@NRT@@2DDD<D

8L2+i,1, L2+i,2 , L2+i,3, L2+i,4 , L2+i,5 , L2+i,6, L2+i,7 , L2+i,8, L2+i,9 , L2+i,10 , L2+i,11,
L2+i,12, L2+i,13 , L2+i,14, L2+i,15, L2+i,16 , L2+i,17 , L2+i,18, L2+i,19 , L2+i,20 , L2+i,21<

Accounting for the orthogonality conditions, 8u 1,2 , v1,2 < must be orthogonal to ccleftVec[[1]] and 8u 2,2 , v2,2 < must be
orthogonal to ccleftVec[[2]], for the resonance terms, RT, we express u i,2 and vi,2 in the following function form:

8ratio1, ratio2< = Table@-ccleftVec@@i, 1DD, 8i, 2<D

8-I w1, -I w2 <

sol2Form = 8u1,2 -> Function@8T0 , T1, T2 <,

RTsymbolList1.RT@@1DD + HRTsymbolList1.RT@@1DD •. conjugateRuleL +
[email protected]@@1DD •• EvaluateD, v1,2 -> Function@8T0 , T1 , T2<,
ratio1 * RTsymbolList1.RT@@1DD + Hratio1 * RTsymbolList1.RT@@1DD •. conjugateRuleL +
[email protected]@@1DD •• EvaluateD, u2,2 -> Function@8T0 , T1 , T2<,
RTsymbolList2.RT@@2DD + HRTsymbolList2.RT@@2DD •. conjugateRuleL +
[email protected]@@2DD •• EvaluateD, v2,2 -> Function@8T0 , T1 , T2<,
ratio2 * RTsymbolList2.RT@@2DD + Hratio2 * RTsymbolList2.RT@@2DD •. conjugateRuleL +
[email protected]@@2DD •• EvaluateD<;

Substituting sol2Form into order2Eqm and collecting the coefficients of like NRT, we obtain two sets of algebraic equations
as
160 Chapter6.nb

eq611b = HCoefficient@Subtract žž # •. sol2Form, NRT@@1DDD •. Exp@_ T0 + _.D -> 0 & •ž

order2Eqm@@81, 2<DD •• FlattenL == 0 •• Thread

8-I W L1,1 - L2,1 == 0, I W L1,2 - L2,2 == 0, -I W L1,3 - L2,3 == 0, I W L1,4 - L2,4 == 0, -L2,5 == 0,
-2 I W L1,6 - L2,6 == 0, 2 I W L1,7 - L2,7 == 0, I W L1,8 + I w1 L1,8 - L2,8 == 0, 2 I w1 L1,9 - L2,9 == 0,
I w2 L1,10 - L2,10 == 0, -I W L1,11 + I w2 L1,11 - L2,11 == 0, I W L1,12 + I w2 L1,12 - L2,12 == 0,
I w1 L1,13 + I w2 L1,13 - L2,13 == 0, 2 I w2 L1,14 - L2,14 == 0, -I W L1,15 - I w1 L1,15 - L2,15 == 0,
-L2,16 == 0, -2 I w1 L1,17 - L2,17 == 0, -I w2 L1,18 - L2,18 == 0, -I W L1,19 - I w2 L1,19 - L2,19 == 0,
I W L1,20 - I w2 L1,20 - L2,20 == 0, -L2,21 == 0, -I w1 L1,22 - I w2 L1,22 - L2,22 == 0,
-2 I w2 L1,23 - L2,23 == 0, w21 L1,1 - I W L2,1 == 0, w21 L1,2 + I W L2,2 == 0,
-2 I W m1 + w21 L1,3 - I W L2,3 == 0, 2 I W m1 + w21 L1,4 + I W L2,4 == 0, 6 d1 + w21 L1,5 == 0,
3 d1 + w21 L1,6 - 2 I W L2,6 == 0, 3 d1 + w21 L1,7 + 2 I W L2,7 == 0, 6 d1 + w21 L1,8 + I W L2,8 + I w1 L2,8 == 0,
3 d1 + w21 L1,9 + 2 I w1 L2,9 == 0, w21 L1,10 + I w2 L2,10 == 0, 2 d2 + w21 L1,11 - I W L2,11 + I w2 L2,11 == 0,
2 d2 + w21 L1,12 + I W L2,12 + I w2 L2,12 == 0, 2 d2 + w21 L1,13 + I w1 L2,13 + I w2 L2,13 == 0,
d3 + w21 L1,14 + 2 I w2 L2,14 == 0, 6 d1 + w21 L1,15 - I W L2,15 - I w1 L2,15 == 0,
6 d1 + w21 L1,16 == 0, 3 d1 + w21 L1,17 - 2 I w1 L2,17 == 0, w21 L1,18 - I w2 L2,18 == 0,
2 d2 + w21 L1,19 - I W L2,19 - I w2 L2,19 == 0, 2 d2 + w21 L1,20 + I W L2,20 - I w2 L2,20 == 0,
2 d3 + w21 L1,21 == 0, 2 d2 + w21 L1,22 - I w1 L2,22 - I w2 L2,22 == 0, d3 + w21 L1,23 - 2 I w2 L2,23 == 0<

eq611c = HCoefficient@Subtract žž # •. sol2Form, NRT@@2DDD •. Exp@_ T0 + _.D -> 0 & •ž

order2Eqm@@83, 4<DD •• FlattenL == 0 •• Thread

8-L4,1 == 0, -2 I W L3,2 - L4,2 == 0, 2 I W L3,3 - L4,3 == 0, I w1 L3,4 - L4,4 == 0,

-I W L3,5 + I w1 L3,5 - L4,5 == 0, I W L3,6 + I w1 L3,6 - L4,6 == 0, -I W L3,7 + I w2 L3,7 - L4,7 == 0,
I W L3,8 + I w2 L3,8 - L4,8 == 0, I w1 L3,9 + I w2 L3,9 - L4,9 == 0, 2 I w2 L3,10 - L4,10 == 0,
-I w1 L3,11 - L4,11 == 0, -I W L3,12 - I w1 L3,12 - L4,12 == 0, I W L3,13 - I w1 L3,13 - L4,13 == 0, -L4,14 == 0,
-I w1 L3,15 + I w2 L3,15 - L4,15 == 0, -I W L3,16 - I w2 L3,16 - L4,16 == 0, I W L3,17 - I w2 L3,17 - L4,17 == 0,
I w1 L3,18 - I w2 L3,18 - L4,18 == 0, -L4,19 == 0, -I w1 L3,20 - I w2 L3,20 - L4,20 == 0,
-2 I w2 L3,21 - L4,21 == 0, 2 d2 + w22 L3,1 == 0, d2 + w22 L3,2 - 2 I W L4,2 == 0, d2 + w22 L3,3 + 2 I W L4,3 == 0,
w22 L3,4 + I w1 L4,4 == 0, 2 d2 + w22 L3,5 - I W L4,5 + I w1 L4,5 == 0, 2 d2 + w22 L3,6 + I W L4,6 + I w1 L4,6 == 0,
2 d3 + w22 L3,7 - I W L4,7 + I w2 L4,7 == 0, 2 d3 + w22 L3,8 + I W L4,8 + I w2 L4,8 == 0,
2 d3 + w22 L3,9 + I w1 L4,9 + I w2 L4,9 == 0, 3 d4 + w22 L3,10 + 2 I w2 L4,10 == 0, w22 L3,11 - I w1 L4,11 == 0,
2 d2 + w22 L3,12 - I W L4,12 - I w1 L4,12 == 0, 2 d2 + w22 L3,13 + I W L4,13 - I w1 L4,13 == 0, 2 d2 + w22 L3,14 == 0,
2 d3 + w22 L3,15 - I w1 L4,15 + I w2 L4,15 == 0, 2 d3 + w22 L3,16 - I W L4,16 - I w2 L4,16 == 0,
2 d3 + w22 L3,17 + I W L4,17 - I w2 L4,17 == 0, 2 d3 + w22 L3,18 + I w1 L4,18 - I w2 L4,18 == 0,
6 d4 + w22 L3,19 == 0, 2 d3 + w22 L3,20 - I w1 L4,20 - I w2 L4,20 == 0, 3 d4 + w22 L3,21 - 2 I w2 L4,21 == 0<

Solving eq611b for the NRTsymbolList1, we obtain

Chapter6.nb 161

coef11 = Solve@eq611b, Array@NRTsymbolList1, 2D •• FlattenD@@1DD

:L2,5 ® 0, L2,16 ® 0, L2,21 ® 0, L1,5 ® - €€€€€€€€€€ , L1,16 ® - €€€€€€€€€€ , L1,21 ® - €€€€€€€€€€ , L2,1 ® 0, L2,2 ® 0,
6 d1 6 d1 2 d3
2 2
w1 w1 w21
2 W2 m1 2 W2 m1 6 I W d1 6 I W d1 2 I d1
L2,3 ® - €€€€€€€€€€€€€€€€ , L2,4 ® - €€€€€€€€€€€€€€€€ , L2,6 ® - €€€€€€€€€€€€€€€€€€€€ , L2,7 ® €€€€€€€€€€€€€€€€€€€€ , L2,9 ® €€€€€€€€€€€€€€ ,
W2 - w21 W2 - w21 4 W2 - w21 4 W2 - w21 w1
2 I d3 w2 2 I d1 2 I d3 w2
L2,10 ® 0, L2,14 ® - €€€€€€€€€€€€€€€€€€€€ , L2,17 ® - €€€€€€€€€€€€€€ , L2,18 ® 0, L2,23 ® €€€€€€€€€€€€€€€€€€€€ , L1,1 ® 0, L1,2 ® 0,
w1 - 4 w2
2 2 w1 w21 - 4 w22
2IWm 2IWm 3 d1 3 d1 d1
L1,3 ® - €€€€€€€€€€€€€€€€1€ , L1,4 ® €€€€€€€€€€€€€€€€1€ , L1,6 ® €€€€€€€€€€€€€€€€
€€€€ , L1,7 ® €€€€€€€€€€€€€€€€
€€€€ , L1,9 ® €€€€€€ , L1,10 ® 0,
2 2 2 2
w21
6 I d1 HW + w1L
W - w1
2 W - w1
2 4 W - w1
2 4 W - w1
2

W HW + 2 w1L
d3 d d3
€€€€ , L1,17 ® €€€€1€€ , L1,18 ® 0, L1,23 ® - €€€€€€€€€€€€€€€€
L1,14 ® - €€€€€€€€€€€€€€€€ €€€€ , L2,8 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

2 I d2 HW - w2L 2 I d2 HW + w2 L 2 I d2 Hw1 + w2 L
w21 - 4 w22 w21 w21 - 4 w22

w2 H2 w1 + w2 L
L2,11 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L2,12 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L2,13 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

6 I d1 HW + w1L 2 I d2 HW + w2L 2 I d2 HW - w2 L
2 2 2 2
W - w1 - 2 W w2 + w2
2 W - w1 + 2 W w2 + w2
2

W HW + 2 w1L
L2,15 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L2,19 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L2,20 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ ,

2 I d2 Hw1 + w2 L
W - w1 + 2 W w2 + w2
2 2 2 W - w21 - 2 W w2 + w22
2

w2 H2 w1 + w2 L W HW + 2 w1L
6 d1 2 d2
L2,22 ® - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ , L1,8 ® €€€€€€€€€€€€€€€€
€€€€€€€€€€ , L1,11 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€ ,
-W2 + w1 + 2 W w2 - w22
2

2 d2 2 d2 6 d1
w2 H2 w1 + w2 L W HW + 2 w1 L
L1,12 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ , L1,13 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L1,15 ® €€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2 2
-W2 + w1 - 2 W w2 - w2

€€€€€€€€€€€€€ , L1,22 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

H2 w1 + w2 L
2 d2 2 d2 2 d2
L1,19 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ , L1,20 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
-W + w1 - 2 W w2 - w2
2 2 2 -W + w1 + 2 W w2 - w2
2 2 2 w 2

Solving eq611c for the NRTsymbolList2, we obtain

coef12 = Solve@eq611c, Array@NRTsymbolList2, 2D •• FlattenD@@1DD

:L4,1 ® 0, L4,14 ® 0, L4,19 ® 0, L3,1 ® - €€€€€€€€€€ , L3,14 ® - €€€€€€€€€€ , L3,19 ® - €€€€€€€€€€ , L4,2 ® - €€€€€€€€€€€€€€€€€€€€ ,
2 d2 2 d2 6 d4 2 I W d2
2 2 2
w2 w2 w2 4 W2 - w22
2 I W d2 2 I d4 2 I d4 d2
L4,3 ® €€€€€€€€€€€€€€€€€2€€€ , L4,4 ® 0, L4,10 ® €€€€€€€€€€€€€€ , L4,11 ® 0, L4,21 ® - €€€€€€€€€€€€€ , L3,2 ® €€€€€€€€€€€€€€€€€2€€€ ,
w2 w2
2 I d2 HW - w1L
4 W - w2
2 4 W - w2
2

d2 d4 d4
L3,3 ® €€€€€€€€€€€€€€€€
€€€€ , L3,4 ® 0, L3,10 ® €€€€€€ , L3,11 ® 0, L3,21 ® €€€€€€ , L4,5 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ ,

2 I d2 HW + w1 L 2 I d3 HW - w2 L 2 I d3 HW + w2 L 2 I d3 Hw1 + w2 L
4 W2 - w22 w22 w22 W2 - 2 W w1 + w21 - w22

HW L HW L w1 Hw1 + 2 w2 L
L4,6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L4,7 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L4,8 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L4,9 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
W - w W + w
2 I d2 HW + w1L 2 I d2 HW - w1 L 2 I d3 Hw1 - w2L
2 2
W + 2 W w1 + w1 - w2
2 2 2 2 2

w1 Hw1 - 2 w2L
L4,12 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L4,13 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L4,15 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

2 I d3 HW + w2L 2 I d3 HW - w2L 2 I d3 Hw1 - w2 L 2 I d3 Hw1 + w2L

W + 2 W w1 + w1 - w2
2 2 2 W - 2 W w1 + w1 - w2
2 2 2

W HW + 2 w2L W HW - 2 w2 L w1 Hw1 - 2 w2 L w1 Hw1 + 2 w2L

L4,16 ® - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ , L4,17 ® €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ , L4,18 ® €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ , L4,20 ® - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ ,

W HW - 2 w2 L W HW + 2 w2 L
2 d2 2 d2 2 d3 2 d3
L3,5 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L3,6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L3,7 ® €€€€€€€€€€€€€€€€€€€€€€€€€€ , L3,8 ® €€€€€€€€€€€€€€€€€€€€€€€€€€ ,
W - 2 W w1 + w1 - w2
2 2 2 W + 2 W w1 + w1 - w2
2 2 2

w1 Hw1 + 2 w2L w1 Hw1 - 2 w2 L

2 d3 2 d2 2 d2 2 d3
L3,9 ® €€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€ , L3,12 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L3,13 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ , L3,15 ® €€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€ ,
2 2 2 2
W2 + 2 W w1 + w1 - w2 W2 - 2 W w1 + w1 - w2

L3,16 ® €€€€€€€€€€€€€€€€€€€€€€€€€€ , L3,17 ® €€€€€€€€€€€€€€€€€€€€€€€€€€ , L3,18 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , L3,20 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

W HW + 2 w2 L W HW - 2 w2 L w1 Hw1 - 2 w2L w1 Hw1 + 2 w2 L
2 d3 2 d3 2 d3 2 d3
162 Chapter6.nb

Substituting sol2Form into order2Eqm and collecting the coefficients of like RT, we obtain two sets of algebraic equations as

eq1 = MapIndexed@Coefficient@order2Eqm@@1, 1DD •. sol2Form, #1D == r1,#2@@1DD &, RT@@1DDD •.

Exp@_ T0 + _.D -> 0

82 I w1 G1,1 == r1,1, I W G1,2 == r1,2, I w2 G1,3 == r1,3 <

eq2 = MapIndexed@Coefficient@order2Eqm@@3, 1DD •. sol2Form, #1D == r3,#2@@1DD &, RT@@2DDD •.

Exp@_ T0 + _.D -> 0

8I W G3,1 + I w2 G3,1 == r3,1, I W G3,2 + I w2 G3,2 == r3,2, 2 I w1 G3,3 + I w2 G3,3 == r3,3 , 2 I w2 G3,4 == r3,4<

Solving eq1 and eq2 for the RTsymbolList1 and RTsymbolList2, we obtain

coef21 = Solve@eq1, RTsymbolList1D@@1DD •. r1Rule

:G1,1 ® - €€€€€€€€1€€ , G1,2 ® - €€€€€€€€1€€ , G1,3 ® - €€€€€€€€

€€€€ >
Im 3d d2
2 w1 W w1 w1 w2

coef22 = Solve@eq2, RTsymbolList2D@@1DD •. r3Rule

:G3,4 ® - €€€€€€€€€€ , G3,1 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , G3,2 ® 0, G3,3 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >

4 w2 HW + w2 L 2 w2 H2 w1 + w2 L
I m2 f2 d2
2 w2

Combining these coefficients using the function Join yields

coeffs = Join@coef11, coef12, coef21, coef22,

coef21 •. conjugateRule, coef22 •. conjugateRuleD •. HomgRule@@1DD •. e -> 0L;

Then, we express the solution of the second-order equations in pure function form as

sol2 = sol2Form •. Function@8T0, T1 , T2 <, a_D :>

Function@8T0, T1, T2 < •• Evaluate, a •. coeffs •• Expand •• EvaluateD;
sol2 •. displayRule

9u1,2 ® FunctionA8T0 , T1 , T2<,

6 d1 L2 E-2 I T0 W-2 I t1 d1 L21 E2 I T0 W+2 I t1 d1 L21 4 I E-IT0 W-I t1 L1 m1 4 I EI T0 W+I t1 L1 m1
- €€€€€€€€€€€€€1€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ -
w21 5 w21 5 w21 3 w1 3 w1
3 E-IT0 W-I t1+I T0 w1 d1 L1 A1 3 EI T0 W+I t1+I T0 w1 d1 L1 A1 I EI T0 w1 m1 A1 E2 I T0 w1 d1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ -
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w21 4 w21 2 w1 w21
2 E-IT0 W-I t1+I T0 w2 d2 L1 A2 2 EI T0 W+I t1+I T0 w2 d2 L1 A2 EI T0 w1 +I T0 w2 d2 A1 A2 E2 I T0 w2 d3 A22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
w21 15 w21 4 w21 15 w21
• • • •
3 E-IT0 W-I t1-I T0 w1 d1 L1 A1 3 EI T0 W+I t1-I T0 w1 d1 L1 A1 I E-I T0 w1 m1 A1 6 d1 A1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€ -
4 w21 2 w21 2 w1 w21
• •2 • •
E-I T0 w1 +I T0 w2 d2 A2 A1 E-2 I T0 w1 d1 A1 2 E-I T0 W-I t1 -I T0 w2 d2 L1 A2 2 EI T0 W+I t1-I T0 w2 d2 L1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ -
2 w21 w21 15 w21 w21

E
Chapter6.nb 163

€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ E,
• • • • •2
EI T0 w1 -IT0 w2 d2 A1 A2 2 d3 A2 A2 E-I T0 w1-I T0 w2 d2 A1 A2 E-2 I T0 w2 d3 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w21 w21 4 w21 15 w21

v1,2 ® FunctionA8T0 , T1 , T2<, - €€€€ E-I T0 W-I t1 L1 m1 - €€€€ EI T0 W+I t1 L1 m1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

8 8 4 I E-2 I T0 W-2 I t1 d1 L21
€€€€€€€€€€€€€€€€ +
3 3 5 w1
4 I E2I T0 W+2 It1 d1 L21 1 3 I E-I T0 W-It1 +I T0 w1 d1 L1 A1 9 I EIT0 W+I t1+I T0 w1 d1 L1 A1
€€€€€€€€€€€€€€ - €€€€ EI T0 w1 m1 A1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
5 w1 2 2 w1 4 w1
2 I E2I T0 w1 d1 A21 8 I EIT0 W+I t1+I T0 w2 d2 L1 A2 3 I EI T0 w1 +I T0 w2 d2 A1 A2 4 I E2 I T0 w2 d3 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€2€ -
€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w1 15 w1 4 w1 15 w1
-I W-I t -I w • W+I t -I w •
1 • 9IE T0 1 T 0 1
d1 L1 A1 3IE I T0 1 T 0 1
d1 L1 A1
€€€€ E-I T0 w1 m1 A1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€€€ +
2 4 w1 2 w1
• •2 •
I E-IT0 w1+I T0 w2 d2 A2 A1 2 I E-2 I T0 w1 d1 A 8 I E-I T0 W-I t1 -I T0 w2 d2 L1 A2
€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2 w1 w1 15 w1

€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ E,
• • • •2
I EI T0 w1 -I T0 w2 d2 A1 A2 3 I E-I T0 w1 -I T0 w2 d2 A1 A2 4 I E-2 I T0 w2 d3 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w1 4 w1 15 w1

u2,2 ® FunctionA8T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€2€ + €€€€€€€€€€€€€€€€€€€€€€€€€2€€ - €€€€€€€€€1€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

E-I T0 W-I t2 f EI T0 W+I t2 f d2 L2 E-2 I T0 W-2 I t1 d2 L21
€€€€€€€€ +
32 w21 32 w21 2 w21 12 w21
E2 IT0 W+2 I t1 d2 L21 2 E-I T0 W-It1 +I T0 w1 d2 L1 A1 2 EI T0 W+I t1+I T0 w1 d2 L1 A1 E2 I T0 w1 d2 A21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
12 w21 3 w21 5 w21 16 w21
E-I T0 W-I t1+I T0 w2 d3 L1 A2 EI T0 W+I t1 +IT0 w2 d3 L1 A2 I EI T0 w2 m2 A2 2 EI T0 w1+I T0 w2 d3 A1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ +
2 w1 2
6 w1 2 4 w1 5 w21
• • •
E2 IT0 w2 d4 A22 2 E-I T0 W-I t1-I T0 w1 d2 L1 A1 2 EI T0 W+I t1 -IT0 w1 d2 L1 A1 d2 A1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€ -
4 w21 5 w21 3 w21 2 w21
• • 2 • •
2 E-IT0 w1+I T0 w2 d3 A2 A1 E-2 I T0 w1 d2 A1 E-I T0 W-I t1 -I T0 w2 d3 L1 A2 EIT0 W+I t1-I T0 w2 d3 L1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ +
3 w21 16 w21 6 w21 2 w21

€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€2€€ E,
• • • • • •2
I E-IT0 w2 m A2 2 EI T0 w1 -IT0 w2 d3 A1 A2 3 d4 A2 A2 2 E-I T0 w1-I T0 w2 d3 A1 A2 E-2 I T0 w2 d4 A
€€€€€€€€€€€€€€€€€€€€€€€€2€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 w1 3 w21 2 w21 5 w21 4 w21

v2,2 ® FunctionA8T0 , T1 , T2<, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

I E-2I T0 W-2 It1 d2 L21 I E2 I T0 W+2 I t1 d2 L21 I E-I T0 W-I t2 f2 w2
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€ -
3 w1 3 w1 32 w21
I EI T0 W+I t2 f2 w2 2 I E-I T0 W-I t1 +I T0 w1 d2 L1 A1 6 I EI T0 W+I t1+I T0 w1 d2 L1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
32 w21 3 w1 5 w1
I E2 IT0 w1 d2 w2 A2 2 I EI T0 W+I t1 +I T0 w2 d3 L1 A2 EIT0 w2 m2 w2 A2 6 I EI T0 w1 +I T0 w2 d3 A1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€1€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ +
16 w21 3 w1 4 w1 5 w1
• •
I E2 IT0 w2 d4 A22 6 I E-IT0 W-I t1-I T0 w1 d2 L1 A1 2 I EI T0 W+I t1-I T0 w1 d2 L1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w1 5 w1 3 w1
• •2 •
2 I E-I T0 w1+I T0 w2 d3 A2 A1 I E-2 I T0 w1 d2 w2 A1 2 I E-I T0 W-I t1 -I T0 w2 d3 L1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
3 w1 16 w21 3 w1

€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€2€€ E=
• • • • •2
E-I T0 w2 m2 w2 A2 2 I EI T0 w1-I T0 w2 d3 A1 A2 6 I E-I T0 w1 -I T0 w2 d3 A1 A2 I E-2 I T0 w2 d4 A
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 w1 3 w1 5 w1 w1

Ÿ Third-Order Problem
164 Chapter6.nb

Substituting the first- and second-order solutions into the third-order problem, eqOrder[3], yields

order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

Substituting the expRule1 into the right-hand sides of order3Eq and collecting the terms that could produce secular terms,
the terms proportional to EI w1 T 0 and EI w2 T 0 , we have

ST21 = Coefficient@order3Eq@@#, 2DD •. expRule1@1D, Exp@I w1 T0DD & •ž 81, 2<;

ST21 •. displayRule

EI T1 s1 HD1A1 L A2 d2 3 EI T1 s1+I T1 s2 +I t1 HD1A1 L d1 L1 I HD1A1 L m1 EI T1 s1 HD1 A2 L d2 A1

:-HD2A1 L + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• • •
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ ,
2 w21 2 w21 2 w1 2 w21

-6 E-I T1 s2 -It1 A1 A2 a2 L1 - 24 A1 a1 L21 + €€€€ HD1 A1L m1 + A1 m21 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

1 27 E-I T1 s2-I t1 A1 A2 d1 d2 L1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 2 w21
I EI T1 s1 HD1A1 L A2 d2
•
23 E-IT1 s2-I t1 A1 A2 d2 d3 L1 81 A1 d21 L21 23 A1 d22 L21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ -
15 w21 2 w21 15 w21 2 w1

3 I EI T1 s1 +I T1 s2+I t1 HD1 A1L d1 L1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - I HD2 A1L w1 - 12 A21 a1 A1 + €€€€€€€€€€€€€€€€
• • •
• 30 A21 d21 A1 EI T1 s1+I T1 s2 +I t2 f2 d2 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ +
2 w1 w21 16 w21
I EI T1 s1 HD1 A2 L d2 A1
• • • •
9 A21 d22 A1 2 I EI T1 s1 A2 d2 m1 A1 14 I EI T1 s1+I T1 s2 +I t1 d1 L1 m1 A1
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
8 w21 2 w1 w1 w1
• • •
I EI T1 s1 A2 d2 m2 A1 • • A1 A2 d22 A2 12 A1 A2 d1 d3 A2
€€€€€€€€€€ - 4 A1 A2 a3 A2 - 6 EI T1 s2 +I t1 A1 a2 L1 A2 + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 w1 2 w21 w21

€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
• • • •
8 A1 A2 d23 A2 3 A1 A2 d2 d4 A2 27 EI T1 s2 +I t1 A1 d1 d2 L1 A2 23 EI T1 s2 +It1 A1 d2 d3 L1 A2
€€€€€€€€€€€€€€€€€€€€€€€€
€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
15 w21 w21 2 w21 15 w21
Chapter6.nb 165

ST22 = Coefficient@order3Eq@@#, 2DD •. expRule1@2D, Exp@I w2 T0DD & •ž 83, 4<;

ST22 •. displayRule

E-I T1 s1 HD1 A1 L A1 d2 I HD1 A2L m2

:-HD2A2 L + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€ ,
8 w1 2 4 w1
58 E-I T1 s2-I t1 A22 d2 d3 L1
-3 E-I T1 s2 -It1 A22 a4 L1 - 4 A2 a3 L21 - 3 EI T1 s2 +I t1 a2 L31 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ +
15 w21
5 E-I T1 s2 -I t1 A22 d3 d4 L1 56 A2 d22 L21 12 A2 d1 d3 L21 2 A2 d23 L21 3 A2 d2 d4 L21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€ +
2 w21 15 w21 w21 3 w21 w21

€€€€€€ - I HD2A2 L w2 -
58 EI T1 s2 +I t1 d1 d2 L31 5 EI T1 s2 +It1 d2 d3 L31 I E-I T1 s1 A21 d2 m1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
5 w21 6 w21 w1

I E-I T1 s1 HD1 A1L A1 d2 w2 I EI T1 s2+I t2 f2 m2 w2 I E-I T1 s1 A21 d2 m2 w2 HD1A2 L m2 w2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€ +
8 w1 2
16 w1 2
8 w1 2 4 w1
• •
A2 m22 w2 • I T1 s2 +I t1 • A1 A2 d22 A1 12 A1 A2 d1 d3 A1
€€€€€€€€€€€€€€€€€€ - 4 A1 A2 a3 A1 - 6 E A1 a2 L1 A1 + €€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€ +
2 w1 2 w21 w21
• • • •
8 A1 A2 d23 A1 3 A1 A2 d2 d4 A1 27 EI T1 s2 +I t1 A1 d1 d2 L1 A1 23 EI T1 s2 +It1 A1 d2 d3 L1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
15 w21 w21 2 w21 15 w21
• •
2 • I T1 s2 +I t1 • 2 I T1 s2+2 I t1 2• 58 A22 d23 A2 15 A22 d24 A2
12 A2 a5 A2 - 6 E A2 a4 L1 A2 - 2 E a3 L1 A2 + €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ +
15 w21 2 w21
• • •
116 EI T1 s2 +I t1 A2 d2 d3 L1 A2 5 EI T1 s2 +It1 A2 d3 d4 L1 A2 4 E2 I T1 s2 +2 I t1 d22 L21 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
15 w21 w21 w21

€€€€€€€€€€€€€€€€€€€€ >
• • •
2 E2 IT1 s2+2 I t1 d1 d3 L21 A2 E2 I T1 s2 +2 I t1 d23 L21 A2 E2 I T1 s2 +2 It1 d2 d4 L21 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
5 w21 w21 2 w21

Then, the solvability conditions demand that ST21 and ST22 be orthogonal to solutions of their corresponding adjoints, the
components of ccleftVec. The result is
166 Chapter6.nb

SCond2 = 8ccleftVec@@1DD.ST21 == 0, ccleftVec@@2DD.ST22 == 0< •• ExpandAll;

SCond2 •. displayRule

:-6 E-I T1 s2 -It1 A1 A2 a2 L1 - 24 A1 a1 L21 + A1 m21 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

27 E-I T1 s2 -I t1 A1 A2 d1 d2 L1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 w21

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ - 2 I HD2 A1L w1 - 12 A1 a1 A1 + €€€€€€€€€€€€€€€€€€€€€€€€ -

•
23 E-IT1 s2-I t1 A1 A2 d2 d3 L1 81 A1 d21 L21 23 A1 d22 L21 2 • 30 A21 d21 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
15 w21 2 w21 15 w21 w21
• • • •
EI T1 s1 +IT1 s2+I t2 f2 d2 A1 9 A21 d22 A1 2 I EI T1 s1 A2 d2 m1 A1 14 I EI T1 s1 +I T1 s2 +It1 d1 L1 m1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
16 w21 8 w21 w1 w1
I T1 s1 • 2• •
IE A2 d2 m2 A1 • • A1 A2 d2 A2 12 A1 A2 d1 d3 A2
€€€€€€€€€€ - 4 A1 A2 a3 A2 - 6 EI T1 s2 +I t1 A1 a2 L1 A2 + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 w1 2 w21 w21
• • • •
8 A1 A2 d23 A2 3 A1 A2 d2 d4 A2 27 EI T1 s2 +I t1 A1 d1 d2 L1 A2 23 EI T1 s2 +It1 A1 d2 d3 L1 A2
€€€€€€€€€€€€€€€€€€€€€€€€
€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ == 0,
15 w21 w21 2 w21 15 w21
58 E-I T1 s2-I t1 A22 d2 d3 L1
-3 E-I T1 s2 -It1 A22 a4 L1 - 4 A2 a3 L21 - 3 EI T1 s2 +I t1 a2 L31 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ +
15 w21
5 E-IT1 s2-I t1 A22 d3 d4 L1 56 A2 d22 L21 12 A2 d1 d3 L2 2 A2 d23 L21 3 A2 d2 d4 L2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€1€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€1€€€ +
2 w21 15 w21 w21 3 w21 w21

€€€€€€ - 2 I HD2A2 L w2 +
58 EI T1 s2 +I t1 d1 d2 L31 5 EI T1 s2 +It1 d2 d3 L31 I E-I T1 s1 A21 d2 m1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
5 w21 6 w21 w1
I EI T1 s2 +I t2 f2 m2 w2 I E-I T1 s1 A21 d2 m2 w2 A2 m2 w2 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€ + €€€€€€€€€2€€€€€€€€€ - 4 A1 A2 a3 A1 -
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
16 w1 2
8 w1 2 2 w1
2• • •
• A1 A2 d2 A1 12 A1 A2 d1 d3 A1 8 A1 A2 d23 A1
6 EI T1 s2 +I t1 A1 a2 L1 A1 + €€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
2 w21 w21 15 w21
• • •
3 A1 A2 d2 d4 A1 27 EI T1 s2 +I t1 A1 d1 d2 L1 A1 23 EI T1 s2 +I t1 A1 d2 d3 L1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w21 2 w21 15 w21
• •
• • • 58 A22 d23 A2 15 A22 d24 A2
12 A22 a5 A2 - 6 EI T1 s2 +I t1 A2 a4 L1 A2 - 2 E2 I T1 s2+2 I t1 a3 L21 A2 + €€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€ +
15 w21 2 w21
• • •
116 EIT1 s2+I t1 A2 d2 d3 L1 A2 5 EI T1 s2 +I t1 A2 d3 d4 L1 A2 4 E2 I T1 s2 +2 I t1 d22 L21 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ -
15 w21 w21 w21
• • •
2 E2 IT1 s2+2 I t1 d1 d3 L21 A2 E2 I T1 s2 +2 I t1 d23 L21 A2 E2 I T1 s2 +2 It1 d2 d4 L21 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ == 0>
5 w21 w21 2 w21

Solving SCond2 for D2 A1 and D2 A2 , we have

Chapter6.nb 167

H0,1L H0,1L
SCond2Rule1 = SolveASCond2, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD •• ExpandAll;
SCond2Rule1 •. displayRule

:D2 A1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
27 I E-I T1 s2-I t1 A1 A2 d1 d2 L1 23 I E-I T1 s2-I t1 A1 A2 d2 d3 L1 81 I A1 d21 L21 23 I A1 d22 L21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€
€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ €€€€ +
4 w1 3 30 w1 3 4 w1
3 30 w31
• •
3 I E-IT1 s2-I t1 A1 A2 a2 L1 12 I A1 a1 L2 I A1 m2 15 I A21 d21 A1 I EI T1 s1 +I T1 s2 +I t2 f2 d2 A1
€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€1€€€ - €€€€€€€€€€€€€1€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w1 w1 2 w1 w31 32 w31
• • • • •
9 I A21 d22 A1 EI T1 s1 A2 d2 m1 A1 7 EI T1 s1 +I T1 s2 +I t1 d1 L1 m1 A1 EIT1 s1 A2 d2 m2 A1 6 I A21 a1 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ -
16 w31 w21 w21 4 w21 w1
• • • • •
I A1 A2 d22 A2 6 I A1 A2 d1 d3 A2 4 I A1 A2 d2 A2 3 I A1 A2 d2 d4 A2 27 I EI T1 s2+I t1 A1 d1 d2 L1 A2
€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€3€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
4 w31 w31 15 w31 2 w31 4 w31
• • •
23 I EIT1 s2+I t1 A1 d2 d3 L1 A2 2 I A1 A2 a3 A2 3 I EI T1 s2+I t1 A1 a2 L1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€ ,
30 w31 w1 w1

EI T1 s2 +I t2 f2 m2 E-I T1 s1 A21 d2 m2 I A m2 3 I E-I T1 s2 -It1 A22 a4 L1 2 I A2 a3 L21

D2 A2 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ - €€€€€€€€2€€€€€2€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€ - €€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ +
32 w21 16 w21 4 w1 2 w2 w2
3 I EIT1 s2+I t1 a2 L31 29 I E-I T1 s2 -I t1 A22 d2 d3 L1 5 I E-I T1 s2-I t1 A22 d3 d4 L1 28 I A2 d22 L21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€
€€€€ -
2 w2 15 w21 w2 4 w21 w2 15 w21 w2
6 I A2 d1 d3 L21 I A2 d23 L21 3 I A2 d2 d4 L21 29 I EI T1 s2 +I t1 d1 d2 L31 5 I EI T1 s2 +I t1 d2 d3 L31
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ +
w21 w2 3 w21 w2 2 w21 w2 5 w21 w2 12 w21 w2
• • • •
E-I T1 s1 A21 d2 m1 2 I A1 A2 a3 A1 3 I EI T1 s2 +I t1 A1 a2 L1 A1 I A1 A2 d22 A1 6 I A1 A2 d1 d3 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€
€ €€
€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ -
2 w1 w2 w2 w2 4 w21 w2 w21 w2
• • • •
4 I A1 A2 d23 A1 3 I A1 A2 d2 d4 A1 27 I EI T1 s2+I t1 A1 d1 d2 L1 A1 23 I EI T1 s2+I t1 A1 d2 d3 L1 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
15 w21 w2 2 w21 w2 4 w21 w2 30 w21 w2
• • • • •
6 I A22 a5 A2 3 I EI T1 s2 +It1 A2 a4 L1 A2 I E2 I T1 s2+2 I t1 a3 L21 A2 29 I A22 d23 A2 15 I A22 d24 A2
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w2 w2 w2 15 w21 w2 4 w21 w2
• • •
58 I EIT1 s2+I t1 A2 d2 d3 L1 A2 5 I EIT1 s2+I t1 A2 d3 d4 L1 A2 2 I E2I T1 s2+2 I t1 d22 L21 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€ +
15 w21 w2 2 w21 w2 w21 w2

€€€€€€€€€€€€€€€€€€€€€€€€ >
• • •
I E2 IT1 s2+2 I t1 d1 d3 L21 A2 I E2 I T1 s2 +2 It1 d23 L21 A2 I E2 I T1 s2 +2I t1 d2 d4 L21 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
5 w21 w2 2 w21 w2 4 w21 w2

Ÿ Reconstitution

Using the method of reconstitution, Ak ¢ = e D1 Ak + e2 D2 Ak + º, we combine the partial-differential equations SCond1-

Rule1 and SCond2Rule1 into the following two ordinary-differential equations governing the modulation of the complex-
valued functions Ak :
168 Chapter6.nb

moduEq = Table@2 I w k A k ¢, 8k, 2<D ==

HTable@2 I w k dt@1D@ A k @T1, T2 DD, 8k, 2<D •. SCond1Rule1 •. SCond2Rule1 ••
Collect@#, eD &L •• Thread;
moduEq •. displayRule

i 3 I EI T1 s1 +I T1 s2 +I t1 d1 L1 A1 y
:2 I w1 A¢1 == 2 I e w1 j j €€€€€ z
z
• •
j 1 1 z+
I EI T1 s1 A2 d2 A1

k {
-A m + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€
w w
i
1 1

2 I e2 w1 j
j
j- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
27 I E -I T1 s 2 -I t1 A1 A2 d1 d2 L1 23 I E -I T1 s 2 -I t1 A1 A2 d2 d3 L1

k
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
3
4 w1 30 w31
81 I A1 d21 L21 23 I A1 d22 L21 3 I E-I T1 s2 -It1 A1 A2 a2 L1 12 I A1 a1 L2 I A1 m2
€€€€€€€€€€€€€€€€€€€€€€€€
€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€1€€€ - €€€€€€€€€€€€€1€€€ -
4 w31 30 w31 w1 w1 2 w1
• • • •
15 I A21 d21 A1 I EI T1 s1 +IT1 s2+I t2 f2 d2 A1 9 I A21 d22 A1 EI T1 s1 A2 d2 m1 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€ -
w31 32 w31 16 w31 w21
• • • •
7 EIT1 s1+I T1 s2+I t1 d1 L1 m1 A1 EI T1 s1 A2 d2 m2 A1 6 I A21 a1 A1 I A1 A2 d22 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w21 4 w21 w1 4 w31
• • • •
6 I A1 A2 d1 d3 A2 4 I A1 A2 d23 A2 3 I A1 A2 d2 d4 A2 27 I EIT1 s2+I t1 A1 d1 d2 L1 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
w31 15 w31 2 w31 4 w31
3 I EI T1 s2 +I t1 A1 a2 L1 A2 y
€€€€€€€€€€€€€€€€€€€€€€ z
z,
• • •
z
23 I EI T1 s2+I t1 A1 d2 d3 L1 A2 2 I A1 A2 a3 A2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
30 w31 w1 w1

i I E-I T1 s1 A21 d2 y
2 I w2 A¢2 == 2 I e j j
j €€€€€€€€€€€€€€€€ z
z
z w2 +
I EI T1 s2 +It2 f2

k {
-A 2 m 2 - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€€€€
€ + €€€€€€€€€€€€€€€€
4 w 2 w
i
2 2

2 I e2 w2 j
j
j €€€€€€€€€€€€€€€€€€€€€€€€
IT1 s2+I t2 -I T1 s1 2
E f m E A d m I A m 2
3 I E-I T1 s2-I t1 A22 a4 L1 2 I A2 a3 L2

k
2 2 1 2 2 2 2
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€1€ +
+ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
32 w1 2
16 w1 2 4 w1 2 w2 w2

3 I EIT1 s2+I t1 a2 L31 29 I E-I T1 s2 -I t1 A22 d2 d3 L1 5 I E-I T1 s2-I t1 A22 d3 d4 L1 28 I A2 d22 L21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€
€€€€ -
2 w2 15 w21 w2 4 w21 w2 15 w21 w2
6 I A2 d1 d3 L21 I A2 d23 L21 3 I A2 d2 d4 L21 29 I EI T1 s2 +I t1 d1 d2 L31 5 I EI T1 s2 +I t1 d2 d3 L31
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ +
w21 w2 3 w21 w2 2 w21 w2 5 w21 w2 12 w21 w2
• • • •
E-IT1 s1 A21 d2 m1 2 I A1 A2 a3 A1 3 I EI T1 s2 +I t1 A1 a2 L1 A1 I A1 A2 d22 A1 6 I A1 A2 d1 d3 A1
€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€
€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
2 w1 w2 w2 w2 4 w21 w2 w21 w2
• • • •
4 I A1 A2 d2 A1 3 I A1 A2 d2 d4 A1 27 I EI T1 s2+I t1 A1 d1 d2 L1 A1 23 I EI T1 s2+I t1 A1 d2 d3 L1 A1
€€€€€€€€€€€€€€€€€€€€€€€€3€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
15 w21 w2 2 w21 w2 4 w21 w2 30 w21 w2
• • • • •
6 I A22 a5 A2 3 I EI T1 s2 +I t1 A2 a4 L1 A2 I E2 I T1 s2 +2 I t1 a3 L21 A2 29 I A22 d23 A2 15 I A22 d24 A2
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€ -
w2 w2 w2 15 w21 w2 4 w21 w2
• • •
58 I EI T1 s2+I t1 A2 d2 d3 L1 A2 5 I EI T1 s2+I t1 A2 d3 d4 L1 A2 2 I E2 I T1 s2+2 I t1 d22 L21 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€ +
15 w21 w2 2 w21 w2 w21 w2
I E2 I T1 s2 +2 I t1 d2 d4 L21 A2 y
€€€€€€€€€€€€€€€€€€€€€€€€ z
z
z>
• • •
I E2I T1 s2+2 I t1 d1 d3 L21 A2 I E2 I T1 s2 +2 It1 d23 L21 A2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
5 w21 w2 2 w21 w2 4 w21 w2
Chapter6.nb 169

Ÿ 6.1.2 The function MMS

According to the procedures described in the previous section, we can build a function named MMS (Method of Multiple
Scales) specifically for eq61b. A more general function (a Package) can be created by considering as arguments the govern -
ing equations, symbols for dependent variables, independent variable, excitation amplitudes and frequencies, and all other
related quantities which allow the program to identify their respective meanings. We will use MMS to solve for different
resonance cases in the following sections.

MMS@scaling_List, ResonanceCond : 8__Equal<D :=

ModuleA8<,
omgList = 8w1, w2<;
omgRule = Solve@ResonanceCond, 8W, #< •• FlattenD@@1DD & •ž omgList •• Reverse;
multiScales = 8ui_@tD -> ui žž timeScales, vi_@tD -> vi žž timeScales,
Derivative@1D@u_D@tD -> dt@1D@u žž timeScalesD, t -> T0<;
eqa = Heq61b •. scaling •. multiScales •. solRule •• TrigToExp •• ExpandAllL •.
n_•;n>3
e -> 0;
eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eqa •• Transpose;
eqOrder@i_D :=
I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •. u_ k_,1 -> u k,i M == I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •.

u_ k_,1 -> u k,i M - H#@@1DD & •ž eqEps@@iDDL •• Thread;

H* First-Order Problem *L
linearSys = #@@1DD & •ž eqOrder@1D;
expr1 = linearSys •. 9ui_,1 -> IP EI wi #1 &M, vi_,1 -> IQ EI wi #1
&M= •. Exp@a_D -> 1 ••
Partition@#, 2D &;
coefMat = Outer@D, #, 8P, Q<D & •ž expr1;
hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose;
rightVec = # • #@@1DD & •ž HNullSpace@#D@@1DD & •ž coefMatL;
leftVec = NullSpace@hermitian@#DD@@1DD & •ž coefMat;
ccleftVec = leftVec •. conjugateRule;
order1Eq = eqOrder@1D •. u_i_,1 -> Hui,1@#1D &L;

sol1p =
DSolve@order1Eq, 8u1,1@T0 D, v1,1@T0 D, u2,1@T0 D, v2,1@T0 D<, T0 D@@1DD •. C@_D -> 0 •• Simplify;
fRule = 9fi_ -> 2 Li Iw2i - W2M=;
sol1u =
TableAui,1 -> FunctionA8T0 , T1, T2 <, Ai @T1, T2 D Exp@I wi T0 D + Ai @T1, T2 D Exp@-I wi T0 D +
•

Hui,1@T0 D •. sol1p •. fRule •• Simplify •• ExpandL •• EvaluateE, 8i, 2<E;

sol1v = Table@vi,1 -> Function@8T0, T1, T2 <,
D@ui,1 žž timeScales, T0D •. sol1u •• EvaluateD, 8i, 2<D;
sol1 = Join@sol1u, sol1vD;
170 Chapter6.nb

H* Second-Order Problem *L
order2Eq = eqOrder@2D •. sol1 •• ExpandAll;
expRule1@i_D :=
Exp@a_D :> Exp@Expand@a •. omgRule@@iDDD •. e n_. T0 :> timeScales@@n + 1DDD;
ST11 = Coefficient@order2Eq@@#, 2DD •. expRule1@1D, Exp@I w1 T0DD & •ž 81, 2<;
ST12 = Coefficient@order2Eq@@#, 2DD •. expRule1@2D, Exp@I w2 T0DD & •ž 83, 4<;
SCond1 = 8ccleftVec@@1DD.ST11 == 0, ccleftVec@@2DD.ST12 == 0<;
SCond1Rule1 = SolveASCond1, 9AH1,0L @T1, T2 D, AH1,0L@T1 , T2 D=E@@1DD •• ExpandAll;
sigRule = Solve@ResonanceCond, 8s1 , s2 <D@@1DD;
1 2

expRule2 = Exp@a_D :> ExpAa •. 9T1 -> e T0, T2 -> e2 T0= •. sigRule •• ExpandE;
SCond1Rule2 = SCond1Rule1 •. expRule2;
ccSCond1Rule2 = SCond1Rule2 •. conjugateRule;
order2Eqm = order2Eq •. SCond1Rule2 •. ccSCond1Rule2 •• ExpandAll;
IfAUnion@#@@2DD & •ž order2EqmD === 80<,
sol2 = 8u1,2 -> H0 &L, v1,2 -> H0 &L, u2,2 -> H0 &L, v2,2 -> H0 &L<,
basicH = TableA9Ai@T1 , T2D EI wi T0 , Ai@T1 , T2 D E-I wi T0 =, 8i, 2<E •• Flatten;
•

collectForm =
Join@basicH, If@list1 = List žž Plus žž H#@@2DD & •ž eqOrder@1DL •. c_ fi_ Ea_ -> Li Ea ;
Head@list1D === List, list1, 8<DD;
possibleTerms = JoinAcollectForm,
IfAlist1 = List žž Plus žž I#@@2DD & •ž eqOrder@2D •. u_i_,j_ -> H0 &LM •. c_ fi_ Ea_ -> Ea ;

Head@list1D === List, list1, 8<E,

Outer@Times, collectForm, collectFormD •• Flatten •• UnionE;
ResonantTerms@i_D :=
I# •. 8a_ •; a =!= 0 -> 1< & •ž IE-Iwi T0 possibleTerms •. expRule1@iD •. Exp@_ T0 + _.D -> 0MM
possibleTerms •• Union •• Rest;
RT = Array@ResonantTerms, 2D;
r1Rule = MapIndexed@
r1,#2@@1DD -> Coefficient@order2Eqm@@1, 2DD, #1D &, RT@@1DDD •. Exp@_ T0 + _.D -> 0;
H* r2Rule=MapIndexed@r2,#2@@1DD->Coefficient@order2Eqm@@2,2DD,#1D &,
RT@@1DDD•.Exp@_ T0+_.D->0; *L
r3Rule = MapIndexed@r3,#2@@1DD -> Coefficient@order2Eqm@@3, 2DD, #1D &, RT@@2DDD •.
Exp@_ T0 + _.D -> 0;
H* r4Rule=MapIndexed@r4,#2@@1DD->Coefficient@order2Eqm@@4,2DD,#1D &,
RT@@2DDD•.Exp@_ T0+_.D->0; *L
NRT = Complement@possibleTerms, Join@#, # •. conjugateRuleDD & •ž RT;
RTsymbolList1 = Table@G1,j , 8j, Length@RT@@1DDD<D;
RTsymbolList2 = Table@G3,j , 8j, Length@RT@@2DDD<D;
NRTsymbolList1@i_D = Table@Li,j , 8j, Length@NRT@@1DDD<D;
NRTsymbolList2@i_D = Table@Li+2,j, 8j, Length@NRT@@2DDD<D;
8ratio1, ratio2< = Table@-ccleftVec@@i, 1DD, 8i, 2<D;
sol2Form = 8u1,2 -> Function@8T0, T1 , T2 <,
Chapter6.nb 171

RTsymbolList1.RT@@1DD + HRTsymbolList1.RT@@1DD •. conjugateRuleL +

[email protected]@@1DD •• EvaluateD, v1,2 -> Function@8T0 , T1 , T2<,
ratio1 * RTsymbolList1.RT@@1DD + Hratio1 * RTsymbolList1.RT@@1DD •. conjugateRuleL +
[email protected]@@1DD •• EvaluateD, u2,2 -> Function@8T0 , T1 , T2<,
RTsymbolList2.RT@@2DD + HRTsymbolList2.RT@@2DD •. conjugateRuleL +
[email protected]@@2DD •• EvaluateD, v2,2 -> Function@8T0 , T1 , T2<,
ratio2 * RTsymbolList2.RT@@2DD + Hratio2 * RTsymbolList2.RT@@2DD •. conjugateRuleL +
[email protected]@@2DD •• EvaluateD<;
eqb = HCoefficient@Subtract žž # •. sol2Form, NRT@@1DDD •. Exp@_ T0 + _.D -> 0 & •ž
order2Eqm@@81, 2<DD •• FlattenL == 0 •• Thread;
eqc = HCoefficient@Subtract žž # •. sol2Form, NRT@@2DDD •. Exp@_ T0 + _.D -> 0 & •ž
order2Eqm@@83, 4<DD •• FlattenL == 0 •• Thread;
coef11 = Solve@eqb, Array@NRTsymbolList1, 2D •• FlattenD@@1DD;
coef12 = Solve@eqc, Array@NRTsymbolList2, 2D •• FlattenD@@1DD;
eqd = MapIndexed@Coefficient@order2Eqm@@1, 1DD •. sol2Form, #1D == r1,#2@@1DD &,
RT@@1DDD •. Exp@_ T0 + _.D -> 0;
eqe = MapIndexed@Coefficient@order2Eqm@@3, 1DD •. sol2Form, #1D == r3,#2@@1DD &,
RT@@2DDD •. Exp@_ T0 + _.D -> 0;
coef21 = Solve@eqd, RTsymbolList1D@@1DD •. r1Rule;
coef22 = Solve@eqe, RTsymbolList2D@@1DD •. r3Rule;
coeffs = Join@coef11, coef12, coef21, coef22,
coef21 •. conjugateRule, coef22 •. conjugateRuleD •. HomgRule@@1DD •. e -> 0L;
sol2 = sol2Form •. Function@8T0 , T1 , T2<, a_D :>
Function@8T0, T1 , T2 < •• Evaluate, a •. coeffs •• Expand •• EvaluateD
E;

H* Third-Order Problem *L
order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;
ST21 = Coefficient@order3Eq@@#, 2DD •. expRule1@1D, Exp@I w1 T0DD & •ž 81, 2<;
ST22 = Coefficient@order3Eq@@#, 2DD •. expRule1@2D, Exp@I w2 T0DD & •ž 83, 4<;
SCond2 = 8ccleftVec@@1DD.ST21 == 0, ccleftVec@@2DD.ST22 == 0< •• ExpandAll;
H0,1L H0,1L
SCond2Rule1 = SolveASCond2, 9A1 @T1, T2 D, A2 @T1 , T2 D=E@@1DD •• ExpandAll;

H* Reconstitution *L
moduEq =
Table@2 I w k A k ¢, 8k, 2<D == HTable@2 I w k dt@1D@ A k@T1 , T2DD, 8k, 2<D •. SCond1Rule1 •.
SCond2Rule1 •• Collect@#, eD &L •• Thread;
Print@"The second-order approximate solution:"D;
PrintATableA
ui @tD == Iui žž timeScales •. solRule •. e3 -> 0 •. sol1 •. sol2 •. displayRuleM, 8i, 2<EE;
IfAOr žž Table@HFi • fi •. scalingL === e, 8i, 2<D,
Print@"where"D;
172 Chapter6.nb

fi
2 Iwi - W2M
PrintA"Li== €€€€€€€€€€€€€€€€€€€€€€€€€€€ "E
2

E;
Print@"\nThe modulation equations:"D;
Print@moduEq •. displayRuleD
E

As an example, we check the case in Section 6.1.1:

scaling1 = 9m̀ n_ -> e m n , F1 -> e f1, F2 -> e2 f2 =;

ResonanceCond1 = 8w2 == 2 w1 + e s1 , W == w2 + e s2 <;

MMS@scaling1, ResonanceCond1D •• Timing

The second-order approximate solution:

:u 1 @tD == e IE I T0 w1 A 1 + E -I T0 W-I t1 L 1 + E I T0 W+I t1 L 1 + E -I T0 w1 A 1 M +

•

i
jE
e2 j
j €€€€€€€€€€€€€€€€
2 I T0 w1 A 2 d E I T0 w1+I T0 w2 A 1 A 2 d 2 E 2 I T0 w2 A 22 d 3 3 E -I T0 W-I t1 +I T0 w1 A 1 d 1 L 1

k
1 1
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€ +
w 21 4 w 21 15 w 21 2 w 21
3 E I T0 W+I t1+I T0 w1 A 1 d 1 L 1 2 E -I T0 W-I t1+I T0 w2 A 2 d 2 L 1 2 E I T0 W+I t1 +I T0 w2 A 2 d 2 L 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ -
4 w1 2 w1 2 15 w 21
6 d 1 L 21 E -2 I T0 W-2 I t1 d 1 L 21 E 2 I T0 W+2 I t1 d 1 L 21 I E I T0 w1 A 1 m 1 4 I E -I T0 W-I t1 L 1 m 1
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ € - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ +
w1 2 5 w1 2 5 w1 2 2 w1 3 w1
• • •
4 I E I T0 W+I t1 L 1 m 1 6 A1 d1 A1 E -I T0 w1 +I T0 w2 A 2 d 2 A 1 3 E -I T0 W-I t1 -I T0 w1 d 1 L 1 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€€€ -
3 w1 w 21 2 w 21 4 w 21
• • •2 • •
3 E I T0 W+I t1-I T0 w1 d 1 L 1 A 1 I E -I T0 w1 m 1 A 1 E -2 I T0 w1 d 1 A 1 E I T0 w1 -I T0 w2 A 1 d 2 A2 2 A2 d3 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€ €€€€€€ +
2 w 21 2 w1 w 21 2 w 21 w 21

E -2 I T0 w2 d 3 A 2 y z
€€€€€€€€€ €€€€€€ z
• • • • •2
z,
2 E -I T0 W-I t1-I T0 w2 d 2 L 1 A 2 2 E I T0 W+I t1-I T0 w2 d 2 L 1 A 2 E -I T0 w1 -I T0 w2 d 2 A 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ + €€€€€€€€€€€€€€€€
15 w 21 w 21 4 w 21 15 w 21

i
j E 0
u 2 @tD == e IE I T0 w2 A 2 + E -I T0 w2 A 2M + e 2 j
j
j €€€€€€€€€€€€€€€€
-I T W-I t2 f
• E I T0 W+I t2 f 2 E 2 I T0 w1 A 21 d 2

k
2
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ +
32 w 1 2 32 w 1 2 16 w 21

2 E I T0 w1+I T0 w2 A 1 A 2 d 3 E 2 I T0 w2 A 22 d 4 2 E -I T0 W-I t1 +I T0 w1 A 1 d 2 L 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€ +
5 w1 2 4 w1 2 3 w 21
2 E I T0 W+I t1+I T0 w1 A 1 d 2 L 1 E -I T0 W-I t1+I T0 w2 A 2 d 3 L 1 E I T0 W+I t1 +I T0 w2 A 2 d 3 L 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ -
5 w1 2 2 w1 2 6 w 21
•
d 2 L 21 E-2 I T0 W-2 I t1 d 2 L 21 E 2 I T0 W+2 I t1 d 2 L 21 I E I T0 w2 A 2 m 2 A1 d2 A1
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€ €€€€ €€€€€ -
2 w 21 12 w 21 12 w 21 4 w1 2 w 21
-I T w +I T w
0 2 A d A
• -I T W-I t -I T w
0 1 d L A
• I T W+I t -I T 0 w1 d L A
•
2E 0 1 2 3 1 2E 0 1
2 1 1 2E 0 1
2 1 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€ -
3 w 21 5 w 21 3 w 21
• 2 • • •
E -2 I T0 w1 d 2 A 1 2 E I T0 w1 -I T0 w2 A 1 d3 A 2 3 A2 d4 A2 E -I T0 W-I t1-I T0 w2 d3 L 1 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ -
16 w 21 3 w 21 2 w 21 6 w21

E -2 I T0 w2 d 4 A 2 y z
€€€€€€€€€ €€€€€€ z
z>
• • • • •2
z
E I T0 W+I t1-I T0 w2 d 3 L 1 A 2 I E -I T0 w2 m 2 A 2 2 E -I T0 w1 -I T0 w2 d 3 A 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€
2 w 21 4 w1 5 w 21 4 w 21

where
Chapter6.nb 173

2 Hw 2i - W 2 L
fi
L i == €€€€€€€€€€€€€€€€
€€€€€ €€€€

The modulation equations:

i 3 I E I T1 s1 +I T1 s2 +I t1 d 1 L 1 A 1 y
:2 I w 1 A ¢1 == 2 I e w 1 j
j € €€€€ z
z
• •
j z+
I E I T1 s1 A 2 d 2 A 1

k {
- A 1 m 1 + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€
w w

i 27 I E -I T1 s2 -I t1 A 1 A 2 d 1 d 2 L 1
1 1

2 I e 2 w1 j j
j - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
23 I E -I T1 s2 -I t1 A 1 A 2 d 2 d 3 L 1

k
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ €€€€ -
4 w1 3 30 w 31
81 I A1 d 21 L21 23 I A 1 d 22 L 21 3 I E -I T1 s2 -I t1 A 1 A 2 a 2 L 1 12 I A 1 a 1 L 21 I A 1 m 21
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€ €€€€ €€€ -
4 w 31 30 w 31 w1 w1 2 w1
• • • •
15 I A21 d 21 A1 I E I T1 s1 +I T1 s2+I t2 f 2 d 2 A 1 9 I A 21 d 22 A 1 E I T1 s1 A 2 d 2 m 1 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€ -
w1 3 32 w 1 3 16 w 1 3 w 21
• • • •
7 E I T1 s1+I T1 s2+I t1 d 1 L 1 m 1 A 1 E I T1 s1 A2 d 2 m 2 A 1 6 I A 21 a 1 A 1 I A 1 A 2 d 22 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
w 21 4 w 21 w1 4 w 31
• • • •
6 I A1 A2 d1 d3 A2 4 I A 1 A 2 d 23 A 2 3 I A1 A2 d2 d4 A2 27 I E I T1 s2+I t1 A 1 d 1 d 2 L 1 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
w 31 15 w 31 2 w 31 4 w 31
3 I E I T1 s2 +I t1 A 1 a 2 L 1 A 2 y
€€€€€ € zz,
• • •
z
23 I EI T1 s2+I t1 A 1 d 2 d 3 L 1 A 2 2 I A 1 A 2 a3 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
30 w 31 w1 w1

i 1 2 y
2 I w 2 A ¢2 == 2 I e j - A 2 m 2 - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€ z w 2 +
I T1 s2 +I t2 f -I T1 s1 A 2 d

k {
I E 2 I E
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€

i
4 w2 2 w2

2 I e 2 w2 j j
j €€€€€€€€€€€€€€€€
E I T1 s2+I t2 f m E -I T1 s1 A 2 d m I A m 2 3 I E -I T1 s2-I t1 A 22 a 4 L 1 2 I A 2 a 3 L 21

k
2 2 1 2 2 2 2
€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€€ - €€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ +
32 w 21 16 w 21 4 w1 2 w2 w2

3 I E I T1 s2+I t1 a 2 L 31 29 I E -I T1 s2 -I t1 A 22 d 2 d 3 L 1 5 I E -I T1 s2-I t1 A 22 d 3 d 4 L 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€ -
2 w2 15 w 1 w 2
2 4 w 21 w 2
28 I A2 d 22 L21 6 I A 2 d 1 d 3 L 21 I A 2 d 23 L 21 3 I A 2 d 2 d 4 L 21 29 I E I T1 s2+I t1 d 1 d 2 L 31
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € -
15 w 1 w 2
2 w1 w2
2 3 w1 w2 2 2 w1 w2
2 5 w 21 w 2
• •
5 I E I T1 s2+I t1 d 2 d 3 L 31 E -I T1 s1 A 21 d 2 m 1 2 I A1 A2 a3 A1 3 I E I T1 s2 +I t1 A 1 a 2 L 1 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€ -
12 w 21 w 2 2 w1 w2 w2 w2
• • • •
I A1 A2 d2 A1 2 6 I A1 A2 d1 d3 A1 4 I A1 A2 d3 A1 2 3 I A1 A2 d2 d4 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ -
4 w 21 w 2 w 21 w 2 15 w21 w 2 2 w 21 w 2
• • •
27 I EI T1 s2+I t1 A 1 d 1 d 2 L 1 A 1 23 I E I T1 s2+I t1 A 1 d 2 d 3 L 1 A 1 6 I A 22 a 5 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ +
4 w 21 w 2 30 w 21 w 2 w2
• • • •
3 I E I T1 s2+I t1 A 2 a 4 L 1 A 2 I E 2 I T1 s2 +2 I t1 a 3 L 21 A 2 29 I A 22 d 23 A 2 15 I A 22 d 24 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ €€€ -
w2 w2 15 w 21 w 2 4 w 21 w 2
• • •
58 I EI T1 s2+I t1 A 2 d 2 d 3 L 1 A 2 5 I E I T1 s2+I t1 A 2 d 3 d 4 L 1 A 2 2 I E 2 I T1 s2+2 I t1 d 22 L 21 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€ +
15 w 21 w 2 2 w 21 w 2 w 21 w 2
I E 2 I T1 s2 +2 I t1 d 2 d 4 L 21 A 2 y
€€€€€€€€ z
z
z>
• • •
I E 2 I T1 s2+2 I t1 d 1 d 3 L 21 A 2 I E 2 I T1 s2 +2 I t1 d 23 L 21 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
5 w 21 w 2 2 w 21 w 2 4 w 21 w 2

882.98 Second, Null<

174 Chapter6.nb

Ÿ 6.1.3 W » w1 and w2 » 2 w1

In this case, we have a combination of a primary resonance of the first mode, a superharmonic resonance of the second
mode, and a two-to-one internal resonance. In order to bring the effects of damping, forcing, and nonlinearity at the same
order, we let

scaling3 = 9m̀ n_ -> e m n , F1 -> e2 f1 , F2 -> e f2 =;

To describe quantitatively the nearness of the resonances, we introduce the detuning parameters s1 and s2 defined accord-
ing to

ResonanceCond3 = 8w2 == 2 w1 + e s1 , W == w1 + e s2 <;

Using MMS, we obtain the second-order approximate solution and the two equations governing the modulation of the
complex-valued functions Ak :

MMS@scaling3, ResonanceCond3D •• Timing

The second-order approximate solution:

Chapter6.nb 175

:u 1 @tD == e IE I T0 w1 A 1 + E -I T0 w1 A 1M +
•

i E -I T0 W-I t1 f 1
j
e2 j
j €€€€€€€€€€€€€€€€
E I T0 W+I t1 f 1 E 2 I T0 w1 A 21 d 1 E I T0 w1+I T0 w2 A 1 A 2 d 2 E 2 I T0 w2 A 22 d 3

k
€€€€€€€€
€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€ + €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ -
8 w 21 8 w 21 w 21 4 w21 15 w 21

2 E -I T0 W-I t2+I T0 w1 A 1 d 2 L 2 2 E I T0 W+I t2+I T0 w1 A 1 d 2 L 2 E -I T0 W-I t2 +I T0 w2 A 2 d 3 L 2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ +
w12 3 w1 2 2 w 21
E I T0 W+I t2+I T0 w2 A 2 d 3 L 2 2 d 3 L 22 E -2 I T0 W-2 I t2 d 3 L 22 E2 I T0 W+2 I t2 d3 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€2€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ € -
4 w1 2 w1 3 w1 2 3 w 21
• • •
I E I T0 w1 A 1 m 1 6 A1 d1 A1 E -I T0 w1 +I T0 w2 A 2 d 2 A 1 2 E -I T0 W-I t2-I T0 w1 d 2 L 2 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ €€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ -
2 w1 w 21 2 w 21 3 w 21
• • •2 • •
2 E I T0 W+I t2-I T0 w1 d 2 L 2 A 1 I E -I T0 w1 m 1 A 1 E -2 I T0 w1 d 1 A 1 E I T0 w1 -I T0 w2 A 1 d 2 A2 2 A2 d3 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€ €€€€€€ +
w 21 2 w1 w 21 2 w 21 w 21

E -2 I T0 w2 d 3 A 2 y z
€€€€€€€€€ €€€€€€ z
• • • • •
z,
2
E -I T0 W-I t2-I T0 w2 d 3 L 2 A 2 E I T0 W+I t2 -I T0 w2 d 3 L 2 A 2 E -I T0 w1 -I T0 w2 d2 A 1 A2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€ + €€€€€€€€€€€€€€€€
4 w 21 2 w 21 4 w 21 15 w 21

u 2 @tD == e IE I T0 w2
A2 + E -I T0 W-I t2
L2 + E I T0 W+I t2
L2 + E -I T0 w2 •
A2 M +
i
j E
e2 j
j - €€€€€€€€€€€€€€€€
2 I T0 w1 A 2 d 2 E I T0 w1 +I T0 w2 A 1 A 2 d 3 E 2 I T0 w2 A 22 d 4

k
1 2
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ -
16 w 1 2 5 w1 2 4 w 21

E -I T0 W-I t2+I T0 w1 A 1 d 3 L 2 E I T0 W+I t2 +I T0 w1 A 1 d 3 L 2 2 E -I T0 W-I t2 +I T0 w2 A 2 d 4 L 2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€ +
2 w 21 8 w 21 w 21
6 E I T0 W+I t2+I T0 w2 A 2 d 4 L 2 3 d 4 L 22 3 E -2 I T0 W-2 I t2 d 4 L 22 3 E 2 I T0 W+2 I t2 d 4 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ - €€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€€€ -
5 w 21 2 w 21 16 w 21 16 w 21
•
I E I T0 w2 A 2 m 2 2 I E -I T0 W-I t2 L 2 m 2 2 I E I T0 W+I t2 L 2 m 2 A1 d2 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€ €€€€€€€€€ -
4 w1 3 w1 3 w1 2 w 21
• • •
2 E -I T0 w1+I T0 w2 A 2 d 3 A 1 E -I T0 W-I t2 -I T0 w1 d 3 L 2 A 1 E I T0 W+I t2 -I T0 w1 d 3 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ -
3 w 21 8 w 21 2 w 21
•2 • • •
E -2 I T0 w1 d 2 A 1 2 E I T0 w1 -I T0 w2 A 1 d3 A 2 3 A2 d4 A2 6 E-I T0 W-I t2 -I T0 w2 d 4 L 2 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ -
16 w 21 3 w 21 2 w 21 5 w 21

E -2 I T0 w2 d 4 A 2 y z
€€€€€€€€€ €€€€€€ z
z>
• • • • •2
2 E I T0 W+I t2-I T0 w2 d 4 L 2 A 2 I E -I T0 w2 m 2 A 2 2 E -I T0 w1 -I T0 w2 d 3 A 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€
w 21 4 w1 5 w 21 4 w 21

where

2 Hw i - W 2 L
fi
L i == €€€€€€€€€€€€€€€€
€€€€€ €€€€
2

The modulation equations:

176 Chapter6.nb

:2 I w 1 A ¢1 ==

i
j I E I T1 s1 A 2 d 2 A 1 y
2 I e w1 j €€€€ z
z
•
j - A 1 m 1 - €€€€€€€€€€€€€€€€ z + 2 I e w1
I E I T1 s2+I t1 f 1 I E I T1 s1-I T1 s2 -I t2 A 2 d 3 L 2

k {
2
€€€€€€€€€ €€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
4 w1 w1 w1

i
j
j
j €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
I E I T1 s1-I T1 s2-I t1 A 2 f 1 d 2 5 I E -I T1 s2-I t2 A 21 d 1 d 2 L 2 9 I E -I T1 s2 -I t2 A 21 d 2 d 3 L 2 4 I A 1 d 22 L 22

k
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€ €€€€€ -
8 w1 3 w13 16 w 1 3 3 w 31
6 I A 1 d 1 d 3 L 22 5 I A 1 d 23 L 22 3 I A 1 d 2 d 4 L 22 5 I E I T1 s2 +I t2 d 2 d 3 L 32 27 I E I T1 s2 +I t2 d 3 d 4 L 32
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € +
w13 8 w1 3 2 w1 3 3 w1 3 16 w 31
E I T1 s2 +I t1 f 1 m 1 E I T1 s1-I T1 s2 -I t2 A 2 d 3 L 2 m 1 5 E I T1 s1 -I T1 s2-I t2 A 2 d 3 L 2 m 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ +
8 w1 2 2 w1 2 12 w 21
•
3 I E -I T1 s2-I t2 A 21 a 2 L 2 2 I A 1 a 3 L 22 3 I E I T1 s2+I t2 a 4 L 32 I A 1 m 21 15 I A 21 d 21 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ € - €€€€€€€€€ €€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
2 w1 w1 2 w1 2 w1 w 31
• s +I t • s +I t •
9 I A1 d2 A1
2 2 10 I E 1 2 I T 2 A d d L A
1 1 2 2 1 9IE 1 2 I T 2 A d d L A
1 2 3 2 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
16 w 1 3 w1 3 8 w1 3

• • •
2 I E 2 I T1 s2+2 I t2 d 22 L 22 A 1 I E 2 I T1 s2+2 I t2 d 1 d 3 L 22 A 1 I E 2 I T1 s2+2 I t2 d 23 L 22 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€
€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€
€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ € €€€€€€€
€ -
w 31 w 31 2 w 31
• • •
3 I E 2 I T1 s2+2 I t2 d 2 d 4 L 22 A 1 E I T1 s1 A 2 d 2 m 1 A 1 E I T1 s1 A 2 d 2 m 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ € +
16 w 31 w 21 4 w 21
• • • •
6 I A 21 a 1 A 1 3 I E I T1 s2 +I t2 A1 a 2 L2 A 1 I E 2 I T1 s2 +2 I t2 a 3 L 22 A 1 I A 1 A 2 d 22 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
w1 w1 w1 4 w 31
• • • s +I t •
6 I A1 A2 d1 d3 A2 4 I A1 A2 d3 A2 2 3 I A1 A2 d2 d4 A2 9IE 1 2 I T 2 A2 d2 d3 L2 A2
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
w 31 15 w 31 2 w 31 4 w 31
3 I E I T1 s2 +I t2 A 2 a 4 L 2 A 2 y
€€€€€ € z
z
• • •
z , 2 I w 2 A 2 ==
23 I EI T1 s2+I t2 A 2 d 3 d 4 L 2 A 2 2 I A 1 A 2 a3 A 2

{
¢
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
10 w 31 w1 w1

i 3 I E -I T1 s1+2 I T1 s2+2 I t2 d 4 L 22 y
2 I e j - A 2 m 2 + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€ z w 2 +
I E -I T1 s1 A 21 d 2 I E -I T1 s1 +I T1 s2 +I t2 A 1 d 3 L 2
k {
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
2 w2 w2 2 w2
i
2 I e 2 w2 j j
j - €€€€€€€€€€€€€€€€
-I T1 s1 A 2 d m -I T1 s1 +I T1 s2 +I t2 A d L m -I T1 s1+2 I T1 s2 +2 I t2 d L 2 m
E E 3 E

k
1 2 2 1 3 2 2 4 2 2
€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€ -
16 w 2
1 8 w 2
1 16 w 2
1

I A 2 m 22 3 I E -I T1 s2 -I t2 A 1 A2 a 4 L2 12 I A 2 a 5 L 22 I E -I T1 s1 +I T1 s2 +I t1 A 1 f 1 d 2
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€ +
4 w1 w2 w2 8 w 21 w 2
I E -I T1 s1+2 I T1 s2+I t1+I t2 f 1 d 3 L 2 9 I E -I T1 s2-I t2 A 1 A 2 d 2 d 3 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ -
8 w1 w2
2 4 w 21 w 2
23 I E-I T1 s2-I t2 A 1 A 2 d 3 d 4 L 2 9 I A 2 d 23 L 22 69 I A 2 d 24 L 22 E -I T1 s1 A 21 d 2 m 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € +
10 w 1 w 2
2 4 w1 w2
2 10 w 1 w 2
2 2 w1 w2
E -I T1 s1+I T1 s2+I t2 A 1 d 3 L 2 m 1 2 E -I T1 s1 +I T1 s2 +I t2 A1 d 3 L2 m 2 2 E -I T1 s1 +2 I T1 s2 +2 I t2 d 4 L 22 m 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€ +
2 w1 w2 3 w1 w2 w1 w2
• • • •
2 I A1 A2 a3 A1 3 I E I T1 s2 +I t2 A 2 a4 L 2 A 1 I A 1 A 2 d 22 A 1 6 I A1 A2 d1 d3 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ € - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ -
w2 w2 4 w 21 w 2 w 21 w 2
• • •
4 I A 1 A 2 d 23 A 1 3 I A1 A2 d2 d4 A1 9 I E I T1 s2+I t2 A 2 d 2 d 3 L 2 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
15 w 21 w 2 2 w21 w 2 4 w 21 w 2
15 I A 22 d 24 A 2 y z>
€€€€€€€€€ €€€ z
• • • •
z
23 I EI T1 s2+I t2 A 2 d 3 d 4 L 2 A 1 6 I A 22 a 5 A2 29 I A 22 d 23 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
10 w 21 w 2 w2 15 w 21 w 2 4 w 21 w 2

860.36 Second, Null<

Chapter6.nb 177

Ÿ 6.1.4 W » w1 + w2 and w2 » 2 w1

In this case, we have a combination resonance and a two-to-one internal resonance. We let

scaling4 = 8m̀ n_ -> e m n , F1 -> e f1, F2 -> e f2<;

ResonanceCond4 = 8w2 == 2 w1 + e s1 , W == w1 + w2 + e s2 <;

Using MMS, we obtain the second-order approximate solution and the two equations governing the modulation of the
complex-valued functions Ak :

MMS@scaling4, ResonanceCond4D •• Timing

The second-order approximate solution:

:u 1 @tD == e IE I T0 w1 A 1 + E -I T0 W-I t1 L 1 + E I T0 W+I t1 L 1 + E -I T0 w1 A 1 M +

•

i
jE
e2 j
j €€€€€€€€€€€€€€€€
2 I T0 w1 A 2 d E I T0 w1+I T0 w2 A 21 d 2 E 2 I T0 w2 A 21 d 3 2 E -I T0 W-I t1+I T0 w1 A 1 d 1 L 1

k
1 1
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ +
w12 4 w1 2 15 w 1 2 w 21

2 E I T0 W+I t1+I T0 w1 A 1 d 1 L 1 E -I T0 W-I t1+I T0 w2 A 1 d 2 L 1 E I T0 W+I t1 +I T0 w2 A 1 d 2 L 1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ -
5 w 21 2 w 21 12 w 21
6 d 1 L 21 3 E -2 I T0 W-2 I t1 d 1 L 21 3 E 2 I T0 W+2 I t1 d 1 L21 2 E -I T0 W-I t2+I T0 w1 A 1 d 2 L 2
€€€€€€€€€2€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€
€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ +
w1 35 w21 35 w 21 3 w 21
2 E I T0 W+I t2+I T0 w1 A 1 d 2 L 2 E -I T0 W-I t2+I T0 w2 A 1 d 3 L 2 E I T0 W+I t2 +I T0 w2 A 1 d 3 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ +
15 w 21 2 w 21 12 w 21
2 E -2 I T0 W-I t1 -I t2 d 2 L 1 L 2 2 E I t1 -I t2 d2 L 1 L 2 2 E -I t1 +I t2 d 2 L1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ +
35 w 21 w 21 w 21
2 E 2 I T0 W+I t1+I t2 d 2 L 1 L 2 2 d 3 L 22 E -2 I T0 W-2 I t2 d 3 L22 E 2 I T0 W+2 I t2 d 3 L 22 I E I T0 w1 A 1 m 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€ €€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
35 w 21 w 21 35 w 21 35 w 21 2 w1
• •
3 I E -I T0 W-I t1 L 1 m 1 3 I E I T0 W+I t1 L 1 m 1 6 A1 d1 A1 E I T0 w1-I T0 w2 A 1 d 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€€ - €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ -
4 w1 4 w1 w 21 2 w 21
• • • •
E -I T0 w1+I T0 w2 A 1 d 2 A 1 2 A1 d3 A1 2 E -I T0 W-I t1 -I T0 w1 d 1 L 1 A 1 2 E I T0 W+I t1-I T0 w1 d 1 L 1 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€ +
2 w 21 w 21 5 w 21 w 21
• • •
E -I T0 W-I t1-I T0 w2 d 2 L 1 A 1 E I T0 W+I t1 -I T0 w2 d 2 L 1 A 1 2 E -I T0 W-I t2 -I T0 w1 d 2 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€€€ +
12 w 1 2 2 w1 2 15 w 21
• • •
2 E I T0 W+I t2-I T0 w1 d 2 L 2 A 1 E -I T0 W-I t2-I T0 w2 d 3 L 2 A 1 E I T0 W+I t2 -I T0 w2 d 3 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ +
3 w 21 12 w 21 2 w 21

E -2 I T0 w2 d 3 A 1 y z
€€€€€€€€€ €€€€€€ z
• • • •2
z,
2 2
I E -I T0 w1 m 1 A1 E -2 I T0 w1 d 1 A 1 E -I T0 w1 -I T0 w2 d 2 A 1

{
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€
2 w1 w 21 4 w 21 15 w 21

u 2 @tD == e IE I T0 w2 A 2 + E -I T0 W-I t2 L 2 + E I T0 W+I t2 L 2 + E -I T0 w2 A 2 M +

•

i E 2 I T0 w1 A 22 d 2
j
e2 j
j
j
2 E I T0 w1 +I T0 w2 A 22 d 3 E 2 I T0 w2 A 22 d 4 E -I T0 W-I t1 +I T0 w1 A 2 d 2 L 1

k
- €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
€ € + €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ +
16 w 21 5 w 21 4 w 21 8 w 21

E I T0 W+I t1+I T0 w1 A 2 d 2 L 1 2 E -I T0 W-I t1+I T0 w2 A 2 d 3 L 1 2 E I T0 W+I t1 +I T0 w2 A 2 d 3 L 1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ -
6 w1 2 3 w1 2 21 w 21
d 2 L 21 E-2 I T0 W-2 I t1 d 2 L 21 E 2 I T0 W+2 I t1 d 2 L 21 E -I T0 W-I t2 +I T0 w1 A 2 d 3 L 2
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ +
2 w1 2 32 w 1 2 32 w 1 2 8 w 21
178 Chapter6.nb

E I T0 W+I t2+I T0 w1 A 2 d 3 L 2 2 E -I T0 W-I t2+I T0 w2 A 2 d 4 L 2 2 E I T0 W+I t2 +I T0 w2 A 2 d 4 L 2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ +
6 w 21 w 21 7 w 21
E -2 I T0 W-I t1 -I t2 d 3 L 1 L 2 E I t1 -I t2 d 3 L 1 L 2 E -I t1+I t2 d 3 L 1 L 2 E 2 I T0 W+I t1 +I t2 d 3 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ -
16 w 21 2 w 21 2 w 21 16 w 21
3 d 4 L 22 3 E -2 I T0 W-2 I t2 d 4 L 22 3 E 2 I T0 W+2 I t2 d 4 L22 I E I T0 w2 A 2 m 2
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
2 w 21 32 w21 32 w 21 4 w1
• •
6 I E -I T0 W-I t2 L 2 m 2 6 I E I T0 W+I t2 L 2 m 2 A2 d2 A2 2 E I T0 w1-I T0 w2 A 2 d 3 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€€ - €€€€€€€€€ €€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ -
5 w1 5 w1 2 w 21 3 w 21
• • • •
2 E -I T0 w1+I T0 w2 A 2 d 3 A 2 3 A2 d4 A2 E -I T0 W-I t1 -I T0 w1 d 2 L 1 A 2 E I T0 W+I t1-I T0 w1 d 2 L 1 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ +
3 w 21 2 w 21 6 w 21 8 w 21
• • •
2 E -I T0 W-I t1-I T0 w2 d 3 L 1 A 2 2 E I T0 W+I t1-I T0 w2 d 3 L 1 A 2 E -I T0 W-I t2 -I T0 w1 d 3 L 2 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ -
21 w 21 3 w 21 6 w 21
• • •
E I T0 W+I t2-I T0 w1 d 3 L 2 A 2 2 E -I T0 W-I t2-I T0 w2 d 4 L 2 A 2 2 E I T0 W+I t2 -I T0 w2 d 4 L 2 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€ +
8 w 21 7 w 21 w 21

E -2 I T0 w2 d 4 A 2 y z
€€€€€€€€€ €€€€€€ z
z>
• • 2 • 2 •2
I E -I T0 w2 m 2 A2 E -2 I T0 w1 d 2 A 2 2 E -I T0 w1 -I T0 w2 d 3 A 2

{
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
4 w1 16 w 21 5 w 21 4 w 21

where

2 Hw 2i - W 2 L
fi
L i == €€€€€€€€€€€€€€€€
€€€€€ €€€€

The modulation equations:

:2 I w 1 A ¢1 ==

i
j IE 1 1
2 I e2 w1 j
j - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
I T s -I T1 s2 -I t1 A 2 d 2 L I E I T1 s1 -I T1 s2 -I t1 A 22 d 1 d 3 L 1 2 I E I T1 s1-I T1 s2-I t1 A 22 d 23 L 1

k
2 2 1
€€€€€€€€€€€€€€€€
€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ +
2 w 31 5 w 31 3 w 31

I E I T1 s1-I T1 s2-I t1 A 22 d 2 d 4 L 1 54 I A 1 d 21 L 21 11 I A 1 d 22 L 21 13 I E I T1 s1 -I T1 s2 -I t2 A 22 d 2 d 3 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ -
4 w1 3 5 w1 3 24 w 1 3 30 w 31
7 I E I T1 s1-I T1 s2-I t2 A 22 d 3 d 4 L 2 18 I E I t1-I t2 A 1 d 1 d 2 L 1 L 2 18 I E -I t1 +I t2 A1 d 1 d2 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ -
4 w1 3 5 w1 3 5 w 31
11 I EI t1 -I t2 A 1 d 2 d 3 L 1 L 2 11 I E -I t1+I t2 A 1 d 2 d 3 L 1 L 2 4 I A 1 d 22 L 22 6 I A 1 d 1 d 3 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ +
24 w 1 3 24 w 1 3 5 w1 3 w 31
I A 1 d 23 L 22 3 I A 1 d 2 d 4 L 22 I E I T1 s1-I T1 s2-I t1 A 22 a 3 L 1 12 I A 1 a 1 L 21
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ +
24 w 1 3 2 w1 3 w1 w1
3 I E I T1 s1-I T1 s2-I t2 A 22 a 4 L 2 3 I E I t1 -I t2 A 1 a 2 L 1 L 2 3 I E -I t1 +I t2 A 1 a 2 L 1 L 2 2 I A 1 a 3 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ -
2 w1 w1 w1 w1
• • • • •
I A 1 m 21 15 I A 21 d 21 A 1 9 I A 21 d 22 A 1 E I T1 s1 A 2 d 2 m 1 A 1 E I T1 s1 A 2 d 2 m 2 A 1 6 I A 21 a 1 A 1
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ € + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ +
2 w1 w 31 16 w 31 w 21 4 w 21 w1
• 2 • 2 •2
9 I E I T1 s1+I T1 s2+I t1 d 21 L 1 A 1 3 I E I T1 s1+I T1 s2 +I t1 d 22 L 1 A 1 3 I E I T1 s1 +I T1 s2 +I t2 d 1 d 2 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € -
w 31 16 w 31 w 31
•2 •2 •2
3 I E I T1 s1+I T1 s2+I t2 d 2 d 3 L 2 A 1 6 I E I T1 s1 +I T1 s2 +I t1 a 1 L 1 A 1 3 I E I T1 s1+I T1 s2+I t2 a 2 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ -
16 w 31 w1 2 w1
• • • • •
I A1 A2 d2 A2 2 6 I A1 A2 d1 d3 A2 4 I A1 A2 d3 A2 2 3 I A1 A2 d2 d4 A2 5 E T1 s2+I t1 d 2 L 1 m 1 A 2
I
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € -
4 w 31 w 31 15 w 31 2 w 31 4 w 21
y
Chapter6.nb 179

2 I A1 A2 a3 A2 y z
€€€€€€€€€ €€€€ €€ z
• • • •
z+
E I T1 s2 +I t2 d 3 L 2 m 1 A 2 E I T1 s2+I t1 d 2 L 1 m 2 A 2 29 E I T1 s2 +I t2 d 3 L 2 m 2 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€
2 w 21 4 w 21 20 w 21 w1

i y
2 I e w1 j
3 2 2 z
j €€€€€€€€€ €€€€ z
I T1 s1 A d • I T1 s2 +I t1 d L • I T1 s2 +I t2 d L •

j - A 1 m 1 + €€€€€€€€€€€€€€€€ z,
I E A I E A I E A

k {
2 2 1 2 1 2
€€€€€€€€€€€€€€€€
€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w w w

i 3 2 A1 y
1 1 1

2 I w 2 A ¢2 == 2 I e w 2 j j - A 2 m 2 + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ z

z+
I T1 s2+I t1 d L • I T1 s2+I t2 d L •

j z
-I T1 s1 A 2 d
I E I E A I E

k {
1 2 2 1 1
€€€€€€€€€ €€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 w w w

i E -I T1 s1 A 21 d 2 m 2
2 2 2

2 I e 2 w2 j
j
j - €€€€€€€€€€€€€€€€
I A 2 m 22 2 I A 2 a 3 L 21 3 I E I t1-I t2 A 2 a 4 L 1 L 2

k
€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€€ +
16 w 1 2 4 w1 w2 w2

3 I E -I t1 +I t2 A 2 a 4 L 1 L 2 12 I A 2 a 5 L 22 5 I A 2 d 22 L 21 6 I A 2 d 1 d 3 L 21 4 I A 2 d 23 L 21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ -
w2 w2 12 w 21 w 2 w 21 w 2 7 w 21 w 2
3 I A 2 d 2 d 4 L 21 29 I E I t1 -I t2 A2 d 2 d3 L 1 L2 29 I E -I t1 +I t2 A 2 d 2 d 3 L 1 L 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ -
2 w 21 w 2 12 w 21 w 2 12 w 21 w 2
45 I EI t1 -I t2 A 2 d 3 d 4 L 1 L 2 45 I E -I t1+I t2 A 2 d 3 d 4 L 1 L 2 29 I A 2 d 23 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
14 w 1 w 2 2 14 w 1 w 2 2 12 w 21 w 2
• •
135 I A 2 d 24 L 22 E -I T1 s1 A 21 d 2 m 1 E I T1 s2+I t1 d 2 L 1 m 2 A 1 E I T1 s2 +I t2 d 3 L 2 m 2 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € +
14 w 21 w 2 2 w1 w2 8 w 21 8 w 21
• • • • •
2 I A1 A2 a3 A1 I A 1 A 2 d 22 A 1 6 I A1 A2 d1 d3 A1 4 I A 1 A 2 d 23 A 1 3 I A1 A2 d2 d4 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ -
w2 4 w 21 w 2 w 21 w 2 15 w 21 w 2 2 w 21 w 2
• • • •
5 E I T1 s2+I t1 d 2 L 1 m 1 A 1 E I T1 s2+I t2 d 3 L 2 m 1 A 1 6 E I T1 s2 +I t2 d 3 L 2 m 2 A 1 6 I A 22 a 5 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ +
4 w1 w2 2 w1 w2 5 w1 w2 w2
• • •
2 I E -I T1 s1+I T1 s2+I t1 A 1 a 3 L 1 A 2 3 I E -I T1 s1 +I T1 s2+I t2 A 1 a 4 L 2 A 2 29 I A 22 d 23 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
w2 w2 15 w 21 w 2
• • •
15 I A22 d 24 A2 I E -I T1 s1 +I T1 s2+I t1 A 1 d 22 L 1 A 2 2 I E -I T1 s1 +I T1 s2 +I t1 A1 d 1 d3 L 1 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
4 w 21 w 2 w 21 w 2 5 w 21 w2
• •
4 I E -I T1 s1+I T1 s2+I t1 A 1 d 23 L 1 A 2 I E -I T1 s1 +I T1 s2+I t1 A 1 d 2 d 4 L 1 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€ -
3 w 21 w 2 2 w 21 w 2
7 I E -I T1 s1 +I T1 s2 +I t2 A 1 d 3 d 4 L 2 A 2 y
€€€€€€€€€€ zz>
• •
z
13 I E-I T1 s1+I T1 s2+I t2 A 1 d 2 d 3 L 2 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
15 w21 w 2 2 w 21 w 2

8192.266 Second, Null<

Ÿ 6.1.5 W » w2 and w2 » 3 w1

In this case, we have a primary resonance of the second mode, a subharmonic resonance of the first mode, and a three-to-one
internal resonance. We let

scaling5 = 9m̀ n_ -> e2 m n , F1 -> e f1 , F2 -> e3 f2=;

ResonanceCond5 = 9w2 == 3 w1 + e2 s1, W == w2 + e2 s2 =;

Using MMS, we obtain the second-order approximate solution and the two equations governing the modulation of the
complex-valued functions Ak :

MMS@scaling5, ResonanceCond5D •• Timing

The second-order approximate solution:

180 Chapter6.nb

:u 1 @tD == e IE I T0 w1 A 1 + E -I T0 W-I t1 L 1 + E I T0 W+I t1 L 1 + E -I T0 w1 A 1 M +

•

i E 2 I T0 w1 A 21 d 1
j
e2 j
j €€€€€€€€€€€€€€€€
2 E I T0 w1+I T0 w2 A 1 A 2 d 2 E 2 I T0 w2 A 22 d 3 2 E -I T0 W-I t1 +I T0 w1 A 1 d1 L 1

k
€€€€€€€€
€ €€€
€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ €€€
€ + €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€ +
w 21 15 w 21 35 w 21 w 21

2 E I T0 W+I t1+I T0 w1 A 1 d 1 L 1 2 E -I T0 W-I t1+I T0 w2 A 2 d 2 L 1 2 E I T0 W+I t1 +I T0 w2 A 2 d 2 L 1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€ -
5 w1 2 w12 35 w 21
•
6 d 1 L 21 3 E -2 I T0 W-2 I t1 d 1 L 21 3 E 2 I T0 W+2 I t1 d 1 L21 6 A1 d1 A1
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€ +
w 21 35 w21 35 w 21 w 21
• • •
2 E -I T0 w1+I T0 w2 A 2 d 2 A 1 2 E -I T0 W-I t1 -I T0 w1 d 1 L1 A 1 2 E I T0 W+I t1 -I T0 w1 d 1 L 1 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€ +
3 w 21 5 w 21 w 21
•2 • • •
E -2 I T0 w1 d 1 A 1 2 E I T0 w1 -I T0 w2 A 1 d2 A 2 2 A2 d3 A2 2 E-I T0 W-I t1 -I T0 w2 d 2 L 1 A 2
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ -
w 21 3 w 21 w 21 35 w 21

E -2 I T0 w2 d 3 A 2 y z
z,
• • • •2
€€€€€€€€€ €€€€ €€ z
2 E I T0 W+I t1-I T0 w2 d 2 L 1 A 2 2 E -I T0 w1 -I T0 w2 d 2 A 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€
w 21 15 w 21 35 w 21

i
j E
u 2 @tD == e IE I T0 w2 A 2 + E -I T0 w2 A 2M + e 2 j
j - €€€€€€€€€€€€€€€€
2 I T0 w1 A 2 d
• 2 E I T0 w1+I T0 w2 A 1 A 2 d 3 E 2 I T0 w2 A22 d 4

k
1 2
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ -
5 w1 2 7 w1 2 9 w 21

2 E -I T0 W-I t1+I T0 w1 A 1 d 2 L 1 2 E I T0 W+I t1+I T0 w1 A 1 d 2 L 1 2 E -I T0 W-I t1 +I T0 w2 A 2 d3 L 1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€ +
5 w 21 7 w 21 9 w 21
•
2 E I T0 W+I t1+I T0 w2 A 2 d 3 L 1 2 d 2 L 21 E -2 I T0 W-2 I t1 d 2 L 21 E 2 I T0 W+2 I t1 d 2 L 21 2 A1 d2 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ - €€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ € - €€€€€€€€€€€€€€€€
€€€€€ -
27 w 21 9 w 21 27 w 21 27 w 21 9 w 21
• • •
2 E -I T0 w1+I T0 w2 A 2 d 3 A 1 2 E -I T0 W-I t1 -I T0 w1 d 2 L1 A 1 2 E I T0 W+I t1 -I T0 w1 d 2 L 1 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€ -
5 w 21 7 w 21 5 w 21
•2 • • •
E -2 I T0 w1 d 2 A 1 2 E I T0 w1 -I T0 w2 A 1 d3 A 2 2 A2 d4 A2 2 E-I T0 W-I t1 -I T0 w2 d 3 L 1 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ -
5 w 21 5 w 21 3 w 21 27 w 21

E -2 I T0 w2 d 4 A 2 y z
€€€€€€€€€ €€€€ €€ z
z>
• • • •2
2 E I T0 W+I t1-I T0 w2 d 3 L 1 A 2 2 E -I T0 w1 -I T0 w2 d 3 A 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€
9 w 21 7 w 21 9 w 21

where

2 Hw 2i - W 2 L
fi
L i == €€€€€€€€€€€€€€€€
€€€€€ €€€€

The modulation equations:

Chapter6.nb 181

:2 I w 1 A ¢1 ==
i
j
2 I e2 w1 j j - A 1 m 1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
18 I E -I T2 s2 -I t1 A 1 A 2 d 1 d 2 L1 106 I E -I T2 s2-I t1 A 1 A 2 d 2 d 3 L 1 54 I A1 d 21 L21

k
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€
€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ € €€€€€€€
€ €€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
5 w 31 315 w 31 5 w 31
• •
106 I A1 d 22 L21 3 I E -I T2 s2-I t1 A 1 A 2 a 2 L 1 12 I A 1 a 1 L 21 15 I A 21 d 21 A 1 19 I A 21 d 22 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ +
315 w 31 w1 w1 w 31 45 w 31
• • 2 • 2 •2
6 I A 21 a 1 A 1 3 I E I T2 s1 A 2 d 1 d 2 A 1 3 I E I T2 s1 A 2 d 2 d 3 A 1 9 I E I T2 s1 +I T2 s2 +I t1 d 21 L 1 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ -
w1 w 31 5 w 31 w 31
•2 •2 •2 •
3 I E I T2 s1+I T2 s2+I t1 d 22 L 1 A 1 3 I E I T2 s1 A 2 a 2 A 1 6 I E I T2 s1 +I T2 s2 +I t1 a 1 L 1 A 1 4 I A 1 A 2 d 22 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
5 w 31 2 w1 w1 5 w 31
• • • •
6 I A1 A2 d1 d3 A2 4 I A 1 A 2 d 23 A 2 2 I A1 A2 d2 d4 A2 18 I E I T2 s2+I t1 A 1 d 1 d 2 L 1 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
w1 3 35 w 1 3 3 w1 3 5 w 31
3 I E I T2 s2 +I t1 A 1 a 2 L 1 A 2 z y
€€€€€ € z
• • •
z,
106 I EI T2 s2+I t1 A 1 d 2 d 3 L 1 A 2 2 I A 1 A 2 a3 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
315 w 31 w1 w1

i
2 I w 2 A ¢2 == 2 I e 2 w 2 j
j
j - A 2 m 2 - €€€€€€€€€€€€€€€€
I E I T2 s2 +I t2 f 2 I E -I T2 s1 A 31 a 2 3 I E -I T2 s2 -I t1 A 22 a 4 L 1

k
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€€ +
4 w2 2 w2 2 w2

2 I A 2 a 3 L 21 3 I E I T2 s2 +I t1 a 2 L 31 I E -I T2 s1 A 31 d 1 d 2 I E -I T2 s1 A 31 d 2 d 3
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ € - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ -
w2 2 w2 w 21 w 2 5 w 21 w 2
69 I E -I T2 s2-I t1 A 22 d 2 d 3 L 1 5 I E -I T2 s2 -I t1 A 22 d 3 d 4 L 1 68 I A 2 d 22 L 21 6 I A 2 d 1 d3 L 21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ -
35 w 1 w 2
2 9 w1 w2
2 35 w 1 w 2
2 w 21 w 2
4 I A 2 d 23 L 21 2 I A 2 d 2 d 4 L 21 207 I E I T2 s2+I t1 d 1 d 2 L 31 5 I E I T2 s2+I t1 d 2 d 3 L 31
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€ +
27 w 1 w 2
2 3 w1 w2
2 35 w 1 w 22 27 w21 w 2
• • • • •
2 I A1 A2 a3 A1 3 I E I T2 s2+I t1 A 1 a 2 L 1 A 1 4 I A 1 A 2 d 22 A 1 6 I A1 A2 d1 d3 A1 4 I A 1 A 2 d 23 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ € + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
w2 w2 5 w 21 w 2 w 21 w 2 35 w 21 w 2
• • •
2 I A1 A2 d2 d4 A1 18 I E I T2 s2 +I t1 A1 d 1 d 2 L 1 A 1 106 I E I T2 s2 +I t1 A 1 d 2 d 3 L 1 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ +
3 w 21 w 2 5 w 21 w 2 315 w 21 w 2
• • • • •
6 I A 22 a 5 A 2 3 I E I T2 s2 +I t1 A 2 a 4 L 1 A 2 I E 2 I T2 s2 +2 I t1 a 3 L 21 A 2 69 I A 22 d 23 A 2 5 I A 22 d 24 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€ €€€€€ -
w2 w2 w2 35 w 21 w 2 3 w 21 w2
• • •
138 I EI T2 s2+I t1 A 2 d 2 d 3 L 1 A 2 10 I E I T2 s2+I t1 A 2 d 3 d 4 L 1 A 2 2 I E 2 I T2 s2+2 I t1 d 22 L 21 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€ +
35 w 21 w 2 9 w 21 w 2 w 21 w 2
I E 2 I T2 s2+2 I t1 d 2 d 4 L 21 A 2 y
€€€€€€€€ z
z
z>
• • •
3 I E 2 I T2 s2+2 I t1 d 1 d 3 L 21 A 2 2 I E 2 I T2 s2 +2 I t1 d 23 L 21 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
35 w 21 w 2 9 w 21 w 2 9 w 21 w 2

866.906 Second, Null<

Ÿ 6.1.6 W » w2 - 2 w1 and w2 » 3 w1

In this case, we have a primary resonance of the first mode, a superharmonic resonance of the second mode, and a three-to-
one internal resonance. We let

scaling6 = 9m̀ n_ -> e2 m n , F1 -> e3 f1, F2 -> e f2=;

ResonanceCond6 = 9w2 == 3 w1 + e2 s1, W == w2 - 2 w1 + e2 s2 =;

182 Chapter6.nb

Using MMS, we obtain the second-order approximate solution and the two equations governing the modulation of the
complex-valued functions Ak :

MMS@scaling6, ResonanceCond6D •• Timing

The second-order approximate solution:

:u 1 @tD == e IE I T0 w1 A 1 + E -I T0 w1 A 1M +
•

i E 2 I T0 w1 A 21 d 1
j
e2 j
j €€€€€€€€€€€€€€€€
2 E I T0 w1+I T0 w2 A 21 d 2 E 2 I T0 w2 A 21 d 3 2 E -I T0 W-I t2+I T0 w1 A1 d 2 L2

k
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ +
w 21 15 w 21 35 w 21 w 21
2 E I T0 W+I t2+I T0 w1 A 1 d 2 L 2 2 E -I T0 W-I t2+I T0 w2 A 1 d 3 L 2 2 E I T0 W+I t2 +I T0 w2 A 1 d 3 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€ -
3 w1 2 3 w1 2 15 w 21
•
2 d 3 L 22 E -2 I T0 W-2 I t2 d 3 L 22 E 2 I T0 W+2 I t2 d 3 L 22 6 A1 d1 A1
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ € - €€€€€€€€€€€€€€€€ €€€€€€ +
w 21 3 w 21 3 w 21 w 21
• • •
2 E I T0 w1-I T0 w2 A 1 d 2 A 1 2 E -I T0 w1 +I T0 w2 A 1 d 2 A 1 2 A1 d3 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€ €€€€€€ +
3 w 21 3 w 21 w 21
• • •
2 E -I T0 W-I t2-I T0 w1 d 2 L 2 A 1 2 E I T0 W+I t2-I T0 w1 d 2 L 2 A 1 2 E -I T0 W-I t2 -I T0 w2 d 3 L2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€ € €€€€ +
3 w 21 w 21 15 w 21

E -2 I T0 w2 d 3 A 1 z y
€€€€€€€€€ €€€€€€ z
• •2 •2 •2
z,
2 E I T0 W+I t2-I T0 w2 d 3 L 2 A 1 E -2 I T0 w1 d 1 A 1 2 E -I T0 w1 -I T0 w2 d 2 A 1

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
3 w1 2 w12 15 w 1 2 35 w 1 2

u 2 @tD == e IE I T0 w2
A2 + E -I T0 W-I t2
L2 + E I T0 W+I t2
L2 + E -I T0 w2 •
A2 M +
i
j
j - €€€€€€€€€€€€€€€€
e2 j
E 2 I T0 w1 A 22 d 2 2 E I T0 w1 +I T0 w2 A 22 d 3 E 2 I T0 w2 A 22 d 4

k
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ -
5 w1 2 7 w12 9 w 21
2 E -I T0 W-I t2+I T0 w1 A 2 d 3 L 2 2 E I T0 W+I t2+I T0 w1 A 2 d 3 L 2 6 E -I T0 W-I t2 +I T0 w2 A 2 d4 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€ +
9 w 21 5 w 21 5 w 21
6 E I T0 W+I t2+I T0 w2 A 2 d 4 L 2 2 d 4 L 22 3 E -2 I T0 W-2 I t2 d 4 L 22 3 E 2 I T0 W+2 I t2 d 4 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ -
7 w 21 3 w 21 5 w 21 5 w 21
• • • •
2 A2 d2 A2 2 E I T0 w1 -I T0 w2 A 2 d 3 A 2 2 E -I T0 w1+I T0 w2 A 2 d 3 A 2 2 A2 d4 A2
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€ -
9 w 21 5 w 21 5 w 21 3 w 21
• • •
2 E -I T0 W-I t2-I T0 w1 d 3 L 2 A 2 2 E I T0 W+I t2-I T0 w1 d 3 L 2 A 2 6 E -I T0 W-I t2 -I T0 w2 d 4 L2 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ €€€€ -
5 w 21 9 w 21 7 w 21

E -2 I T0 w2 d 4 A 2 y z
€€€€€€€€€ €€€€€€ zz>
• •2 •2 •2
z
6 E I T0 W+I t2-I T0 w2 d 4 L 2 A 2 E -2 I T0 w1 d 2 A 2 2 E -I T0 w1 -I T0 w2 d 3 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
5 w 21 5 w 21 7 w 21 9 w 21

where

2 Hw i - W 2 L
fi
L i == €€€€€€€€€€€€€€€€
2
€€€€€ €€€€

The modulation equations:

:2 I w 1 A ¢1 ==

i
j
2 I e2 w1 j
j - A 1 m 1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
j
5 I E -I T2 s1 -I T2 s2-I t2 A 21 d 1 d 2 L 2 19 I E -I T2 s1 -I T2 s2 -I t2 A21 d 2 d3 L 2 4 I A 1 d 22 L 22

k
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ - €€€€€€€€€€€€€€€€
€€€€€ €€€€€ -
w13 45 w 1 3 3 w 31

6 I A 1 d 1 d 3 L 22 I E -I T2 s1-2 I T2 s2 -2 I t2 A 2 d 2 d 3 L 22 28 I A 1 d 23 L 22
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ -
w 31 w 31 45 w 31
Chapter6.nb 183

2 I A 1 d 2 d 4 L 22 9 I E -I T2 s1 -2 I T2 s2 -2 I t2 A 2 d 3 d 4 L 22 5 I E I T2 s1+I T2 s2 +I t2 d 2 d 3 L 32
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ -
3 w 31 5 w 31 3 w 31
19 I E I T2 s1+I T2 s2+I t2 d 3 d 4 L 32 I E I T2 s1+I T2 s2 +I t1 f 1 3 I E -I T2 s1 -I T2 s2 -I t2 A 21 a 2 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ € +
15 w 31 4 w1 2 w1
•
2 I A 1 a 3 L 22 3 I E -I T2 s1 -2 I T2 s2-2 I t2 A 2 a 4 L 22 3 I E I T2 s1+I T2 s2 +I t2 a 4 L 32 15 I A 21 d 21 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ €€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
w1 2 w1 2 w1 w 31
• • •
19 I A 21 d 22 A 1 10 I E I T2 s1 +I T2 s2+I t2 A 1 d 1 d 2 L 2 A 1 4 I E -I T2 s2-I t2 A 2 d 22 L 2 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€ +
45 w 1 3 w13 3 w 31
• • •
2 I E -I T2 s2-I t2 A 2 d 1 d 3 L 2 A 1 38 I E I T2 s1+I T2 s2+I t2 A 1 d 2 d 3 L 2 A 1 4 I E -I T2 s2 -I t2 A 2 d 23 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€€ -
w 31 45 w 31 5 w 31
• • •
6 I E -I T2 s2-I t2 A 2 d 2 d 4 L 2 A 1 2 I E 2 I T2 s1+2 I T2 s2 +2 I t2 d 22 L 22 A 1 I E 2 I T2 s1 +2 I T2 s2+2 I t2 d 1 d 3 L 22 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ € -
5 w 31 w 31 w 31
s +2 s +2 t • s +2 s +2 t • •
2IE 2 I T 2 1 I T 2 2 I 2 d L A
2 2 3IE 2 I T 2 1 I T 2 2 I 2 d d L A 2 6 I A1 a1 A1
2
3 2 1 2 4 2 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ +
9 w1 3 5 w1 3 w1
• • •
3 I E I T2 s1+I T2 s2+I t2 A 1 a 2 L 2 A 1 2 I E -I T2 s2 -I t2 A 2 a 3 L 2 A 1 I E 2 I T2 s1 +2 I T2 s2 +2 I t2 a 3 L 22 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€€ +
w1 w1 w1
s •2 s •2 s •2 •
3IE I T 2 1 A2 d1 d2 A1 3IE I T 2 1 A2 d2 d3 A1 3IE I T 2 1 A2 a2 A1 4 I A1 A 2 d 22 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
w 31 5 w 31 2 w1 5 w 31
• • • •
6 I A1 A2 d1 d3 A2 4 I A 1 A 2 d 23 A 2 2 I A1 A2 d2 d4 A2 6 I E I T2 s1+I T2 s2+I t2 A 2 d 2 d 3 L 2 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ -
w 31 35 w 31 3 w 31 5 w 31
3 I E I T2 s1 +I T2 s2 +I t2 A 2 a 4 L 2 A 2 y z
z,
• • •
€€ z
106 I EI T2 s1+I T2 s2+I t2 A 2 d 3 d 4 L 2 A 2 2 I A1 A2 a3 A2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ € €€€€
€ + €€€€€€€€€€€€€€€€ €€€€€€€€ € €€€
€ €€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ € €€€€€€€
105 w 31 w1 w1

i
2 I w 2 A ¢2 == 2 I e 2 w 2 j j
j - A 2 m 2 + €€€€€€€€€€€€€€€€
I E -I T2 s1 A 31 a 2 I E I T2 s2 +I t2 A 21 a 3 L 2 3 I E -I T2 s1 -I T2 s2 -I t2 A 1 A 2 a 4 L 2

k
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ +
2 w2 w2 w2
3 I E I T2 s1+2 I T2 s2+2 I t2 A 1 a 4 L 22 12 I A 2 a 5 L 22 2 I E 2 I T2 s1 +3 I T2 s2 +3 I t2 a 5 L 32
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ € +
2 w2 w2 w2
I E -I T2 s1 A 31 d 1 d 2 I E -I T2 s1 A 31 d 2 d 3 2 I E I T2 s2 +I t2 A 21 d 22 L 2 I E I T2 s2 +I t2 A 21 d 1 d 3 L 2
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € -
w 21 w 2 5 w 21 w 2 3 w21 w 2 w 21 w 2
6 I E -I T2 s1-I T2 s2-I t2 A 1 A 2 d 2 d 3 L 2 2 I E I T2 s2+I t2 A 21 d 23 L 2 3 I E I T2 s2+I t2 A 21 d 2 d 4 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ € -
5 w 21 w 2 5 w 21 w 2 5 w 21 w 2
106 I E-I T2 s1-I T2 s2-I t2 A 1 A 2 d 3 d 4 L 2 I E I T2 s1+2 I T2 s2 +2 I t2 A 1 d 2 d 3 L 22 6 I A 2 d 23 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
€€€€€ €€€€€ -
105 w1 w 22 w1 w2
2 5 w 21 w 2
9 I E I T2 s1+2 I T2 s2+2 I t2 A 1 d 3 d 4 L 22 106 I A 2 d 24 L 22 I E 2 I T2 s1 +3 I T2 s2+3 I t2 d 23 L 32
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€ -
5 w1 w2
2 35 w 1 w 22 3 w 21 w 2
• • •
9 I E 2 I T2 s1+3 I T2 s2 +3 I t2 d 24 L 32 2 I A1 A 2 a 3 A 1 3 I E I T2 s1 +I T2 s2 +I t2 A 2 a 4 L 2 A 1 4 I A1 A 2 d22 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ -
5 w 21 w 2 w2 w2 5 w 21 w2
• • • s +I s +I t •
6 I A1 A2 d1 d3 A1 4 I A1 A2 d3 A1 2 2 I A1 A2 d2 d4 A1 6IE 2 1 I T T 2 2 2 A d d L A
2 2 3 2 1
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ -
w 21 w 2 35 w 21 w 2 3 w 21 w 2 5 w 21 w 2
5 I A 22 d 24 A 2 y
€€€€€€€€€ z
z
z>
• • • •
106 I EI T2 s1+I T2 s2+I t2 A 2 d 3 d 4 L 2 A 1 6 I A 22 a 5 A 2 69 I A 22 d 23 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€
105 w 21 w 2 w2 35 w 21 w 2 3 w 21 w 2

872.154 Second, Null<

184 Chapter6.nb

Ÿ 6.1.7 W » w1 + w2 and w2 » w1

In this case, we have a subharmonic resonance of either mode and a one-to-one internal resonance. We let

scaling7 = 8m̀ n_ -> e m n , F1 -> e f1, F2 -> e f2<;

ResonanceCond7 = 8w2 == w1 + e s1 , W == w1 + w2 + e s2 <;

Using MMS, we obtain the second-order approximate solution and the two equations governing the modulation of the
complex-valued functions Ak :

MMS@scaling7, ResonanceCond7D •• Timing

The second-order approximate solution:

:u 1 @tD == e IE I T0 w1 A 1 + E -I T0 W-I t1 L 1 + E I T0 W+I t1 L 1 + E -I T0 w1 A 1 M +

•

i
jE
e2 j
j €€€€€€€€€€€€€€€€
2 I T0 w1 A 2 d 2 E I T0 w1+I T0 w2 A 21 d 2 E 2 I T0 w2 A 21 d 3 3 E -I T0 W-I t1+I T0 w1 A1 d 1 L1

k
1 1
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ +
w12 3 w1 2 3 w1 2 2 w21

3 E I T0 W+I t1+I T0 w1 A 1 d 1 L 1 E -I T0 W-I t1+I T0 w2 A 1 d 2 L 1 E I T0 W+I t1 +I T0 w2 A 1 d 2 L 1 6 d 1 L 21

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€ €€€€€€€ +
4 w 21 2 w 21 4 w 21 w 21
E -2 I T0 W-2 I t1 d 1 L 21 E 2 I T0 W+2 I t1 d 1 L 21 E -I T0 W-I t2 +I T0 w1 A 1 d 2 L 2 E I T0 W+I t2+I T0 w1 A1 d 2 L2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€
€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€
€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ -
5 w 21 5 w 21 2 w 21 4 w 21
E -I T0 W-I t2+I T0 w2 A 1 d 3 L 2 E I T0 W+I t2 +I T0 w2 A 1 d 3 L 2 2 E -2 I T0 W-I t1 -I t2 d 2 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ -
2 w 21 4 w 21 15 w 21
2 E I t1 -I t2 d 2 L 1 L 2 2 E -I t1+I t2 d 2 L 1 L 2 2 E 2 I T0 W+I t1+I t2 d 2 L 1 L 2 2 d 3 L 22
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€ €€€€ €€€ +
w 21 w 21 15 w 21 w 21
E -2 I T0 W-2 I t2 d 3 L 22 E 2 I T0 W+2 I t2 d 3 L 22 I E I T0 w1 A 1 m 1 4 I E -I T0 W-I t1 L 1 m 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ € - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ € +
15 w 21 15 w 21 2 w1 3 w1
• • •
4 I E I T0 W+I t1 L 1 m 1 6 A1 d1 A1 2 E I T0 w1 -I T0 w2 A 1 d 2 A 1 2 E -I T0 w1 +I T0 w2 A 1 d 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € -
3 w1 w 21 w 21 w 21
• • •
2 A1 d3 A1 3 E -I T0 W-I t1-I T0 w1 d 1 L 1 A 1 3 E I T0 W+I t1 -I T0 w1 d 1 L 1 A 1
€€€€€€€€€€€€€€€€
€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ €€ +
w1 2 4 w1 2 2 w 21
• • •
E -I T0 W-I t1-I T0 w2 d 2 L 1 A 1 E I T0 W+I t1 -I T0 w2 d 2 L 1 A 1 E -I T0 W-I t2 -I T0 w1 d 2 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ -
4 w 21 2 w 21 4 w 21
• • •
E I T0 W+I t2-I T0 w1 d 2 L 2 A 1 E -I T0 W-I t2 -I T0 w2 d 3 L 2 A 1 E I T0 W+I t2 -I T0 w2 d 3 L 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ +
2 w 21 4 w 21 2 w 21

E -2 I T0 w2 d 3 A 1 y z
€€€€€€€€€ €€€€€€ z
• • • •2
z,
2 2
I E -I T0 w1 m 1 A1 E -2 I T0 w1 d 1 A 1 2 E -I T0 w1 -I T0 w2 d 2 A 1

{
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
2 w1 w 21 3 w 21 3 w 21

u 2 @tD == e IE I T0 w2 A 2 + E -I T0 W-I t2 L 2 + E I T0 W+I t2 L 2 + E -I T0 w2 A 2 M +

•

i E 2 I T0 w1 A 22 d 2
j
e2 j
j
j
2 E I T0 w1+I T0 w2 A 22 d 3 E 2 I T0 w2 A 22 d 4 E -I T0 W-I t1+I T0 w1 A 2 d 2 L 1

k
€€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
€ € + €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ +
3 w 21 3 w 21 w 21 2 w 21

E I T0 W+I t1+I T0 w1 A 2 d 2 L 1 E -I T0 W-I t1 +I T0 w2 A 2 d 3 L 1 E I T0 W+I t1 +I T0 w2 A 2 d 3 L 1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ -
4 w1 2 2 w1 2 4 w 21
2 d 2 L 21 E -2 I T0 W-2 I t1 d 2 L 21 E 2 I T0 W+2 I t1 d 2 L 21 E -I T0 W-I t2+I T0 w1 A 2 d 3 L 2
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ +
w1 2 15 w 1 2 15 w 1 2 2 w 21
Chapter6.nb 185

E I T0 W+I t2+I T0 w1 A 2 d 3 L 2 3 E -I T0 W-I t2+I T0 w2 A 2 d 4 L 2 3 E I T0 W+I t2 +I T0 w2 A 2 d 4 L 2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€ +
4 w 21 2 w 21 4 w 21
2 E -2 I T0 W-I t1 -I t2 d 3 L 1 L 2 2 E I t1 -I t2 d3 L 1 L 2 2 E -I t1 +I t2 d 3 L1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ +
15 w 21 w 21 w 21
2 E 2 I T0 W+I t1+I t2 d 3 L 1 L 2 6 d 4 L 22 E -2 I T0 W-2 I t2 d 4 L22 E 2 I T0 W+2 I t2 d 4 L 22 I E I T0 w2 A 2 m 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€ €€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
15 w 21 w 21 5 w 21 5 w 21 2 w1
• •
4 I E -I T0 W-I t2 L 2 m 2 4 I E I T0 W+I t2 L 2 m 2 2 A2 d2 A2 2 E I T0 w1-I T0 w2 A 2 d 3 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€€ - €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ -
3 w1 3 w1 w 21 w 21
• • • •
2 E -I T0 w1+I T0 w2 A 2 d 3 A 2 6 A2 d4 A2 E -I T0 W-I t1 -I T0 w1 d 2 L 1 A 2 E I T0 W+I t1-I T0 w1 d 2 L 1 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ +
w 21 w 21 4 w 21 2 w 21
• • •
E -I T0 W-I t1-I T0 w2 d 3 L 1 A 2 E I T0 W+I t1 -I T0 w2 d 3 L 1 A 2 E -I T0 W-I t2 -I T0 w1 d 3 L 2 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ -
4 w1 2 2 w1 2 4 w 21
• • •
E I T0 W+I t2-I T0 w1 d 3 L 2 A 2 3 E -I T0 W-I t2-I T0 w2 d 4 L 2 A 2 3 E I T0 W+I t2 -I T0 w2 d 4 L 2 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€ +
2 w 21 4 w 21 2 w 21

E -2 I T0 w2 d 4 A 2 y z
€€€€€€€€€ €€€€€€ z
z>
• • 2 • 2 •2
I E -I T0 w2 m 2 A2 E -2 I T0 w1 d 2 A 2 2 E -I T0 w1 -I T0 w2 d 3 A 2

{
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€
2 w1 3 w 21 3 w 21 w 21

where

2 Hw 2i - W 2 L
fi
L i == €€€€€€€€€€€€€€€€
€€€€€ €€€€

The modulation equations:

i 81 I A 1 d 21 L 21
:2 I w 1 A ¢1 == 2 I e 2 w 1 j j
j
27 I E I T1 s1 A 2 d 1 d 2 L 21 9 I A 1 d 22 L 21 9 I E I T1 s1 A 2 d 2 d 3 L 21

k
- €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€
€ €€€
€ - €€€€€€€€€€€€€€€€
€€€€
€ €€€€
€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ -
4 w 31 4 w 31 4 w 31 4 w 31
27 I EI t1 -I t2 A 1 d 1 d 2 L 1 L 2 27 I E -I t1+I t2 A 1 d 1 d 2 L 1 L 2 7 I E I T1 s1+I t1 -I t2 A 2 d 22 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ € -
4 w1 3 4 w1 3 4 w 31
5 I E I T1 s1-I t1 +I t2 A 2 d 22 L 1 L 2 3 I E I T1 s1+I t1 -I t2 A 2 d 1 d 3 L 1 L 2 3 I E I T1 s1 -I t1 +I t2 A 2 d 1 d 3 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€ -
2 w1 3 2 w1 3 4 w 31
9 I E I t1 -I t2 A 1 d 2 d 3 L 1 L 2 9 I E -I t1+I t2 A 1 d 2 d 3 L 1 L 2 7 I E I T1 s1+I t1 -I t2 A 2 d 23 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ -
4 w1 3 4 w1 3 4 w 31
5 I E I T1 s1-I t1 +I t2 A 2 d 23 L 1 L 2 3 I E I T1 s1+I t1 -I t2 A 2 d 2 d 4 L 1 L 2 3 I E I T1 s1 -I t1 +I t2 A 2 d 2 d 4 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€ -
2 w 31 2 w 31 4 w 31
I A 1 d 22 L 22 6 I A 1 d 1 d 3 L 22 9 I E I T1 s1 A 2 d 2 d 3 L 22 I A 1 d 23 L 22 6 I A 1 d 2 d 4 L 22
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ -
4 w 31 w 31 4 w 31 4 w 31 w 31
27 I EI T1 s1 A 2 d 3 d 4 L 22 12 I A 1 a 1 L 21 3 I E I T1 s1 A 2 a 2 L 21 3 I E I t1 -I t2 A 1 a 2 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ +
4 w 31 w1 w1 w1
3 I E -I t1 +I t2 A 1 a 2 L 1 L 2 2 I E I T1 s1+I t1 -I t2 A 2 a 3 L 1 L 2 2 I E I T1 s1-I t1+I t2 A 2 a 3 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ €€€€ € +
w1 w1 w1
• •
2 I A 1 a 3 L 22 3 I E I T1 s1 A 2 a 4 L 22 I A 1 m 21 15 I A 21 d 21 A 1 10 I E I T1 s1 A 1 A 2 d 1 d 2 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€ €€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ € -
w1 w1 2 w1 w 31 w 31
• s • s • s •
5 I A1 d2 A1
2 2 2IE 2 I T 1 1 A d A1
2
2 2
2 IE 2 I T 1 1 A d1 d3 A1
2
2 10 I E 1 1 A 1 A 2 d 2 d 3 A 1
I T
€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ € -
3 w 31 w 31 w 31 3 w 31
• • •
2 I E 2 I T1 s1 A 22 d 23 A 1 I E 2 I T1 s1 A22 d 2 d4 A 1 7 E I T1 s1+I T1 s2 +I t1 d 1 L 1 m 1 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€ -
w 31 w 31 w 21
186 Chapter6.nb

• • • •
E I T1 s1 +I T1 s2+I t2 d 2 L 2 m 1 A 1 4 E I T1 s1 +I T1 s2 +I t2 d 2 L 2 m 2 A 1 6 I A 21 a 1 A 1 3 I E I T1 s1 A 1 A2 a 2 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ € + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ +
w 21 3 w 21 w1 w1
s • -I s • • •
IE 2 I T 1 1 A2 a3 A1
2 5IE T 1 1 A1 d1 d2 A2
2 4 I A1 A2 d2 A2 2 6 I A1 A2 d1 d3 A2
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ -
w1 w 31 3 w31 w 31
• • • •
5 I E -I T1 s1 A 21 d 2 d 3 A 2 5 I E I T1 s1 A 22 d 2 d 3 A 2 4 I A 1 A 2 d 23 A 2 6 I A1 A2 d2 d4 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ -
3 w 31 3 w 31 3 w 31 w 31
• • • •
5 I E I T1 s1 A 22 d3 d 4 A2 11 E I T1 s2 +I t1 d 2 L 1 m 1 A 2 E I T1 s2+I t2 d 3 L 2 m 1 A 2 E I T1 s2 +I t1 d 2 L 1 m 2 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ -
w 31 6 w 21 2 w 21 2 w 21
3 I E I T1 s1 A 22 a 4 A 2 y
€€€€€ €€€ z
z
• • • •
z+
11 E I T1 s2+I t2 d 3 L 2 m 2 A 2 3 I E -I T1 s1 A 21 a 2 A 2 2 I A1 A2 a3 A2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ €€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
6 w1 2 2 w1 w1 2 w1

i
2 I e w1 j j - A 1 m 1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
I T1 s1 +I T1 s2 +I t1 d L • I T1 s1 +I T1 s2 +I t2 d L •

j
3 I E A I E A

k
1 1 1 2 2 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€€€ +
w1 w1

I E I T1 s2 +I t2 d 3 L 2 A 2 y
€€€€€€€€€ €€€€ z z
• •
z,
I E I T1 s2+I t1 d 2 L 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w1 w1

i
2 I w 2 A ¢2 == 2 I e w 2 j j
• •
j
s +I t
IE I T 1 2 1 d 2 L1 A 1 I E I T1 s2+I t2 d 3 L 2 A 1

k
- A 2 m 2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ € €€€
€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€
€ €€€€ +
w2 w2

3 I E -I T1 s1+I T1 s2 +I t2 d 4 L 2 A 2 y
€€€€€€€€€ €€€€ € z
z
• •
z+
I E -I T1 s1+I T1 s2+I t1 d 3 L 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
w2 w2

i
j - €€€€€€€€
2 I e 2 w2 j j
I A 2 m 22 3 I E -I T1 s1 A 1 a 2 L 21 2 I A 2 a 3 L 21 2 I E -I T1 s1+I t1 -I t2 A 1 a 3 L 1 L 2

k
€ €€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ +
2 w1 w2 w2 w2
2 I E -I T1 s1-I t1 +I t2 A 1 a 3 L 1 L 2 3 I E I t1 -I t2 A 2 a 4 L 1 L 2 3 I E -I t1 +I t2 A2 a 4 L1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ €€€ +
w2 w2 w2
3 I E -I T1 s1 A 1 a 4 L 22 12 I A 2 a 5 L 22 27 I E -I T1 s1 A 1 d 1 d 2 L 21 I A 2 d 22 L 21 6 I A 2 d 1 d 3 L 21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ -
w2 w2 4 w1 w2
2 4 w1 w2
2 w 21 w 2
9 I E -I T1 s1 A 1 d 2 d 3 L 21 I A 2 d 23 L 21 6 I A 2 d 2 d 4 L 21 5 I E -I T1 s1+I t1 -I t2 A 1 d 22 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ -
4 w1 w2
2 4 w1 w2
2 w1 w2
2 2 w 21 w 2
7 I E -I T1 s1-I t1 +I t2 A 1 d 22 L 1 L 2 3 I E -I T1 s1 +I t1-I t2 A 1 d 1 d 3 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ -
4 w1 w2
2 4 w 21 w 2
3 I E -I T1 s1-I t1 +I t2 A 1 d 1 d 3 L 1 L 2 9 I E I t1-I t2 A 2 d 2 d 3 L 1 L 2 9 I E -I t1 +I t2 A2 d 2 d3 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€ €€€€€ -
2 w1 w2
2 4 w1 w2
2 4 w 21 w 2
5 I E -I T1 s1+I t1 -I t2 A 1 d 23 L 1 L 2 7 I E -I T1 s1 -I t1+I t2 A 1 d 23 L 1 L 2 3 I E -I T1 s1 +I t1-I t2 A 1 d 2 d 4 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ -
2 w1 w2
2 4 w1 w2
2 4 w 21 w 2
3 I E -I T1 s1-I t1 +I t2 A 1 d 2 d 4 L 1 L 2 27 I E I t1-I t2 A 2 d 3 d 4 L 1 L 2 27 I E -I t1 +I t2 A 2 d 3 d 4 L 1 L 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€ -
2 w1 w2
2 4 w1 w2
2 4 w 21 w 2
9 I E -I T1 s1 A 1 d 2 d 3 L 22 9 I A 2 d 23 L 22 27 I E -I T1 s1 A 1 d 3 d 4 L 22 81 I A 2 d 24 L 22
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ -
4 w 21 w 2 4 w 21 w 2 4 w 21 w 2 4 w 21 w 2
• • • •
E I T1 s2 +I t1 d 2 L 1 m 2 A 1 E I T1 s2+I t2 d 3 L 2 m 2 A 1 3 I E -I T1 s1 A 21 a 2 A 1 2 I A1 A2 a3 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€ + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ +
2 w 21 2 w 21 2 w2 w2
s • -I s • • •
3IE I T 1 1 A 2 a4 A 1
2 5IE T 1 1 A1 d1 d2 A1
2 4 I A1 A2 d2 A1 2 6 I A1 A2 d1 d3 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€ €€€
€ €€
€ - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€ €€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ -
2 w2 w 21 w 2 3 w 21 w 2 w 21 w 2
• • • •
5 I E -I T1 s1 A 21 d 2 d 3 A 1 5 I E I T1 s1 A 22 d 2 d 3 A 1 4 I A 1 A 2 d 23 A 1 6 I A1 A2 d2 d4 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€ -
3 w 21 w 2 3 w 21 w 2 3 w 21 w 2 w 21 w 2
• • •
5 I E I T1 s1 A 22 d3 d 4 A1 11 E I T1 s2 +I t1 d 2 L 1 m 1 A 1 E I T1 s2+I t2 d 3 L 2 m 1 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € -
w 21 w 2 6 w1 w2 2 w1 w2
Chapter6.nb 187

• • •
4 E I T1 s2+I t2 d 3 L 2 m 2 A 1 E -I T1 s1 +I T1 s2 +I t1 d 3 L 1 m 2 A 2 3 E -I T1 s1+I T1 s2+I t2 d 4 L 2 m 2 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ +
3 w1 w2 2 w 21 2 w 21
• • • •
I E -2 I T1 s1 A 21 a 3 A 2 3 I E -I T1 s1 A 1 A 2 a 4 A 2 6 I A 22 a 5 A 2 2 I E -2 I T1 s1 A 21 d 22 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€ + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ +
w2 w2 w2 w 21 w 2
• • •
I E -2 I T1 s1 A 21 d 1 d 3 A 2 10 I E -I T1 s1 A 1 A 2 d 2 d 3 A 2 2 I E -2 I T1 s1 A 21 d 23 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ -
w 21 w 2 3 w 21 w 2 w 21 w 2
• • • •
5 I A 22 d 23 A 2 I E -2 I T1 s1 A 21 d 2 d 4 A 2 10 I E -I T1 s1 A 1 A 2 d 3 d 4 A 2 15 I A 22 d 24 A 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€ €€€ -
3 w 21 w 2 w 21 w 2 w 21 w 2 w 21 w 2
11 E -I T1 s1 +I T1 s2 +I t2 d 4 L 2 m 2 A 2 y
€€€€€€€€€ €€€€€€€€ €€€ z
z
z>
• • •
4 E -I T1 s1+I T1 s2+I t1 d 3 L 1 m 1 A 2 E -I T1 s1 +I T1 s2 +I t1 d3 L 1 m2 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
3 w1 w2 2 w1 w2 2 w1 w2

8194.159 Second, Null<

Ÿ 6.1.8 W » 2 w2 - w1 and w2 » w1

In this case, we have a primary resonance of either mode and a one-to-one internal resonance. We let

scaling8 = 9m̀ n_ -> e2 m n , F1 -> e3 f1, F2 -> e3 f2 =;

ResonanceCond8 = 9w2 == w1 + e2 s1 , W == 2 w2 - w1 + e2 s2 =;

Using MMS, we obtain the second-order approximate solution and the two equations governing the modulation of the
complex-valued functions Ak :

MMS@scaling8, ResonanceCond8D •• Timing

The second-order approximate solution:

:u 1 @tD == e IE I T0 w1 A 1 + E -I T0 w1 A 1M +
•

i
j
j €€€€€€€€€€€€€€€€
• •
e2 j
E 2 I T0 w1 A 21 d 1 2 E I T0 w1+I T0 w2 A 1 A 2 d 2 E 2 I T0 w2 A 22 d 3 6 A1 d1 A1 2 E -I T0 w1+I T0 w2 A 2 d 2 A 1

k
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ €€€€ + €€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€
€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € +
w 21 3 w 21 3 w 21 w 21 w 21

E -2 I T0 w2 d 3 A 2 y z
€€€€€€€€€ €€€€€€ z
•2 • • • • •2
z,
E -2 I T0 w1 d 1 A 1 2 E I T0 w1 -I T0 w2 A 1 d2 A 2 2 A2 d3 A2 2 E-I T0 w1-I T0 w2 d 2 A 1 A 2

{
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€
w 21 w 21 w 21 3 w 21 3 w 21

i
j E
u 2 @tD == e IE I T0 w2 A 2 + E -I T0 w2 A 2M + e 2 j j €€€€€€€€€€€€€€€€
2 I T0 w1 A 2 d
• 2 E I T0 w1 +I T0 w2 A 1 A 2 d 3

k
1 2
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ +
3 w1 2 3 w 21
• • •2
E 2 I T0 w2 A 22 d 4 2 A1 d2 A1 2 E -I T0 w1 +I T0 w2 A 2 d 3 A 1 E -2 I T0 w1 d 2 A 1
€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ -
w 21 w 21 w 21 3 w 21

E -2 I T0 w2 d 4 A 2 y z
€€€€€€€€€ €€€€€€ z
z>
• • • • •2
2 E I T0 w1-I T0 w2 A 1 d 3 A 2 6 A2 d4 A2 2 E -I T0 w1-I T0 w2 d 3 A 1 A 2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€€ - €€€€€€€€€€€€€€€€
€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€€€ € + €€€€€€€€€€€€€€€€
w 21 w 21 3 w 21 w 21

The modulation equations:

188 Chapter6.nb

:2 I w 1 A ¢1 ==

i
j
2 I e w1 j
• • •
j - A 1 m 1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
I E 2 I T2 s1 +I T2 s2+I t1 f 1 15 I A21 d 21 A1 10 I E I T2 s1 A 1 A 2 d 1 d 2 A 1 5 I A 21 d 22 A 1

k
2
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€ €€€€€ €€€€€ -
4 w1 w 31 w 31 3 w 31
• • •
2 I E 2 I T2 s1 A 22 d 22 A 1 I E 2 I T2 s1 A 22 d 1 d 3 A 1 10 I E I T2 s1 A 1 A 2 d 2 d 3 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ € -
w 31 w 31 3 w 31
• • • •
2 I E 2 I T2 s1 A 22 d 23 A 1 I E 2 I T2 s1 A 22 d 2 d 4 A 1 6 I A 21 a 1 A 1 3 I E I T2 s1 A 1 A 2 a 2 A 1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ € + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € +
w 31 w 31 w1 w1
s • -I s • • •
IE 2 I T 2 1 A a A 2 5IE T 2 1 A d d A2 4 I A1 A2 d2 A2 2 6 I A1 A2 d1 d3 A2
2 3 1 1 1 2 2
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ -
w1 w 31 3 w 31 w 31
• • • •
5 I E -I T2 s1 A 21 d 2 d 3 A 2 5 I E I T2 s1 A 22 d 2 d 3 A 2 4 I A 1 A 2 d 23 A 2 6 I A1 A2 d2 d4 A2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ -
3 w31 3 w 31 3 w 31 w 31
3 I E I T2 s1 A 22 a 4 A 2 z y
€ €€€ z
• • • •
z,
5 I E I T2 s1 A 22 d 3 d 4 A 2 3 I E -I T2 s1 A 21 a 2 A 2 2 I A1 A2 a3 A2

{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€ + €€€€€€€€€€€€€€€€ €€€€€€€€ € €€€
€ €€ + €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€
w 31 2 w1 w1 2 w1

i
2 I w 2 A ¢2 == 2 I e 2 w 2 j j
-I T2 s1 A 2 a • •
j
I T2 s1 +I T2 s2+I t2 f
I E 3 I E A 2 I A 1 2 3 A1
A a

k
2 1 2 1
- A 2 m 2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ € €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ €€ + €€€€€€€€€€€€€€€€€€€€€€€€ € €€€
€ €€ +
4 w2 2 w2 w2
• • • •
3 I E I T2 s1 A 22 a 4 A 1 5 I E -I T2 s1 A 21 d 1 d 2 A 1 4 I A 1 A 2 d 22 A 1 6 I A1 A2 d1 d3 A1
€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ -
2 w2 w 21 w 2 3 w 21 w 2 w 21 w 2
• • • •
5 I E -I T2 s1 A 21 d 2 d 3 A 1 5 I E I T2 s1 A 22 d 2 d 3 A 1 4 I A 1 A 2 d 23 A 1 6 I A1 A2 d2 d4 A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ -
3 w 21 w 2 3 w 21 w 2 3 w 21 w 2 w 21 w 2
• • • •
5 I E I T2 s1 A 22 d 3 d 4 A 1 I E -2 I T2 s1 A 21 a 3 A 2 3 I E -I T2 s1 A 1 A 2 a 4 A 2 6 I A 22 a 5 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ €€€ + €€€€€€€€€€€€€€€€ €€€€€ €€€€€ -
w 21 w 2 w2 w2 w2
-2 s • -2 s • -I s • •
2IE I T 2 1 A1 d2 A2
2 2 IE I T 2 1 A1 d1 d3 A2
2 10 I E T 2 1 A1 A2 d2 d3 A2 2 I E -2 I T2 s1 A 21 d 23 A 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ € - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€ €€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€ -
w 21 w 2 w 21 w 2 3 w 21 w 2 w 21 w 2
15 I A 2 d 4 A 2 y
€€€€€€€€€ €€€ zz>
• • • •
z
-2 s -I s
5 I A2 d3 A2
2 2 IE I T 2 1 A d d A
2 10 I E 2 1 A 1 A 2 d 3 d 4 A 2
T 2 2

{
1 2 4 2
€€€€€€€€€€€€€€€€
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€ €€€€€€€€€ - €€€€€€€€€€€€€€€€
3 w 21 w 2 w 21 w 2 w 21 w 2 w 21 w 2

816.664 Second, Null<

à 6.2 Parametrically Excited Linearly Coupled Systems

Ÿ 6.2.1 Two-Pendulum Oscillator

As an example, we consider a parametrically excited two-pendulum oscillator with a three-to-one internal resonance:

eq621a = 9q1 ²@tD + a q2² @tD + q1 @tD == €€€€ a Hq2@tD - q1@tDL2 q2² @tD + €€€€ q1@tD3 +
1 1

a Hq2 @tD - q1 @tDL q2 ¢ @tD2 - 2 F1 W2 q1 @tD Cos@W tD, q2² @tD + q1 ² @tD + q2 @tD ==
2 6

€€€€ Hq2 @tD - q1 @tDL2 q1 ²@tD + €€€€ q2 @tD3 - Hq2@tD - q1@tDL q1 ¢@tD2 - 2 F1 W2 q2 @tD Cos@W tD=;
1 1
2 6

where

params = 9a -> €€€€€€€ + e2 s1 =;

16
25
Chapter6.nb 189

for small e. Here, the excitation frequency, W, is assumed to be close to twice the natural frequency of the second mode, w2 ,
which is three times the natural frequency of the first mode, w1 . Hence, we define the following lists:

omgList = 8w1 , w2<;

ResonanceCond = 9W == 2 w2 + e2 s2 , w2 == 3 w1=;

omgRule = Solve@ResonanceCond, 8W, #< •• FlattenD@@1DD & •ž omgList •• Reverse

:8W ® e2 s2 + 6 w1, w2 ® 3 w1<, :W ® e2 s2 + 2 w2 , w1 ® €€€€€€ >>

w2
3

In order to bring the effects of forcing and nonlinearity at the same order, we let

scaling = 9F1 -> e2 f1 =;

the two states v1@tD and v2 @tD defined by

For a consistent expansion, we first transform eq621a into a system of four first-order equations. To this end, we introduce

vel = 8q1 ¢@tD -> v1 @tD, q2¢ @tD -> v2@tD<;

Substituting the velocity and acceleration terms, using vel, into eq621a and combining the result with vel, we transform
eq621a into the following set of four first-order equations:

eq621b = 8vel •. Rule -> Equal, eq621a •. D@vel, tD •. vel< •• Transpose •• Flatten

9q¢1 @tD == v1@tD, q1 @tD + v¢1 @tD + a v¢2 @tD ==

-2 W2 Cos@t WD F1 q1@tD + €€€€ q1@tD3 + a v2 @tD2 H-q1@tD + q2@tDL + €€€€ a H-q1@tD + q2@tDL2 v¢2@tD,
1 1

q¢2 @tD == v2@tD, q2 @tD + v¢1 @tD + v¢2 @tD ==

6 2

-2 W2 Cos@t WD F1 q2@tD + €€€€ q2@tD3 - v1 @tD2 H-q1 @tD + q2 @tDL + €€€€ H-q1@tD + q2@tDL2 v¢1@tD=
1 1
6 2

We seek a uniform second-order expansion of the solution of eq621b in the form

solRule =
9qi_ -> ISumAej qi,j@#1, #2, #3D, 8j, 3<E &M, vi_ -> ISumAej vi,j@#1, #2, #3D, 8j, 3<E &M=;

maxOrder = 2;

Using the time scales T0, T1 , and T2, we express the dependent variables and their time derivatives as

multiScales = 8qi_@tD -> qi žž timeScales, vi_ @tD -> vi žž timeScales,

Derivative@1D@u_D@tD -> dt@1D@u žž timeScalesD, t -> T0<;

Substituting the params, scaling, multiScales, and solRule into eq621b, expanding the result for small e, and discarding
terms of order higher than e3, we obtain

eq621c = Heq621b •. params •. scaling •. multiScales •. solRule •• TrigToExp •• ExpandAllL •.

e n_•;n>3 -> 0;
190 Chapter6.nb

Equating coefficients of like powers of e, we obtain

eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eq621c •• Transpose;

To place the linear operator on one side and the nonhomogeneous terms on the other side, we define

eqOrder@i_D := I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •. u_ k_,1 -> u k,iM ==

I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •. u_ k_,1 -> u k,i M - H#@@1DD & •ž eqEps@@iDDL •• Thread

Using the eqOrder[i] and displayRule, we rewrite eqEps in a concise way as

eqOrder@1D •. displayRule
eqOrder@2D •. displayRule
eqOrder@3D •. displayRule

9D0 q1,1 - v1,1 == 0, D0 v1,1 + €€€€€€ HD0v2,1 L + q1,1 == 0, D0q2,1 - v2,1 == 0, D0v1,1 + D0 v2,1 + q2,1 == 0=
16
25

9D0 q1,2 - v1,2 == -HD1q1,1 L, D0v1,2 + €€€€€€€ HD0 v2,2L + q1,2 == -HD1 v1,1 L - €€€€€€ HD1v2,1L,
16 16

D0 q2,2 - v2,2 == -HD1q2,1 L, D0v1,2 + D0 v2,2 + q2,2 == -HD1v1,1 L - D1 v2,1=

25 25

:D0 q1,3 - v1,3 == -HD1q1,2 L - D2 q1,1,

D0 v1,3 + €€€€€€ HD0v2,3 L + q1,3 == -HD1 v1,2L - €€€€€€ HD1v2,2 L - D2 v1,1 - €€€€€€ HD2 v2,1 L -
16 16 16
25 25 25

HD0 v2,1 L s1 - E-I T0 W W2 f1 q1,1 - EI T0 W W2 f1 q1,1 - €€€€€€ v22,1 q1,1 + €€€€€€€ HD0 v2,1L q21,1 + €€€€€€€€€€ +
16 8 q31,1
25 25 6

€€€€€€€ v2,1 q2,1 - €€€€€€ HD0v2,1 L q1,1 q2,1 + €€€€€€ HD0v2,1 L q2,1, D0 q2,3 - v2,3 == -HD1q2,2 L - D2 q2,1,
16 2 16 8 2
25 25 25
D0 v1,3 + D0v2,3 + q2,3 == -HD1 v1,2L - D1v2,2 - D2 v1,1 - D2v2,1 + v21,1 q1,1 + €€€€ HD0v1,1 L q21,1 -
1
2

E-I T0 W W2 f1 q2,1 - EI T0 W W2 f1 q2,1 - v21,1 q2,1 - HD0 v1,1 L q1,1 q2,1 + €€€€ HD0 v1,1L q22,1 + €€€€€€€€€€ >
1 q32,1
2 6

Ÿ First-Order Problem: Linear System

The first-order problem, eqOrder[1], consists of a set of coupled linear homogeneous differential equations. Hence, the
general solution is the solution of the homogeneous equations. To determine the general solution of the homogeneous set,
we rewrite eqOrder[1] as

linearSys = #@@1DD & •ž eqOrder@1D;

linearSys •. displayRule

9D0 q1,1 - v1,1, D0 v1,1 + €€€€€€ HD0v2,1 L + q1,1 , D0q2,1 - v2,1 , D0v1,1 + D0 v2,1 + q2,1=
16
25

Next, we seek a solution of the linearSys in the form

Chapter6.nb 191

assumedForm = 9qi_,1 -> Ici EI w # &M, vi_,1 -> Idi EI w # &M=;

Substituting assumedForm into linearSys and collecting coefficients of ci and d i, we obtain the coefficient matrix as

coefMat = OuterAD, E-I w T0 linearSys •. assumedForm •• Expand, 8c1 , d1, c2, d2 <E

98I w, -1, 0, 0<, 91, I w, 0, €€€€€€€€€€€€€€ =, 80, 0, I w, -1<, 80, I w, 1, I w<=

16 I w
25

whose adjoint is defined by

hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose

The natural frequencies of the two modes involved are

evals = Cases@w •. Solve@Det@coefMatD == 0, wD, n_ •; n > 0D

• !!!
5 • !!!
: €€€€€€€€€€ , 5 >
3

Hence, the right and left eigenvectors of coefMat corresponding to evals are

rightVec = # • #@@1DD & •ž HNullSpace@coefMat •. w -> #D@@1DD & •ž evalsL

I • !!! 5 5 I • !!! •!! 5 I • !!!

::1, €€€€€€€€€€€€€€ , €€€€ , €€€€€€€€€€€€€€€€€€ >, :1, I 5 , - €€€€ , - €€€€€€€€€€€€€€€€€€ >>
5 5 5 5
3 4 12 4 4

leftVec = NullSpace@hermitian@coefMatD •. w -> #D@@1DD & •ž evals

3 I • !5! I • !5!
::- €€€€€€€€€€€€€€€€€€ , €€€€ , - €€€€€€€€€€ , 1>, : €€€€€€€€€€€€€ , - €€€€ , - €€€€€€€€€€ , 1>>
• ! ! • !!!
5 3I 5 I
4 4 5 4 4 5

whose complex conjugates are

ccleftVec = leftVec •. conjugateRule

• !!! • !!!
:: €€€€€€€€€€€€€€€€€ , €€€€ , €€€€€€€€€€ , 1>, :- €€€€€€€€€€€€€€ , - €€€€ , €€€€€€€€€€ , 1>>
4 • !!! • !!!
3I 5 5 3I I 5 5 I
4 5 4 4 5

Based on the rightVec, we can assume the general solution form as

sol1Form = [email protected] @T1 , T2D EI w1 T0 , A2@T1 , T2 D EI w2 T0 =

1 •!! • !!!
9EI T0 w1 A1@T1 , T2D + EI T0 w2 A2 @T1 , T2D, €€€€ I 5 EI T0 w1 A1 @T1, T2D + I 5 EI T0 w2 A2@T1 , T2D,

€€€€ EI T0 w1 A1@T1 , T2 D - €€€€ EI T0 w2 A2@T1 , T2 D, €€€€€€ I • !!!

5 EI T0 w1 A1@T1 , T2D - €€€€ I • !!!
3
5 EIT0 w2 A2 @T1, T2 D=
5 5 5 5

• !!! • !!!
4 4 12 4

where w1 = 5 ‘ 3 and w2 = 5 . Therefore, the general solution of eqOrder[1] can be expressed in pure function form as
192 Chapter6.nb

sol1 = 8q1,1 , v1,1 , q2,1 , v2,1< ->

HFunction@8T0, T1 , T2 <, # + H# •. conjugateRuleL •• EvaluateD & •ž sol1FormL •• Thread;
sol1 •. displayRule

9q1,1 ® Function@8T0 , T1 , T2<, EI T0 w1 A1 + EI T0 w2 A2 + E-I T0 w1 A1 + E-I T0 w2 A2 D, v1,1 ®

• •

1 • !!! • !!! 1 • !!! • !!!

FunctionA8T0, T1, T2 <, €€€€ I 5 EIT0 w1 A1 + I 5 EIT0 w2 A2 - €€€€ I 5 E-I T0 w1 A1 - I 5 E-IT0 w2 A2 E,
• •
3 3

q2,1 ® FunctionA8T0 , T1 , T2<, €€€€ E A1 - €€€€ E-IT0 w2 A2 E,

5 IT0 w1 5 I T0 w2 5 -I T0 w1 • 5 •
A1 - €€€€ E A2 + €€€€ E

v2,1 ® FunctionA8T0 , T1 , T2<,

4 4 4 4

• !!! 5 •!! •!! 5 • !!!

€€€€€€€ I 5 EI T0 w1 A1 - €€€€ I 5 EI T0 w2 A2 - €€€€€€€ I 5 E-I T0 w1 A1 + €€€€ I 5 E-I T0 w2 A2 E=
5 5 • •
12 4 12 4

where A1 and A2 are to be determined from the solvability conditions at the next levels of approximation.

Ÿ Second-Order Problem

Substituting the first-order solution sol1 into the second-order problem, eqOrder[2], yields

order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

order2Eq •. displayRule

:D0 q1,2 - v1,2 == -EI T0 w1 HD1 A1L - EI T0 w2 HD1 A2L - E-I T0 w1 HD1 A1L - E-I T0 w2 HD1 A2L,
• •

D0 v1,2 + €€€€€€ HD0v2,2 L + q1,2 ==

16

4 I EI T0 w1 HD1A1 L 4 I EI T0 w2 HD1A2 L
€€€€€€ - €€€€ I • !!! €€€€€€ - I • !!!
25

5 EIT0 w1 HD1 A1L + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ 5 EI T0 w2 HD1A2 L +

• !!! • !!!
1
- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3

4 I E-IT0 w1 HD1 A1L 4 I E-I T0 w2 HD1A2 L

3 5 5
1 • ! ! -I T0 w1 • !!!
HD L €€€€€€€€ + I 5 E-I T0 w2 HD1A2 L,
• •

• !!! • ! !
• •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ + €€€
€ I 5 E 1 A1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 5 3 5

D0 q2,2 - v2,2 == - €€€€ EI T0 w1 HD1 A1 L + €€€€ EI T0 w2 HD1A2 L - €€€€ E-I T0 w1 HD1A1 L + €€€€ E-I T0 w2 HD1A2 L,
5 5 5 • 5 •
4 4 4 4

- €€€€ I • !!! 5 EI T0 w1 HD1A1 L + €€€€ I • !!! 5 EI T0 w2 HD1A2 L + €€€€ I • !5! E-I T0 w1 HD1 A1L - €€€€ I • !!!
D0 v1,2 + D0v2,2 + q2,2 ==

5 E-I T0 w2 HD1A2 L>

3 1 3 • 1 •
4 4 4 4

For a uniform expansion, we choose D1 A1 and D1 A2 to eliminate the secular or small-divisor terms. To accomplish this, we
first convert the small-divisor terms into secular terms using the rule

expRule1@i_D := Exp@a_D :> ExpAExpand@a •. omgRule@@iDDD •. e2 T0 -> T2 E

To eliminate the terms that produce secular terms (i.e., determine the solvability conditions) from order2Eq, we collect the
terms proportional to EI w1 T 0 and EI w2 T 0 and obtain

ST11 = CoefficientAorder2Eq@@#, 2DD •. expRule1@1D, EI w1 T0 E & •ž Range@4D

3 I AH1,0L @T1, T2D 3 • !!! H1,0L

9-A1 @T1 , T2D, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ @T1 , T2D, - €€€€ I 5 A1 @T1, T2D=
H1,0L 5 H1,0L
• !!!
1
€€€€€€€€ , - €€€€ A1
5 4 4
Chapter6.nb 193

ST12 = CoefficientAorder2Eq@@#, 2DD •. expRule1@2D, EI w2 T0 E & •ž Range@4D

@T1 , T2D
H1,0L
1 • !!! H1,0L
9-A2 @T1 , T2D, - €€€€€€€€€€€€€€€€ @T1 , T2 D, €€€€ I 5 A2 @T1 , T2 D=
H1,0L 5 H1,0L
•€€€€€€€€€€€€€€€€
!!!
I A2
€€€€€ , €€€€ A2
5 4 4

Then, the solvability conditions demand that ST11 and ST12 be orthogonal to every solution of the corresponding adjoint
problem, namely, the components of ccleftVec. The result is

SCond11 = ccleftVec@@1DD.ST11 == 0
• !!! H1,0L
-3 I 5 A1 @T1 , T2 D == 0

SCond12 = ccleftVec@@2DD.ST12 == 0
• !!! H1,0L
I 5 A2 @T1, T2D == 0

Solving SCond11 and SCond12 for D1 A1 and D1 A2, respectively, yields

H1,0L
SCond11Rule = SolveASCond11, A1 @T1 , T2DE@@1DD

9A1 @T1 , T2 D ® 0=
H1,0L

H1,0L
SCond12Rule = SolveASCond12, A2 @T1 , T2DE@@1DD

9AH1,0L
2 @T1 , T2 D ® 0=

SCond1Rule = Join@SCond11Rule, SCond12RuleD;

whose complex conjugate is

ccSCond1Rule = SCond1Rule •. conjugateRule;

Substituting SCond1Rule and ccSCond1Rule into order2Eq, we obtain

order2Eqm = order2Eq •. SCond1Rule •. ccSCond1Rule •• ExpandAll;

order2Eqm •. displayRule

9D0 q1,2 - v1,2 == 0, D0 v1,2 + €€€€€€ HD0v2,2 L + q1,2 == 0, D0q2,2 - v2,2 == 0, D0v1,2 + D0 v2,2 + q2,2 == 0=
16
25

which is a set of homogeneous equations. Then, we express the solution of the second-order equations in pure function form
as

sol2 = 8q1,2 -> H0 &L, v1,2 -> H0 &L, q2,2 -> H0 &L, v2,2 -> H0 &L<;

Ÿ Third-Order Problem

Substituting the first- and second-order solutions into the third-order problem, eqOrder[3], yields

order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;

194 Chapter6.nb

Substituting the expRule1 into the right-hand sides of order3Eq and collecting the terms that could produce secular terms,
the terms proportional to EI w1 T 0 and EI w2 T 0 , we have

ST21 = CoefficientA#@@2DD •. expRule1@1D, EI w1 T0 E & •ž order3Eq;

ST21 •. displayRule

3 I HD2 A1L 5 • !!!

9-HD2A1 L, - €€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€
•

• !!! • !!!
23 • A21 w1 A1 11 •2
5 A1 s1 w1 + €€€€€€€ A21 A1 - €€€€€€€€ €€€€€€€€€€ + €€€€€€€ A2 A1 +
5 12 36 8 5 12

€€€€€€€€€€€€ , - €€€€ HD2A1 L,

•2 •2 • •

• !!! •!! • ! ! • ! !
3 A2 w1 A1 A2 w2 A1 7 • 27 A1 A2 w1 A2 9 A1 A2 w2 A2 5
€€€€€€€€€€€€€€€€
€€€€€€ + €€€€€€€€
€€€€€€€€€€ + €€€€ A1 A2 A2 - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€
4 5 8 5 2 4 5 2 5 4

3 • !!! 1 • !!! 2 3 •!!

- €€€€ I 5 HD2 A1 L + €€€€€€€€€€€€€€€€
2•
965 A1 A1 • 1175 •2 •2
€€€€€ - €€€€€€ 5 A1 w1 A1 - €€€€€€€€€€€€ A2 A1 + €€€€€€€ 5 A2 w1 A1 -

1 •!! 27 • ! ! 9 • !!!
4 1152 32 384 16

€€€€€€€ 5 A2 w2 A1 - €€€€€€ A1 A2 A2 - €€€€€€€ 5 A1 A2 w1 A2 + €€€€ 5 A1 A2 w2 A2=

•2 35 • • •
32 64 16 8

ST22 = CoefficientA#@@2DD •. expRule1@2D, EI w2 T0 E & •ž order3Eq;

ST22 •. displayRule

I HD2 A2L 5 • !!!

:-HD2A2 L, - €€€€€€€€ • !!! • !!! -
A31 A31 w1 3 •
€ €€€€€€€
€ €€€ + €€€€
€ € - €€€€€€€€
€€€€€€€€ €€€€ 5 A2 s1 w2 - €€€€ A1 A2 A1 +
5 36 24 5 4 2

€€€€€€€€€€€ , €€€€ HD2 A2L,

• • •

• !!! • !!! • !!!

3 A1 A2 w1 A1 A1 A2 w2 A1 43 • • 243 A22 w2 A2 5
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€ A22 A2 - EI T2 s2 W2 f1 A2 + €€€€€€€€€€€€€€€€
4 4
1 •!! 1 • !!! 3 3 • !!!
2 5 4 5 8 5

€€€€ I 5 HD2A2 L + €€€€€€€€€€€€€€€

535 A31 35 • •
- €€€€€€ 5 A1 w1 + €€€€€€ A1 A2 A1 + €€€€ 5 A1 A2 w1 A1 -

1 •!! 243 • !!! 2

4 1152 96 64 8

W f1 A2 - €€€€€€€€€ 5 A2 w2 A2 >
• 1315 2 • 5 I T2 s2 2 • •
€€€€€€€ 5 A1 A2 w2 A1 + €€€€€€€€€€€€ A2 A2 + €€€€ E
16 128 4 32

Then, solvability conditions demand that ST21 and ST22 be orthogonal to their corresponding adjoints. The result is

H0,1L
@T1, T2 DE@@1DD •.
• !!! • !!!
SCond21 = SolveAccleftVec@@1DD.ST21 == 0, A1
9w1 -> 5 ’ 3, w2 -> 5 = •• ExpandAll;

SCond21 •. displayRule
• !!!
:D2 A1 ® - €€€€€€€€€ I 5 A1 s1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€ I • !5! A2 A1 + €€€€€€€€€ I • !5! A1 A2 A2 >
• !!!
•
25 353 I 5 A21 A1 11 •2 23 •
432 3456 128 192

H0,1L
@T1, T2 DE@@1DD •.
• !!! • !!! • !!!
SCond22 = SolveAccleftVec@@2DD.ST22 == 0, A2
9W -> 2 5 , w1 -> 5 ’ 3, w2 -> 5 = •• ExpandAll;

SCond22 •. displayRule

9D2 A2 ® €€€€€€€€€ I • !5! A31 + €€€€€€ I • !!!

5 A2 s1 + €€€€€€€ I • !!!
5 A1 A2 A1 - €€€€€€€€€€€€ I • !!!
5 A22 A2 + 10 I • !!!
5 EI T2 s2 f1 A2 =
11 25 23 • 1337 • •
128 16 64 128

SCond2Rule = Join@SCond21, SCond22D;

Chapter6.nb 195

Ÿ Reconstitution

Using the method of reconstitution, Ak ¢ = e D1 Ak + e2 D2 Ak + º, we combine the partial-differential equations SCond1Rule

and SCond2Rule into the following two ordinary-differential equations governing the modulation of the complex-valued
functions Ak :

moduEq = Table@2 I w k A k ¢, 8k, 2<D ==

HTable@2 I w k dt@1D@ A k @T1, T2 DD, 8k, 2<D •. SCond1Rule •. SCond2RuleL •• Thread;
moduEq •. displayRule

i • !!!
j 25 • !!! • !!! • !!! • y
z
:2 I w1 A¢1 == 2 I e2 w1 j j €€ + €€€€€€€€€ I 5 A2 A1 + €€€€€€€€€ I 5 A1 A2 A2z
z
•
j 432 z,
353 I 5 A21 A1 11 •2 23

k {
- €€€€
€ €€€€ I 5 A1 s1 - €€€€€€€€€€€€€€€€
€€€€€€€€
€ €€€€€€€
3456 128 192
2 I w2 A¢2 == 2 I e2 w2
j €€€€€€€€€ I • !!!
i 25 • !!! 23 • !!! 1337 • !!! •! !
5 A31 + €€€€€€ I 5 A2 s1 + €€€€€€ I 5 A1 A2 A1 - €€€€€€€€€€€€ I 5 A22 A2 + 10 I 5 EI T2 s2 f1 A2 y z>
k 128 {
11 • • •
16 64 128

Ÿ 6.2.2 The function MMSC

According to the procedures described in the previous section, we build a function named MMSC (Method of Multiple
Scales for Linearly Coupled Systems) to automate the process.

MMSC@eqs_List, depVar_List, scaling_List, ResonanceCond : 8__Equal<D :=

ModuleA8<,
omgList = 8w1, w2<;
omgRule = Solve@ResonanceCond, 8W, #< •• FlattenD@@1DD & •ž omgList •• Reverse;
mydepVar = 8u1, v1 , u2 , v2<;
solRule = 8ui_ -> HSum@e ^j ui,j@#1, #2, #3D, 8j, 3<D &L,
vi_ -> HSum@e ^j vi,j @#1, #2, #3D, 8j, 3<D &L<;
maxOrder = 2;
multiScales = 8ui_@tD -> ui žž timeScales, vi_@tD -> vi žž timeScales,
Derivative@1D@u_D@tD -> dt@1D@u žž timeScalesD, t -> T0<;
eqa = Heqs •. Thread@depVar -> mydepVarD •. scaling •. multiScales •. solRule ••
TrigToExp •• ExpandAllL •. e^Hn_ •; n > 3L -> 0;
eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eqa •• Transpose;
eqOrder@i_D :=
I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •. u_ k_,1 -> u k,i M == I#@@1DD & •ž eqEps@@1DD •. fs_ -> 0 •.

u_ k_,1 -> u k,i M - H#@@1DD & •ž eqEps@@iDDL •• Thread;

H* First-Order Problem *L
linearSys = #@@1DD & •ž eqOrder@1D;
assumedForm = 8ui_,1 -> Hci E ^HI w #L &L, vi_,1 -> Hdi E ^HI w #L &L<;
coefMat =
196 Chapter6.nb

Outer@D, E^ H-I w T0 L linearSys •. assumedForm •• Expand, 8c1 , d1, c2, d2 <D;

hermitian@mat_? MatrixQD := mat •. conjugateRule •• Transpose;
evals = Cases@w •. Solve@Det@coefMatD == 0, wD, n_ •; n > 0D;
values = Append@omgRule@@1DD •. e -> 0 •. w1 -> evals@@1DD, w1 -> evals@@1DDD;
rightVec = # • #@@1DD & •ž HNullSpace@coefMat •. w -> #D@@1DD & •ž evalsL;
leftVec = NullSpace@hermitian@coefMatD •. w -> #D@@1DD & •ž evals;
ccleftVec = leftVec •. conjugateRule;
order1Eq = eqOrder@1D •. u_i_,1 -> Hui,1@#1D &L;

sol1p =
DSolve@order1Eq, 8u1,1@T0 D, v1,1@T0 D, u2,1@T0 D, v2,1@T0 D<, T0 D@@1DD •. C@_D -> 0 •• Simplify;
fRule = 9fi_ -> 2 Li Iw2i - W^ 2M=;
sol1Form = H#@@2DD & •ž sol1pL + HH# + H# •. conjugateRuleLL & •ž
[email protected] @T1 , T2D E ^HI w1 T0 L, A2 @T1 , T2 D E^HI w2 T0L<LL;
sol1 = 8u1,1, v1,1 , u2,1, v2,1 < -> HFunction@8T0, T1 , T2 <, # •• EvaluateD & •ž
sol1FormL •• Thread;

H* Second-Order Problem *L
order2Eq = eqOrder@2D •. sol1 •• ExpandAll;

Exp@a_D :> Exp@Expand@a •. omgRule@@iDDD •. e ^n_. T0 :> timeScales@@n + 1DDD;

expRule1@i_D :=

ST11 = Coefficient@#@@2DD •. expRule1@1D, E ^HI w1 T0 LD & •ž order2Eq;

ST12 = Coefficient@#@@2DD •. expRule1@2D, E ^HI w2 T0 LD & •ž order2Eq;
SCond1 = 8ccleftVec@@1DD.ST11 == 0, ccleftVec@@2DD.ST12 == 0<;
H1,0L H1,0L
SCond1Rule1 = SolveASCond1, 9A1 @T1, T2 D, A2 @T1 , T2 D=E@@1DD •• ExpandAll;
H* sigRule=Solve@ResonanceCond,8s1 ,s2<D@@1DD; *L
sigRule = Solve@ResonanceCond, s2 D@@1DD;
expRule2 = Exp@a_D :> Exp@a •. 8T1 -> e T0, T2 -> e ^2 T0< •. sigRule •• ExpandD;
SCond1Rule2 = SCond1Rule1 •. expRule2;
ccSCond1Rule2 = SCond1Rule2 •. conjugateRule;
order2Eqm = order2Eq •. SCond1Rule2 •. ccSCond1Rule2 •• ExpandAll;
IfAUnion@#@@2DD & •ž order2EqmD === 80<,
sol2 = 8u1,2 -> H0 &L, v1,2 -> H0 &L, u2,2 -> H0 &L, v2,2 -> H0 &L<,
Return@"The particular solution of the second order
equations needs to be solved, which is not included in this function."D
H*
basicH=
TableA9Ai@T1 ,T2 D E^HI wi T0L,Ai @T1,T2 D E^H-I wi T0 L=,8i,2<E••Flatten;
•

collectForm=Join@basicH,
If@list1=List žž Plus žž H#@@2DD & •ž eqOrder@1DL•.c_ fi_ E^a_->Li E^a;
Head@list1D===List,list1,8<DD;
possibleTerms=JoinAcollectForm,
IfAlist1=List žž Plus žž I#@@2DD & •ž eqOrder@2D•.u_i_,j_ ->H0 &LM•.

E
Chapter6.nb 197

c_ fi_ E^a_->E^a;Head@list1D===List,list1,8<E,
Outer@Times,collectForm,collectFormD••Flatten••UnionE;
ResonantTerms@i_D:=H#•.8a_•;a=!=0->1< & •ž
HE^H-I wi T0 L possibleTerms•.expRule1@iD•.Exp@_ T0+_.D->0LL
possibleTerms••Union••Rest;
RT=8Array@ResonantTerms,2D,Array@ResonantTerms,2D•.conjugateRule<••
Flatten;
NRT=Complement@possibleTerms,RTD;
rRule=Table@MapIndexed@
ri,#2@@1DD ->Coefficient@order2Eqm@@i,2DD,#1D &,RTD,8i,4<D••Flatten;
list1=Join@ccleftVec,Conjugate@ccleftVecDD;
r4Rule=
Table@Solve@list1@@jDD.Table@ri,j ,8i,4<D==0,r4,j D@@1DD,8j,4<D••Flatten;
RTsymbolList@i_D=Table@Gi,j ,8j,Length@RTD<D;

sol2Form=8u1,2 ,v1,2,u2,2 ,v2,2<->

NRTsymbolList@i_D=Table@Li,j ,8j,Length@NRTD<D;

HFunction@8T0,T1 ,T2<,RTsymbolList@#D.RT+NRTsymbolList@#D.NRT••EvaluateD & •ž

Range@4DL••Thread;
coef1=Solve@Coefficient@Subtract žž #•.sol2Form,NRTD==0•.
Exp@_ T0+_.D->0 & •ž order2Eqm,
Array@NRTsymbolList,4D••FlattenD@@1DD•.values••ExpandAll;
eq1=Table@MapIndexed@Coefficient@order2Eqm@@k,1DD•.sol2Form,#1D==
r k,#2@@1DD &,RTD•.Exp@_ T0 +_.D->0,8k,4<D•.r4Rule••Flatten;
coef2=HSolve@eq1,Join žž Array@RTsymbolList,4DD@@1DD••SimplifyL•.values•.
rRule••ExpandAll;
sol2=sol2Form•.Function@8T0 ,T1,T2 <,a_D:>
Function@8T0,T1 ,T2<••Evaluate,a•.coef1•.coef2••Expand••EvaluateD *L
E;

H* Third-Order Problem *L
order3Eq = eqOrder@3D •. sol1 •. sol2 •• ExpandAll;
ST21 = Coefficient@#@@2DD •. expRule1@1D, Exp@I w1 T0DD & •ž order3Eq;
ST22 = Coefficient@#@@2DD •. expRule1@2D, Exp@I w2 T0DD & •ž order3Eq;
SCond2 = 8ccleftVec@@1DD.ST21 == 0, ccleftVec@@2DD.ST22 == 0< •• ExpandAll;
SCond2Rule1 =
H0,1L H0,1L
SolveASCond2, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD •. values •• ExpandAll;

H* Reconstitution *L
moduEq = Table@2 I w k A k ¢, 8k, 2<D ==
HTable@2 I w k dt@1D@ A k @T1, T2 DD, 8k, 2<D •. HSCond1Rule1 •. valuesL •. SCond2Rule1 ••
Collect@#, eD &L •• Thread;
Print@"The second-order approximate solution:"D;
Print@
198 Chapter6.nb

Table@ui @tD == Hui žž timeScales •. solRule •. e^ 3 -> 0 •. sol1 •. sol2 •. displayRuleL,

8i, 2<D •. Thread@mydepVar -> depVarDD;
IfAOr žž Table@HFi • fi •. scalingL === e, 8i, 2<D,
Print@"where"D;
PrintA"Li==fi •H2Hw2i-W^2LL"E
E;
Print@"\nThe modulation equations:"D;
Print@moduEq •. displayRuleD
E

As an example, we check the case in Section 6.2.1:

eqtest = 9
q¢1 @tD == v1 @tD,
v¢1 @tD + a v¢2 @tD + q1 @tD ==

-2 W2 Cos@t WD F1 q1 @tD + €€€€ q1@tD3 + a v2 @tD2 H-q1@tD + q2@tDL + €€€€ a H-q1@tD + q2@tDL2 v¢2@tD,
1 1

q2 @tD == v2 @tD,
6 2
¢

v¢2 @tD + v¢1 @tD + q2 @tD ==

-2 W2 Cos@t WD F1 q2 @tD + €€€€ q2@tD3 - v1 @tD2 H-q1@tD + q2@tDL + €€€€ H-q1@tD + q2@tDL2 v¢1@tD
1 1
6 2
= •. 9a ® €€€€€€€ + e s1 =;
16 2
25

scaling1 = 9F1 -> e2 f1 =;

ResonanceCond1 = 9w2 == 3 w1 , W == 2 w2 + e2 s2 =;

MMSC@eqtest, 8q1 , v1, q2 , v2 <, scaling1, ResonanceCond1D •• Timing

The second-order approximate solution:

:q 1 @tD == e IE I T0 w1 A 1 + E I T0 w2 A 2 + E -I T0 w1 A 1 + E -I T0 w2 A 2M,
• •

q 2 @tD == e i
j €€€€ E I T0 w1 A 1 - €€€€ E I T0 w2 A 2 + €€€€ E -I T0 w1 A 1 - €€€€ E -I T0 w2 A 2 y
z>
k4 {
5 5 5 • 5 •
4 4 4

The modulation equations:

i • !! !
j • y
j - €€€€€ €€€€ I • !!!
:2 I w 1 A ¢1 == 2 I e 2 w 1 j
! • !!!! • !! ! z
€€€€€€€€€ €€€€€€€€€ + €€€€€ €€€€ I 5 A 2 A 1 + €€€€€€€€€ I 5 A 1 A 2 A 2z
•
z,
25 353 I 5 A 21 A 1 11 •2 23

k {
5 A 1 s 1 - €€€€€€€€€€€€€€€€
432 3456 128 192

2 I w 2 A ¢2 == 2 I e 2 w 2

j €€€€€€€€€ I • !! ! • !!!!
i 23 • !!!! • !!!! • ! !!! • y
5 A 31 + €€€€€ € I 5 A 2 s 1 + €€€€€ € I 5 A 1 A 2 A 1 - €€€€€€€€€ €€€ I 5 A 22 A 2 + 10 I 5 E I T2 s2 f 1 A 2 z >
k 128 {
11 25 • 1337 •
16 64 128

87.691 Second, Null<

Chapter 7
Continuous Systems with Cubic
Nonlinearities
Elastic systems, such as beams, plates, and shells, are usually modeled by partial-differential equations with specified
boundary conditions. For small oscillations, the responses of such deformable bodies, continuous or distributed-parameter
systems, can be adequately described by linear equations and boundary conditions. However, as the amplitude of oscilla -
tions increases, nonlinear effects in the governing equations, boundary conditions, or both come into play. The sources of
nonlinearities may be geometric, inertial, material, or damping in nature.

Geometric nonlinearities may be due to one or more of the following: nonlinear relationships among the strains and the
displacements, large rotations, free surfaces in fluids, time-dependent constraints, mid-plane stretching, and large curvatures.
Nonlinear stretching of the midplane of a deformable body accompanies its transverse vibrations if it is supported in such a
way as to restrict the movement of its ends and / or edges. If large-amplitude vibrations are accompanied by large changes in
the curvature, it is necessary to employ a nonlinear relationship between the curvature and the displacement. Inertial nonlin-
earities are caused by concentrated and / or distributed masses, convective accelerations, and Corriolis and centripetal
accelerations. Material nonlinearities occur whenever the constitutive relations are nonlinear, such as the case when the
stresses are nonlinear functions of the strains; the resistive, inductive, and capacitive circuit elements are nonlinear; and the
feedback control forces and moments in servomechanisms are nonlinear. Damping nonlinearities occur due to form drag and
hysteresis.

Since exact solutions are, in general, not available for determining the dynamic responses of nonlinear continuous systems to
external or parametric excitations, recourse has been made to approximate analyses by using either purely numerical tech -
niques, or purely analytical techniques, or a combination of numerical and analytical techniques. Application of purely
numerical techniques to such problems may be costly in terms of computation time and may not reveal some of their intricate
and complicated responses. With purely analytical methods, it may be difficult to treat systems with inhomogeneities or
complicated geometries. With a combination of numerical and analytical techniques, one can determine some of the compli-
cated responses of systems with inhomogeneities and complicated shapes.

The numerical-analytic approaches can be divided into two groups: discretization and direct methods. In the discretization
methods, one postulates the solution in the form

M f HxL q HtL
w[x,t]=Ú m=1 m m

where M is a finite integer. Then, one assumes the spatial functions fm HxL, space discretization, or the temporal functions
q m HtL, time discretization. With this discretization, the qm HtL are usually taken to be harmonic and the method of harmonic
balance is used to obtain a set of nonlinear boundary-value problems for the fm HxL.

With space discretization, the fm HxL (and thus the spatial dependence) are assumed a priori. If the boundary conditions are
homogeneous, the fm HxL are usually taken to be the eigenfunctions of the linearized problem. The method of weighted
residuals or variational principles can then be used to determine a set of ordinary-differential equations governing the q mHtL.
196 Chapter7.nb

The obtained set of ordinary-differential equations can be studied using any of a great number of methods developed for
discrete systems.

The most common implementation of weighted residuals is the Galerkin method. The great majority of recent studies of
forced vibrations assume that the response can be expressed in terms of only the linear modes that are directly or indirectly
excited (Nayfeh and Mook, 1979). For example, if a system is driven near the natural frequency of a linear mode and that mode
is not involved in an internal resonance with other modes, the response is assumed to consist of only that mode. Such an
approach is usually referred to as a single-mode approximation.

In the direct approach, a reduction method, such as the method of multiple scales, is applied directly to the governing partial-
differential equations and associated boundary conditions and no assumptions are made a priori regarding the spatial or
temporal dependence of the response. The principal advantage of this approach is in the treatment of the boundary condi-
tions at higher orders. This approach has been used recently by Nayfeh and Nayfeh (1979), Nayfeh (1975, 1996), Nayfeh and
Asfar (1986), Nayfeh and Bouguerra (1990), Pai and Nayfeh (1990), Raouf and Nayfeh (1990), Nayfeh, Nayfeh, and Mook
(1992), Nayfeh and Nayfeh (1993, 1995), Nayfeh, Nayfeh, and Pakdemirli (1995), Pakdemirli, Nayfeh, and Nayfeh (1995), Chin
and Nayfeh (1996), Nayfeh and Lacarbonara (1997, 1998), Lacarbonara, Nayfeh, and Kreider (1998), Nayfeh, Lacarbonara, and
Chin (1999), and Rega, et al. (1999).

Some of the aforementioned studies show that the discretization and direct approaches yield the same results for systems
with cubic nonlinearities provided that first-order approximations are sought. For systems with quadratic and cubic nonlineari-
ties, the discretization approach might produce quantitative, and in some case qualitative, errors unless many modes are
included in the discretized model. In other words, one has to include as many terms in the discretized model as needed for
convergence. Since we are dealing with systems with cubic nonlinearities in this chapter, we will obtain first-order results
using both the discretization and direct approaches.

à Preliminaries
Off@General::spell1, Integrate::generD

Needs@"Utilities`Notation`"D

To use the method of multiple scales, we introduce different time scales T0 = t , T1 = e t, and T2= e2 t, symbolize them as

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

and form a list of them as follows:

timeScales = 8T0 , T1, T2 <;

In terms of the time scales Tn , the time derivatives become

dt@0D@expr_D := expr; dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, 1<E;

dt@2D@expr_D := Hdt@1D@dt@1D@exprDD •• ExpandL •. ei_•;i>1 -> 0;

In the course of the analysis, we need the complex conjugates of A and G. We define them using the following rule:

conjugateRule = 9A -> A, A -> A, G -> G, G -> G, Complex@0, n_D -> Complex@0, -nD=;
• • • •
Chapter7.nb 197

To manipulate some complicated integrals without Mathematica being choked, we define the following rules:

intRule1 = 8int@fun_, arg2__D :> int@Expand@funD, arg2D<;

intRule2 = 8int@a_ + b_, arg2__D :> int@a, arg2D + int@b, arg2D,
int@e n_. fun_, arg2__D :> e n int@fun, arg2D,
int@a_ fun_, a1_, b1___D :> a int@fun, a1, b1D •; FreeQ@a, First@a1DD,
int@int@a1_, a2_D fun_., a3__D :> int@a1, a2D int@fun, a3D •; FreeQ@Rest@a2D, First@a2DD<;

To represent some of the expressions in a more concise way, we introduce the following display rule:

displayRule = 9Derivative@a_, b__DAw_i_ E@x, __D :>

#2@@1DD-1 &, 8b<E; arg1 =!= 1, arg1, ""E,

SequenceFormAIfAarg1 = Times žž MapIndexedAD #1
wi SequenceForm žž Table@"'",8a<DE,
Derivative@a__D@Ai_D@__D :> SequenceFormATimes žž MapIndexedAD #1
#2@@1DD &, 8a<E, Ai E,

Derivative@a__D@Ai_D@__D :> SequenceFormATimes žž MapIndexedAD #1

#2@@1DD &, 8a<E, Ai E,
• •

w_i_ @x, __D -> wi, Ai_ @__D -> Ai, Ai_ @__D -> Ai,
• •

Exp@a_. + b_. Complex@0, m_D T0 + c_. Complex@0, n_D T0D :> Exp@a + Hm * b + n * cL I T0 D,
int -> Integrate=;

à 7.1 Solvability Conditions and the Concept of Adjoint

In directly attacking continuous systems, one often encounters nonhomogeneous boundary-value problems whose
homogeneous parts have nontrivial solutions. Consequently, the nonhomogeneous problems have solutions only if
solvability or consistency conditions are satisfied (Nayfeh, 1981). In this section, we describe how to determine such
solvability conditions.

We start the discussion with the simple problem

eq1 = y² @xD + p 2 y@xD == p Sin@p xD;

bc1 = 8y@0D == b1 , y@1D == b2<;

The corresponding homogeneous problem has the nontrivial solution

DSolve@8eq1@@1DD == 0<~Join ~ bc1 •. bi_ -> 0, y@xD, xD@@1DD

8y@xD ® -C@1D Sin@p xD<

Hence, the nonhomogeneous problem will not have a solution unless a solvability condition is satisfied. To determine this
solvability condition, we use two approaches.

In the first approach, we find the general solution of eq1 using the function DSolve as

yRule = DSolve@eq1, y@xD, xD@@1DD •• Simplify

:y@xD ® J- €€€€ + C@2DN Cos@p xD - C@1D Sin@p xD>

x
2
198 Chapter7.nb

where C[1] and C[2] are arbitrary constants. Imposing the boundary conditions bc1, we have

bc2 = bc1 •. Flatten@yRule •. 88x -> 0<, 8x -> 1<<D

9C@2D == b1 , €€€€ - C@2D == b2 =

1
2

These equations are inconsistent unless

Equal žž Plus žž HList žž # & •ž bc2L

1
€€€€ == b1 + b2
2

which is the desired solvability condition.

In the second approach, instead of determining the general solution of the homogeneous differential equations and then
enforcing the boundary conditions to determine the solvability conditions, we use the concept of adjoint as described next.
This approach is attractive if one is not interested in determining the solution of the nonhomogeneous problem but inter-
ested only in determining the solvability conditions, as is the case in many applications of perturbation methods.

To determine the adjoint and then the solvability conditions, we need to perform integration by parts of some products. To
accomplish this with Mathematica, we define a function named intByParts according to

intByParts@expr_ •; Head@exprD =!= Equal, depVar_D :=

Module@8intRule1, intRule2, intRule3<,
intRule1 = f_@terms_, x_D :> int@Expand@termsD, xD;
intRule2 = int@term1_ + term2_, x_D :> int@term1, xD + int@term2, xD;
intRule3 = c1_. int@term1_, x_D + c2_. int@term2_, x_D :> int@c1 term1 + c2 term2, xD;
IBP@u_, v_, n_ •; n >= 1, range : 8x_, a_, b_<D := Hu Derivative@n - 1D@vD@xD •. x -> bL -
Hu Derivative@n - 1D@vD@xD •. x -> aL - IBP@D@u, xD, v, n - 1, rangeD;
IBP@u_, v_, 0, range : 8x_, a_, b_<D := int@u v@xD, rangeD;

expr •. intRule1 ••. intRule2 •. int@u_ Derivative@n_D@depVarD@x_D, arg2_D :>

IBP@u, depVar, n, arg2D ••. intRule3 •. int -> Integrate
D

To determine the solvability condition of eq1 and bc1 with this approach, we multiply eq1 by u@xD, integrate the result by
parts from x = 0 to x = 1, and obtain

eq1a = intByParts@int@u@xD #, 8x, 0, 1<D, yD & •ž eq1

à Hp 2 u@xD y@xD + y@xD u² @xDL â x + y@0D u¢ @0D - y@1D u¢ @1D - u@0D y¢ @0D + u@1D y¢ @1D ==
1

p à Sin@p xD u@xD â x
1

To determine the adjoint, we set the coefficient of y@xD in the integrand on the left-hand side of eq1a equal to zero; that is,
Chapter7.nb 199

adjointEq1 = Cases@eq1a@@1DD, Integrate@expr_, _D :> Coefficient@expr, y@xDDD@@1DD == 0

p 2 u@xD + u² @xD == 0

To determine the adjoint boundary conditions, we consider the homogeneous problem (i.e., f @xD = 0, b 1 = 0, and b 2 = 0), use
the adjointEq1, and obtain from eq1a that

bc1a = eq1a •. Integrate -> H0 &L •. Hbc1 •. Equal -> Rule •. bi_ -> 0L

-u@0D y¢@0D + u@1D y¢@1D == 0

To determine the adjoint boundary conditions, we set each of the coefficients of y¢ @0D and y¢ @1D in bc1a equal to zero and
obtain

adjointBC1 =
Solve@Coefficient@bc1a@@1DD, 8y¢@0D, y¢ @1D<D == 0, 8u@0D, u@1D<D@@1DD •. Rule -> Equal

8u@0D == 0, u@1D == 0<

Therefore, the adjoint u is defined by the adjoint system consisting of adjointEq1 and adjointBC1 . Since they are the same as
the homogeneous parts of eq1 and bc1, the problem is said to be self-adjoint. Hence, u@xD = Sin@xD.

adjoint = 8u -> Function@x, Sin@p xDD<

8u ® Function@x, Sin@p xDD<

Once the adjoint problem has been defined, we return to the nonhomogeneous problem to determine the solvability condi-
tion. Substituting for the adjoint and the boundary conditions bc1 into eq1a yields the solvability condition

SolvCond1 = eq1a@@2DD ==
Heq1a@@1DD •. Integrate -> H0 &L •. Hbc1 •. Equal -> RuleL •. HadjointBC1 •. Equal -> RuleLL

p à Sin@p xD u@xD â x == b1 u¢ @0D - b2 u¢@1D

1

SolvCondf = SolvCond1 •. adjoint

p
€€€€ == p b1 + p b2
2

which is the same as the solvability condition obtained with the other method.

Ÿ 7.1.1 Hinged-Clamped Beam

In treating a hinged-clamped beam, we need to determine the adjoint of the boundary-value problem

eq1a = -w2 f@xD - 2 P f²@xD + fH4L@xD == f@xD;

bc1a = 8f@0D == 0, f² @0D == 0, f@1D == 0, f¢@1D == 0<;

200 Chapter7.nb

Multiplying eq1a by u@xD and integrating the result by parts from x = 0 to x = 1, we obtain

eq1b = intByParts@int@u@xD #, 8x, 0, 1<D, fD & •ž eq1a

à H-w2 u@xD f@xD - 2 P f@xD u²@xD + f@xD uH4L@xDL â x - 2 P f@0D u¢@0D + 2 P f@1D u¢ @1D +
1

2 P u@0D f¢@0D - 2 P u@1D f¢@1D - f¢@0D u² @0D + f¢ @1D u²@1D + u¢@0D f² @0D - u¢ @1D f²@1D +
0

f@0D uH3L@0D - f@1D uH3L@1D - u@0D fH3L @0D + u@1D fH3L @1D == à f@xD u@xD â x
1

To determine the adjoint equation, we set the coefficient of f@xD in the integrand on the left-hand side of eq1b equal to zero;
that is,

adjointEq1 = Cases@eq1b@@1DD, Integrate@expr_, _D :> Coefficient@expr, f@xDDD@@1DD == 0

-w2 u@xD - 2 P u²@xD + uH4L@xD == 0

To determine the adjoint boundary conditions, we consider the homogeneous problem, use the adjointEq1, and obtain from
eq1b that

bc1b = eq1b •. Integrate -> H0 &L •. Hbc1a •. Equal -> RuleL

2 P u@0D f¢ @0D - f¢ @0D u² @0D - u¢ @1D f²@1D - u@0D fH3L@0D + u@1D fH3L @1D == 0

To determine the adjoint boundary conditions, we set each of the coefficients of f¢@0D , fH3L@0D , f² @1D, and fH3L@1D in bc1b equal
to zero and obtain

adjointBC1 = SolveACoefficientAbc1b@@1DD, 9f¢ @0D, fH3L@0D, f² @1D, fH3L @1D=E == 0,

8u@0D, u² @0D, u@1D, u¢@1D<E@@1DD •. Rule -> Equal

8u² @0D == 0, u@1D == 0, u¢@1D == 0, u@0D == 0<

Therefore, the adjoint u is defined by the adjoint system consisting of adjointEq1 and adjointBC1 . Since they are the same as
the homogeneous parts of eq1a and bc1a, the problem is said to be self-adjoint.

Once the adjoint problem has been defined, we return to the nonhomogeneous problem to determine the solvability condi-
tion. Substituting for the adjoint and the boundary conditions bc1a into eq1b yields the solvability condition

SolvCond1 = eq1b@@2DD ==
Heq1b@@1DD •. Integrate -> H0 &L •. Hbc1a •. Equal -> RuleL •. HadjointBC1 •. Equal -> RuleLL

à f@xD u@xD â x == 0
1

Ÿ 7.1.2 Cantilever Beam

The boundary-value problem for a cantilever beam can be written as

eq2a = -w2 f@xD + fH4L@xD == f@xD;

Chapter7.nb 201

bc2a = 9f@0D == 0, f¢ @0D == 0, f²@1D == 0, fH3L @1D == 0=;

To determine the adjoint of this problem, we multiply eq2a by u@xD, integrate the result by parts from x = 0 to x = 1, and
obtain

eq2b = intByParts@int@u@xD #, 8x, 0, 1<D, fD & •ž eq2a

à H-w2 u@xD f@xD + f@xD uH4L@xDL â x - f¢ @0D u²@0D + f¢@1D u² @1D + u¢ @0D f²@0D -
1

u¢ @1D f²@1D + f@0D uH3L @0D - f@1D uH3L @1D - u@0D fH3L@0D + u@1D fH3L @1D == à f@xD u@xD â x
1

We set the coefficient of f@xD in the integrand on the left-hand side of eq2b equal to zero and obtain the adjoint equation as

adjointEq2 = Cases@eq2b@@1DD, Integrate@expr_, _D :> Coefficient@expr, f@xDDD@@1DD == 0

-w2 u@xD + uH4L@xD == 0

To determine the adjoint boundary conditions, we consider the homogeneous problem, use the adjointEq2, and obtain from
eq2b that

bc2b = eq2b •. Integrate -> H0 &L •. Hbc2a •. Equal -> RuleL

f¢ @1D u²@1D + u¢@0D f² @0D - f@1D uH3L@1D - u@0D fH3L@0D == 0

To determine the adjoint boundary conditions, we set each of the coefficients of f²@0D , fH3L @0D , f@1D, and f¢ @1D in bc2b equal
to zero and obtain

adjointBC2 = SolveACoefficientAbc2b@@1DD, 9f² @0D, fH3L@0D, f@1D, f¢@1D=E == 0,

9u@0D, u¢ @0D, u²@1D, uH3L @1D=E@@1DD •. Rule -> Equal

8u@0D == 0, u¢ @0D == 0, u²@1D == 0, uH3L @1D == 0<

Therefore, the adjoint u is defined by the adjoint system consisting of adjointEq2 and adjointBC2 . Since they are the same as
the homogeneous parts of eq2a and bc2a, the problem is said to be self-adjoint.

Once the adjoint problem has been defined, we return to the nonhomogeneous problem to determine the solvability condi-
tion. Substituting for the adjoint and the boundary conditions bc2a into eq2b yields the solvability condition

SolvCond2 = eq2b@@2DD ==
Heq2b@@1DD •. Integrate -> H0 &L •. Hbc2a •. Equal -> RuleL •. HadjointBC2 •. Equal -> RuleLL

à f@xD u@xD â x == 0
1

0
202 Chapter7.nb

à 7.2 Hinged-Clamped Beam

Ÿ 7.2.1 EOM and BC's

We consider the nonlinear planar response of a hinged-clamped uniform prismatic beam to a harmonic axial load. The beam is
subjected to a static axial load and one of its ends is restrained by a linear spring. We assume that the curvature and inertia
nonlinearities are much smaller than the nonlinearity caused by the mid-plane stretching. The transverse deflection w@x, tD of
the beam at the position x and time t is governed by the nondimensional integral-partial-differential equation (Nayfeh and
Mook, 1979)

EOM = ¶t,t w@x, tD - 2 P ¶x,x w@x, tD + ¶x,x,x,x w@x, tD ==

-2 e m@xD ¶t w@x, tD + 4 e F ¶x,x w@x, tD Cos@W tD + e a ¶x,x w@x, tD à H¶x w@x, tDL2 â x
1

@x, tD - 2 P w @x, tD + w @x, tD ==

H0,2L H2,0L H4,0L

i 1 H1,0L y H2,0L
w

-2 e m@xD wH0,1L @x, tD + 4 F e Cos@t WD wH2,0L@x, tD + a e j

jà w @x, tD â xz
zw @x, tD
k 0 {
2

and homogeneous boundary conditions

BC = 8w@0, tD == 0, ¶x,x w@x, tD == 0 •. x -> 0, w@1, tD == 0, ¶x w@x, tD == 0 •. x -> 1<

8w@0, tD == 0, wH2,0L @0, tD == 0, w@1, tD == 0, wH1,0L@1, tD == 0<

We assume that the static axial load is such that the lowest two natural frequencies of the beam are in the ratio of three-to-
one; that is, there is a three-to-one internal resonance between the second and first modes. Moreover, we assume that
neither of these two modes is involved in an internal resonance with any other mode. We consider three parametric reso -
nances, namely, principal parametric resonance of the first mode, principal parametric resonance of the second mode, and
combination parametric resonance of the first and second modes.

Ÿ 7.2.2 Direct Attack of the Continuous Problem

In this section, we directly attack the integral-partial-differential equation EOM and associated boundary conditions BC and
seek a first-order uniform expansion of their solution in the form

solRule = w -> IEvaluateASumAej wj @#1, #2, #3D, 8j, 0, 1<EE &M

w ® Hw0 @#1, #2, #3D + e w1 @#1, #2, #3D &L

where the first independent variable stands for x and the last two independent variables stand for the two time scales T0 and
T1 . Substituting the solRule into EOM, transforming the total time derivatives into partial derivatives in terms of T0 and T1 ,
expanding the result for small e, discarding terms of order higher than e, and using the intRule2 to simplify the expansions of
the integrands, we obtain
Chapter7.nb 203

eq722a =
HHJoin@8EOM<, BCD •. Integrate -> int •. 8w@x_, tD -> w@x, T0, T1 D, Derivative@m_, n_D@wD@
x_, tD -> dt@nD@D@w@x, T0, T1 D, 8x, m<DD, t -> T0 < •.
solRule •• ExpandAllL ••. intRule2 •• ExpandAllL •. e n_•;n>1 -> 0;

Equating coefficients of like powers of e in eq722a, we obtain

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq722a •• Transpose;

eqEps •. displayRule

:9-2 P Hw''
0 L + w0
''''
+ D20w0 == 0, w0@0, T0 , T1 D == 0,

wH2,0,0L@0, T0 , T1 D == 0, w0@1, T0 , T1 D == 0, wH1,0,0L @1, T0, T1D == 0=,

:-4 F Cos@T0 WD Hw''

0 L - a Hw0L Hw0 L - 2 P Hw1 L + w1 + D20w1 + 2 HD0 D1w0 L + 2 HD0 w0L m@xD == 0,
0 0
' '' 2 '' ''''

w1 @0, T0, T1D == 0, w1 @0, T0 , T1 D == 0, w1@1, T0 , T1 D == 0, w1 @1, T0, T1D == 0>>

H2,0,0L H1,0,0L

Because in the presence of damping, all modes that are not directly excited by the forcing or indirectly excited by the internal
resonance will decay with time (Nayfeh and Mook, 1979), the solution of eqEps[[1]] can be expressed in terms of the lowest
two linear free-vibration modes; that is,

sol0 =
w0 -> FunctionA8x, T0, T1 <, SumAAi@T1 D fi@xD Exp@I wi T0D + Ai@T1 D fi@xD Exp@-I wi T0 D, 8i, 2<EE;
•

where w1 and w2 are the natural frequencies of these modes. For later use, we define the list

omgList = 8w1 , w2<;

One can easily show that the mode shapes fi@xD are orthogonal. We assume that these modes are normalized so that

Ù0 fi@xD fj @xD â x -> dij

1

where dij is the Kronecker delta function.

Substituting sol0 into the first-order equation, eqEps[[2,1]], and using intRule1 and intRule2, we obtain
204 Chapter7.nb

order1Eq = HeqEps@@1, 1, 1DD •. w0 -> w1 L ==

HHeqEps@@1, 1, 1DD •. w0 -> w1L - HSubtract žž eqEps@@2, 1DD •. sol0 •• TrigToExp ••
ExpandL •. intRule1 ••. intRule2 •• ExpandL;
order1Eq •. displayRule

-2 P Hw''
1 L + w1 + D20w1 == -2 I EI T0 w1 HD1A1 L w1 f1 @xD + 2 I E-I T0 w1 HD1 A1 L w1 f1 @xD -
'''' •

2 I EI T0 w1 A1 w1 m@xD f1 @xD + 2 I E-I T0 w1 w1 A1 m@xD f1 @xD - 2 I EI T0 w2 HD1A2 L w2 f2 @xD +

•

2 I E-I T0 w2 HD1 A2 L w2 f2@xD - 2 I EI T0 w2 A2 w2 m@xD f2 @xD + 2 I E-I T0 w2 w2 A2 m@xD f2 @xD +

• •

i 1 ¢ y 3 ²
2 EI T0 H-W+w1 L F A1 f²1@xD + 2 EIT0 HW+w1L F A1 f²1 @xD + E3 I T0 w1 a j jà f1 @xD2 â xz z A1 f1@xD +
k 0 {
i y i y
2 EI T0 H2 w1 +w2L a jjà f¢1 @xD f¢2@xD â xz z A21 A2 f²1 @xD + EI T0 Hw1 +2 w2L a j jà f¢2 @xD2 â xz z A1 A22 f²1@xD +
k 0 { k 0 {
1 1

i
jà f @xD â xz y
z A A1 f² @xD +
F A1 f1@xD + 2 E F A1 f1 @xD + 3 E aj
I T0 H-W-w1 L • IT0 HW-w1L • ² •
1

k 0 {
² I T0 w1 ¢ 2 2
2E 1 1 1

i
jà f¢1 @xD f¢2 @xD â xy z A1 A2 • i
jà f¢2@xD2 â xy z A22 •
4 EI T0 w2 a j z A1 f²1@xD + EIT0 H-w1+2 w2 L a j z A1 f²1@xD +
k 0 { k 0 {
1 1

i 1 ¢ y i 1 ¢ y
3 E-I T0 w1 a j jà f1@xD2 â xz z A1 A1 f²1 @xD + 2 EI T0 H-2 w1 +w2L a j jà f1 @xD f¢2@xD â xz z A2 A1 f²1 @xD +
•2 •2
k 0 { k 0 {
i y i y
E-3 I T0 w1 a jjà f¢1@xD2 â xz z A1 f²1@xD + 2 EIT0 H2 w1-w2 L a j jà f¢1@xD f¢2 @xD â xz z A21 A2 f²1@xD +
• •
k 0 { k 0 {
1 3 1

i y i y • • ²
2 EI T0 w1 a j jà f¢2 @xD2 â xz z A1 A2 A2 f²1@xD + 4 E-IT0 w2 a j jà f¢1 @xD f¢2@xD â xz z A1 A1 A2 f1 @xD +
•
k 0 { k 0 {
1 1

i y i y •2 • ²
2 E-I T0 w1 a j jà f¢2@xD2 â xz z A2 A1 A2 f²1@xD + 2 EIT0 H-2 w1-w2 L a j jà f¢1@xD f¢2 @xD â xz z A1 A2 f1@xD +
• •
k 0 { k 0 {
1 1

i
jà f @xD â xz y
z A1 A f @xD + E 0 1 2 a j i
jà f @xD â xz y
z A1 A f²@xD +
aj
I T0 Hw1 -2 w2L • H-w L • •
k 0 { k 0 {
1 1
¢ 2 2 ² I T -2 w ¢ 2 2
E 2 2 1 2 2 1

i
jà f¢1 @xD2 â xy z A21 A2 f²2 @xD +
2 EI T0 H-W+w2 L F A2 f²2@xD + 2 EIT0 HW+w2L F A2 f²2 @xD + EI T0 H2 w1 +w2L a j z
k 0 {
1

i y i y 3 ²
2 EI T0 Hw1 +2 w2L a jjà f¢1 @xD f¢2@xD â xz z A1 A22 f²2 @xD + E3 I T0 w2 a j jà f¢2 @xD2 â xz z A2 f2@xD +
k 0 { k 0 {
1 1

i y i y 2• ²
2 EI T0 w2 a j jà f¢1 @xD2 â xz z A1 A2 A1 f²2@xD + 2 EI T0 H-w1+2 w2 L a j jà f¢1@xD f¢2 @xD â xz z A2 A1 f2@xD +
•
k 0 { k 0 {
1 1

i y
EI T0 H-2 w1 +w2L a jjà f¢1 @xD2 â xz z A2 A1 f²2 @xD + 2 EI T0 H-W-w2 L F A2 f²2@xD + 2 EI T0 HW-w2L F A2 f²2 @xD +
• • •
k 0 {
1 2

i y i 1 ¢ y
EI T0 H2 w1 -w2L a j
jà f¢1 @xD2 â xz z A21 A2 f²2 @xD + 4 EI T0 w1 a j jà f1 @xD f¢2@xD â xz z A1 A2 A2 f²2 @xD +
• •
k 0 { k 0 {
1

i
jà f @xD â xz y
z A A2 f @xD + 2 E 0 2 a j i
jà f @xD â xz y
z A1 A1 A2 f²@xD +
aj
• • •
k 0 { k 0 {
1 1
I T0 w2 ¢ 2 2 ² -I T w ¢ 2
3E 2 2 2 1 2

i
jà f¢1@xD f¢2 @xD â xy z A2 • i
jà f¢1@xD2 â xy z•
4 E-I T0 w1 a j z A1 A2 f²2@xD + EIT0 H-2 w1-w2 L a j z A1 A2 f²2@xD +
• 2•

k 0 { k 0 {
1 1

i
jà f¢ @xD f¢ @xD â xy z A1 • i
jà f¢ @xD2 â xy z A2 •
2 EI T0 Hw1 -2 w2L a j z A2 f²2 @xD + 3 E-I T0 w2 a j z A2 f²2 @xD +
k 0 { k 0 {
1 2 1 2
1 2 2

i 1 ¢ y • •2 ² i 1 ¢ y •3 ²
2 EI T0 H-w1 -2 w2L a jjà f1 @xD f¢2 @xD â xz z A1 A2 f2 @xD + E-3 I T0 w2 a j jà f2@xD2 â xz z A2 f2@xD
k 0 { k 0 {

It follows from eqEps[[2]] that the first-order boundary conditions are

Chapter7.nb 205

order1BC = eqEps@@2DD •• Rest

9w1 @0, T0, T1D == 0, w1 @0, T0 , T1 D == 0, w1@1, T0 , T1 D == 0, w1 @1, T0 , T1D == 0=

H2,0,0L H1,0,0L

à Principal Parametric Resonance of the First Mode

In this case, W » 2 w1. To describe the nearness of the internal and principal parametric resonances, we introduce the two
detuning parameters s1 and s2 defined by

ResonanceConds = 8w2 == 3 w1 + e s1 , W == 2 w1 + e s2 <;

and define the following rules:

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

98w2 ® e s1 + 3 w1, W ® e s2 + 2 w1 <, 9w1 ® €€€€ H-e s1 + w2 L, W ® €€€€ H-2 e s1 + 3 e s2 + 2 w2L==

1 1
3 3

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand side of order1Eq and obtain the source of secular terms as

ST = Table@Coefficient@order1Eq@@2DD •. expRule@iD, Exp@I wi T0DD, 8i, 2<D;

ST •. displayRule

:-2 I HD1A1 L w1 f1 @xD - 2 I A1 w1 m@xD f1 @xD + 2 EI T1 s2 F A1 f²1@xD +

•

i 1 ¢ y 2• ² i 1 ¢ y •2 ²
3aj jà f1@xD2 â xz z A1 A1 f1@xD + 2 EI T1 s1 a j jà f1@xD f¢2 @xD â xz z A2 A1 f1@xD +
k 0 { k 0 {
i y
2aj jà f¢2@xD2 â xz z A1 A2 A2 f²1 @xD + 2 EI T1 s1-I T1 s2 F A2 f²2 @xD +
•
k 0 {
1

i y i 1 ¢ y
EI T1 s1 a j
jà f¢1 @xD2 â xz
z A2 A1 f²2 @xD + 4 a j
jà f1@xD f¢2 @xD â xz z A1 A2 A2 f²2 @xD,
1 •2 •
k 0 { k 0 {
-2 I HD1A2 L w2 f2 @xD - 2 I A2 w2 m@xD f2 @xD + 2 E-I T1 s1 +I T1 s2 F A1 f²1@xD +
i 1 ¢ y 3 ² i 1 ¢ y
E-I T1 s1 a j
jà f1 @xD2 â xzz A1 f1@xD + 4 a j jà f1 @xD f¢2@xD â xz z A1 A2 A1 f²1@xD +
•
k 0 { k 0 {
i y i y
2aj jà f¢1@xD2 â xz z A1 A2 A1 f²2 @xD + 3 a j
jà f¢2@xD2 â xz z A22 A2 f²2@xD>
• •
k 0 { k 0 {
1 1

Because the homogeneous part of order1Eq and order1BC has a nontrivial solution, the corresponding nonhomogeneous

lem is self-adjoint and that the solvability conditions demand that ST be orthogonal to f1 @xD and f2@xD, respectively. Impos-
problem has a solution only if solvability conditions are satisfied. It follows from Section 7.1.1 that the homogeneous prob-

ing these conditions, we have

206 Chapter7.nb

SCond =
Table@int@fj@xD ST@@jDD, 8x, 0, 1<D •. intRule1 ••. intRule2 •. int -> Integrate, 8j, 2<D ==
0 •• Thread;
SCond •. displayRule

i 1 y i 1 y i 1 y
:2 EI T1 s1 -I T1 s2 F j
jà f1@xD f²2 @xD â xz z A2 - 2 I j
jà f1 @xD2 â xz z HD1A1 L w1 - 2 I j
jà m@xD f1 @xD2 â xz z A1 w1 +
k 0 { k 0 { k 0 {
i y i y i y
2 EI T1 s2 F j jà f1 @xD f²1 @xD â xz z A1 + 3 a jjà f¢1 @xD2 â xz zjjà f1 @xD f²1@xD â xz z A21 A1 +
• •
k 0 { k 0 {k 0 {
1 1 1

i y i y
2 EI T1 s1 a j jà f¢1 @xD f¢2 @xD â xz zjjà f1@xD f²1 @xD â xz z A2 A1 +
•
k 0 {k 0 {
1 1 2

i
jà f1 @xD â xz y
zji
jà f1@xD f2 @xD â xz y
z A2 A1 + 2 a i jà f¢2@xD2 â xy z
aj j z
•
k 0 {k 0 { k 0 {
1 1 1
I T1 s1 ¢ 2 ² 2
E

i
jà f1@xD f²1 @xD â xy z A1 A2 • i
jà f¢1 @xD f¢2@xD â xy zijà f1@xD f²2 @xD â xy z A1 A2 •
j z A2 + 4 a j z j z
k 0 { k 0 {k 0 {
1 1 1
A2 == 0,

i 1 y i 1 ¢ yji 1 y 3
2 E-I T1 s1 +I T1 s2 F jjà f2@xD f²1 @xD â xz z A1 + E-I T1 s1 a j
jà f1@xD2 â xz z jà f2 @xD f²1@xD â xz z A1 -
k 0 { k 0 {k 0 {
i y i y
2Ij jà f2@xD2 â xz z HD1A2 L w2 - 2 I j
jà m@xD f2 @xD2 â xz z A2 w2 +
k 0 { k 0 {
1 1

i y i y i 1 ¢ y
4aj jà f¢1 @xD f¢2@xD â xz zjjà f2 @xD f²1 @xD â xz z A1 A2 A1 + 2 a j jà f1@xD2 â xz z
•
k 0 {k 0 { k 0 {
1 1

i
j y i y i y
jà f2@xD f²2 @xD â xz z A1 A2 A1 + 3 a jjà f¢2 @xD2 â xz zjjà f2@xD f²2 @xD â xz z A22 A2 == 0>
• •
k 0 { k 0 {k 0 {
1 1 1

To simplify the notation in the solvability conditions, we use the orthonormality of the mode shapes and define the following
parameters:

notationRule = 9à fi_ @xD2 â x -> 1, à m@xD fi_ @xD2 â x -> mi , F à fi_@xD f²j_@xD â x -> -wi si,j f=;
1 1 1

0 0 0

To identify the forms of the nonlinear terms in the solvability conditions, we first identify all of the possible forms of the
nonhomogeneous terms in the first-order problem. To this end, we let

basicTerms = TableA9Ai@T1 D EI wi T0 , Ai @T1 D E-I wi T0 =, 8i, 2<E •• Flatten

•

8EI T0 w1 A1@T1 D, E-I T0 w1 A1 @T1D, EI T0 w2 A2 @T1 D, E-I T0 w2 A2@T1 D<

• •

Then, all of the possible forms of the nonlinear terms in the first-order problem are given by

cubicTerms = Nest@Outer@Times, basicTerms, #D &, basicTerms, 2D •• Flatten •• Union;

cubicTerms •. displayRule

:E3 I T0 w1 A31, EI T0 H2w1 +w2L A21 A2 , EI T0 Hw1 +2 w2L A1 A22 , E3 I T0 w2 A32 , EI T0 w1 A21 A1 ,
•

EI T0 w2 A1 A2 A1, EI T0 H-w1+2 w2L A22 A1 , E-I T0 w1 A1 A1 , EI T0 H-2 w1+w2 L A2 A1, E-3 I T0 w1 A1,
• • •2 •2 •3

EI T0 H2 w1 -w2L A21 A2 , EI T0 w1 A1 A2 A2, EI T0 w2 A22 A2 , E-I T0 w2 A1 A1 A2 , E-I T0 w1 A2 A1 A2 ,

• • • • • • •

EI T0 H-2 w1 -w2L A1 A2 , EI T0 Hw1 -2 w2L A1 A2 , E-I T0 w2 A2 A2, EI T0 H-w1-2 w2L A1 A2 , E-3 I T0 w2 A2 >
•2 • •2 •2 • •2 •3
Chapter7.nb 207

Out of these terms, only the terms that may lead to secular terms appear in the solvability conditions, which can be identified
according to

secularTerms =
IE-I w# T0 cubicTerms •. expRule@#D •. Exp@_ T0 + _.D -> 0 •• Union •• RestM & •ž 81, 2<

::A1 @T1D2 A1@T1 D, EI T1 s1 A2 @T1 D A1 @T1 D , A1 @T1D A2 @T1D A2 @T1D>,

• • 2 •

8E-I T1 s1 A1@T1 D3, A1 @T1D A2 @T1D A1 @T1D, A2@T1 D2 A2 @T1D<>

• •

Next, we define the following parameters:

symbolList = 8-8 w1 8g11 , d1, g12 <, -8 w2 8d2 , g21, g22 <<

88-8 g11 w1 , -8 d1 w1 , -8 g12 w1<, 8-8 d2 w2, -8 g21 w2 , -8 g22 w2 <<

and express SCond in a more concise form as

eqMod = ExpandA- €€€€€€ HHSCond@@#, 1DD •. notationRule •. Thread@secularTerms@@#DD -> 0DL +

1
w#
symbolList@@#DD.secularTerms@@#DDLE == 0 & •ž 81, 2<

:2 I m1 A1@T1 D + 2 EI T1 s1 -I T1 s2 f s1,2 A2@T1 D + 2 EI T1 s2 f s1,1 A1@T1 D + 8 g11 A1@T1 D2 A1 @T1 D +

• •

8 EI T1 s1 d1 A2 @T1D A1 @T1D + 8 g12 A1@T1 D A2@T1 D A2@T1 D + 2 I A¢1 @T1D == 0,

• 2 •

2 E-I T1 s1 +I T1 s2 f s2,1 A1@T1 D + 8 E-I T1 s1 d2 A1 @T1D3 + 2 I m2 A2@T1 D +

8 g21 A1 @T1 D A2 @T1 D A1 @T1 D + 8 g22 A2 @T1 D2 A2 @T1D + 2 I A¢2 @T1 D == 0>
• •

where

-symbolList@@1DD -> H-Coefficient@SCond@@1, 1DD, #D & •ž secularTerms@@1DDL •• Thread ••

TableForm

8 g11 w1 ® -3 a IÙ01 f¢1 @xD2 â xM Ù01 f1@xD f²1 @xD â x

8 d1 w1 ® -2 a IÙ0 f¢1 @xD f¢2@xD â xM Ù0 f1@xD f²1 @xD â x - a IÙ0 f¢1 @xD2 â xM Ù0 f1@xD f²2 @xD â x
1 1 1 1

8 g12 w1 ® -2 a IÙ01 f¢2 @xD2 â xM Ù01 f1@xD f²1 @xD â x - 4 a IÙ01 f¢1 @xD f¢2 @xD â xM Ù01 f1@xD f²2 @xD â x

-symbolList@@2DD -> H-Coefficient@SCond@@2, 1DD, #D & •ž secularTerms@@2DDL •• Thread ••

TableForm

8 d2 w2 ® -a IÙ0 f¢1 @xD2 â xM Ù0 f2 @xD f²1 @xD â x

1 1

8 g21 w2 ® -4 a IÙ01 f¢1 @xD f¢2@xD â xM Ù01 f2@xD f²1 @xD â x - 2 a IÙ01 f¢1@xD2 â xM Ù01 f2@xD f²2 @xD â x

8 g22 w2 ® -3 a IÙ01 f¢2 @xD2 â xM Ù01 f2@xD f²2 @xD â x

Modulation Equations in Polar Form

The complex-valued solvability conditions can be expressed in real-valued form by introducing the polar transformation
208 Chapter7.nb

ruleA = 9Ai_ -> i

j €€€€ ai@#D Exp@I qi@#DD &y Ai_ -> i
z, • j €€€€ ai@#D Exp@-I qi @#DD &y
z=;
k2
1
{ k2
1
{

into eqMod and obtain

expr1 = Expand@eqMod@@#, 1DD Exp@-I q # @T1DD •. ruleAD & •ž 81, 2<

8I m1 a1@T1 D + EI T1 s2-2 I q1@T1 D f s1,1 a1@T1 D + g11 a1@T1 D3 + EI T1 s1 -IT1 s2-I q1 @T1D+I q2@T1 D f s1,2 a2@T1 D +
EI T1 s1 -3 I q1@T1 D+Iq2 @T1D d1 a1 @T1D2 a2 @T1 D + g12 a1 @T1D a2 @T1D2 + I a¢1 @T1D - a1@T1 D q¢1@T1 D,
E-I T1 s1 +I T1 s2 +I q1@T1 D-I q2 @T1D f s2,1 a1 @T1D + E-IT1 s1+3 I q1 @T1D-I q2@T1 D d2 a1@T1 D3 +
I m2 a2@T1 D + g21 a1@T1 D2 a2 @T1D + g22 a2 @T1D3 + I a¢2 @T1D - a2@T1 D q¢2@T1 D<

Next, we separate the real and imaginary parts of expr1 to obtain the equations governing the modulation of the amplitudes
a i and phases qi . To accomplish this, we define the rule

realRule = 8Re@s_D -> s, Im@s_D -> 0<;

Then, the equations governing the amplitudes of motion correspond to the imaginary parts of expr1; that is,

ampEq = Solve@ComplexExpand@Im@#DD == 0 •. realRule & •ž expr1, 8a¢1 @T1D, a¢2 @T1D<D@@1DD •.

Rule -> Equal •• ExpandAll

8a¢1 @T1D == -m1 a1 @T1 D - f Sin@T1 s2 - 2 q1 @T1DD s1,1 a1@T1 D -

f Sin@T1 s1 - T1 s2 - q1 @T1D + q2@T1 DD s1,2 a2 @T1D - Sin@T1 s1 - 3 q1 @T1D + q2@T1 DD d1 a1 @T1D2 a2@T1 D,
a¢2 @T1D == f Sin@T1 s1 - T1 s2 - q1 @T1D + q2@T1 DD s2,1 a1 @T1D +
Sin@T1 s1 - 3 q1@T1 D + q2 @T1DD d2 a1@T1 D3 - m2 a2 @T1D<

Moreover, the equations governing the phases of motion correspond to the real parts of expr1; that is,

phaseEq = Solve@ComplexExpand@Re@#DD == 0 •. realRule & •ž expr1, 8q1 ¢@T1 D, q2¢ @T1D<D@@1DD •.

Rule -> Equal •• ExpandAll

:q¢1 @T1D ==
f Cos@T1 s1 - T1 s2 - q1@T1 D + q2 @T1DD s1,2 a2@T1 D
f Cos@T1 s2 - 2 q1@T1 DD s1,1 + g11 a1@T1 D2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
a1 @T1D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ +

Cos@T1 s1 - 3 q1@T1 D + q2 @T1DD d1 a1@T1 D a2@T1 D + g12 a2@T1 D2,

f Cos@T1 s1 - T1 s2 - q1 @T1 D + q2 @T1DD s2,1 a1 @T1D
q¢2 @T1D == g21 a1@T1 D2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
a2@T1 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ +

Cos@T1 s1 - 3 q1 @T1D + q2@T1 DD d2 a1 @T1D3

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + g22 a2 @T1 D2>
a2@T1 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

These modulation equations are nonautonomous because they depend explicitly on T1.

To determine an autonomous set of modulation equations, we start with expr1 and identify the independent phase argu-
ments in it. To accomplish this, we first identify all possible phase arguments; that is,
Chapter7.nb 209

expTerms = I Cases@expr1, Exp@arg_D -> arg, InfinityD •• Expand

8-T1 s2 + 2 q1 @T1 D, -T1 s1 + T1 s2 + q1@T1 D - q2 @T1D,

-T1 s1 + 3 q1 @T1 D - q2 @T1D, T1 s1 - T1 s2 - q1 @T1 D + q2 @T1D, T1 s1 - 3 q1@T1 D + q2 @T1D<

Out of these arguments, only two are independent because

Outer@D, expTerms, 8q1 @T1D, q2 @T1D<D •• RowReduce

881, 0<, 80, 1<, 80, 0<, 80, 0<, 80, 0<<

Denoting these two independent arguments by g1 and g2 , we have

gammaList = 82 g1@T1 D, g2 @T1D< == expTerms@@81, 2<DD •• Thread

82 g1 @T1D == -T1 s2 + 2 q1 @T1D, g2@T1 D == -T1 s1 + T1 s2 + q1@T1 D - q2 @T1D<

which can be solved for q1 and q2 to obtain

thetaRule = Solve@gammaList, 8q1@T1 D, q2@T1 D<D@@1DD •• ExpandAll

9q1 @T1D ® €€€€€€€€

€€€€ + g1 @T1D, q2 @T1D ® -T1 s1 + €€€€€€€€1€€€€€2€€€ + g1 @T1D - g2 @T1 D=
T1 s2 3T s
2 2

Substituting for the qi in expr1 yields

expr2 = expr1 •. Table@qi -> HEvaluate@thetaRule@@i, 2DD •. T1 -> #D &L, 8i, 2<D •.
Exp@arg_D :> Exp@Expand@argDD •• Expand

9I m1 a1@T1 D - €€€€ s2 a1@T1 D + E-2 I g1 @T1D f s1,1 a1@T1 D + g11 a1@T1 D3 + E-I g2@T1 D f s1,2 a2@T1 D +
1

E-2 I g1@T1 D-I g2 @T1D d1 a1 @T1 D2 a2 @T1D + g12 a1 @T1D a2 @T1D2 + I a¢1 @T1D - a1@T1 D g1¢ @T1 D,
2

EI g2 @T1 D f s2,1 a1@T1 D + E2 I g1 @T1D+I g2@T1 D d2 a1@T1 D3 + I m2 a2 @T1D + s1 a2 @T1 D - €€€€ s2 a2@T1 D +
3

g21 a1@T1 D a2 @T1D + g22 a2 @T1D + I a2 @T1D - a2@T1 D g1@T1 D + a2 @T1 D g2 @T1 D=
2
2 3 ¢ ¢ ¢

Then, the equations governing the amplitudes of motion correspond to the imaginary parts of expr2; that is,

HampEq = Solve@ComplexExpand@Im@#DD == 0 •. realRule & •ž expr2, 8a¢1 @T1 D, a¢2 @T1D<D@@1DD •.

Rule -> Equal •• ExpandAllL •. f_@T1 D -> f

8a¢1 == Sin@2 g1 + g2D a21 a2 d1 - a1 m1 + f Sin@2 g1D a1 s1,1 + f Sin@g2 D a2 s1,2 ,

a¢2 == -Sin@2 g1 + g2 D a31 d2 - a2 m2 - f Sin@g2D a1 s2,1<

Moreover, the equations governing the phases of motion correspond to the real parts of expr2; that is,
210 Chapter7.nb

HphaseEq = Solve@ComplexExpand@Re@#DD == 0 •. realRule & •ž expr2, 8g1¢ @T1 D, g2¢ @T1D<D@@1DD •.

Rule -> Equal •• ExpandAllL •. f_@T1 D -> f

:g2¢ == a21 g11 + a22 g12 - a21 g21 - a22 g22 + Cos@2 g1 + g2D a1 a2 d1 -

Cos@2 g1 + g2D a31 d2 f Cos@g2 D a2 s1,2 f Cos@g2 D a1 s2,1

€€€€€€€€€€€€€ - s1 + s2 + f Cos@2 g1 D s1,1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

D
a2 a1 a2

g1¢ == a21 g11 + a22 g12 + Cos@2 g1 + g2D a1 a2 d1 - €€€€€€ + f Cos@2 g1 D s1,1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
s 2 f Cos@g 2 a2 s1,2
2 a1

The modulation equations are autonomous because they are independent of the independent variable T1 .

Modulation Equations in Cartesian Form and Symmetry Property

In performing bifurcation analyses of the dynamics of the system, one might find it more convenient, in some cases, to
represent the modulation equations in Cartesian rather than polar form. As a byproduct, one can also obtain the symmetries
of the system.To this end, we introduce the Cartesian transformation

cartRule =

9A k_ -> i
j €€€€ Hp k@#D - I q k@#DL Exp@I l k@#DD &y A k_ -> i
z, • j €€€€ Hp k @#D + I q k @#DL Exp@-I l k @#DD &y
z=;
k2
1
{ k2
1
{
realRule = 8Im@x_D -> 0, Re@x_D -> x, l k_¢ @T1D -> n k <;

where the li are chosen to render the resulting modulation equations autonomous. Substituting the cartRule into eqMod
yields

eqModCart@k_D := Exp@-I l k @T1DD Subtract žž eqMod@@kDD •. cartRule •• Expand;

Next, we choose the li to render eqModCart autonomous. To accomplish this, we identify its independent arguments by
identifying first all of its arguments. They are given by

list1 =
-I HCases@#, Exp@x_D -> x, InfinityD & •ž Array@eqModCart, 2D •• Flatten •• UnionL •• Expand

8T1 s2 - 2 l1@T1 D, -T1 s1 + T1 s2 + l1@T1 D - l2 @T1 D,

-T1 s1 + 3 l1 @T1 D - l2 @T1D, T1 s1 - T1 s2 - l1 @T1 D + l2 @T1D, T1 s1 - 3 l1@T1 D + l2 @T1D<

To determine the number of independent relations in list1, we use the function RowReduce and obtain

Outer@D, list1, 8 l1 @T1D, l2 @T1D<D •• RowReduce

881, 0<, 80, 1<, 80, 0<, 80, 0<, 80, 0<<

Hence, there are only two linearly independent arguments in list1. We choose the first two to determine the li. For an
invariant set of modulation equations, we set each of these relations to be an even multiple of p. Solving the resulting
relations for the li, we obtain

list1 = list1@@81, 2<DD;

Chapter7.nb 211

lambdaRule = Solve@list1 + 8m1 , m2< 2 p == 0 •• Thread, Table@ li @T1 D, 8i, 2<DD@@1DD

9l1 @T1D ® €€€€ H2 p m1 + T1 s2L, l2 @T1 D ® €€€€ H2 p m1 + 4 p m2 - 2 T1 s1 + 3 T1 s2 L=

1 1
2 2

where the mi are positive or negative integers. Letting ni = li ¢ yields

D@lambdaRule, T1 D •. realRule

9n1 ® €€€€€€ , n2 ® €€€€ H-2 s1 + 3 s2 L=

s2 1
2 2

Then, separating the real and imaginary parts of eqModCart, we obtain the following Cartesian form of the modulation
equations:

Heqs = Solve@
Flatten@Table@ComplexExpand@8Im@#D, Re@#D<D & •ž HeqModCart@kD •. Exp@_D -> 1L ••.
realRule, 8k, 2<DD == 0 •• Thread, Table@8p k¢ @T1D, q k ¢ @T1 D<, 8k, 2<D ••
FlattenD@@1DD •. Rule -> EqualL •. f_@T1 D -> f •• TableForm

p¢1 == p21 q1 g11 + q31 g11 + p22 q1 g12 + q1 q22 g12 - 2 p1 p2 q1 d1 + p21 q2 d1 - q21 q2 d1 - p1 m1 - q1 n1 - f q1 s1,1 + f q2
q¢1 == -p31 g11 - p1 q21 g11 - p1 p22 g12 - p1 q22 g12 - p21 p2 d1 + p2 q21 d1 - 2 p1 q1 q2 d1 - q1 m1 + p1 n1 - f p1 s1,1 - f p
p¢2 == p21 q2 g21 + q21 q2 g21 + p22 q2 g22 + q32 g22 + 3 p21 q1 d2 - q31 d2 - p2 m2 - q2 n2 + f q1 s2,1
q¢2 == -p21 p2 g21 - p2 q21 g21 - p32 g22 - p2 q22 g22 - p31 d2 + 3 p1 q21 d2 - q2 m2 + p2 n2 - f p1 s2,1

The symmetry property of the system can also be obtained as follows:

phase = Flatten@
Table@Mod@#@@2DD & •ž lambdaRule •. T1 -> 0, 2 pD, 8m1, 0, 10<, 8m2, 0, 10<D, 1D •• Union

880, 0<, 8p, p<<

rotMat@a_D = 88Cos@aD, Sin@aD<, 8-Sin@aD, Cos@aD<<;

symmetryList = Join žž Table@rotMat@#@@iDDD.8pi, qi <, 8i, 2<D & •ž phase

88p1 , q1, p2, q2 <, 8-p1, -q1 , -p2 , -q2<<

Hence, if 8 p 1, q 1 , p 2, q 2 < is a solution of the modulation equations, then 8- p 1, -q 1 , - p 2 , -q 2 < is also another solution of
these equations. If these two sets of solutions are the same, then the system response is symmetric; otherwise, it will be
asymmetric.
212 Chapter7.nb

à The Functions PolarForm and CartesianForm

Following the procedures described in the preceding section, we build two functions, PolarForm and CartesianForm, to
automate the process provided that the complex dependent variables in the modulation equations are written in the form
•••
S i , i = 1, ..., neq and the corresponding complex conjugates are S i. The functions can be slightly modified to allow a more
general form for the modulation equations. To obtain the polar form or Cartesian form of the complex modulation equations,
we simply provide the list of modulation equations and S as two required inputs in PolarForm or CartesianForm, respec-
tively. If S = A, the list of modulation equations is the only required input.

PolarForm@eqs_List, S_Symbol: AD :=
ModuleA8neq = Length@eqsD, t<,
t = Cases@eqs, Si_@a_D -> a, ¥D@@1DD;

polarRule = 9S k_ -> i y, •
j €€€€ a k @#D Exp@I b k @#DD &z i €€€€
S k_ -> j y=;
a k @#D Exp@-I b k@#DD &z
k2
1
{ k2
1
{
realRule = 8Im@x_D -> 0, Re@x_D -> x<;
eqModPolar@k_D := Exp@-I b k @tDD Subtract žž eqs@@kDD •. polarRule •• Expand;
list1 = I Cases@Array@eqModPolar, neqD, Exp@a_D -> a, ¥D •• Expand •• Union;
newLists@list_, n_D :=
With@8v = Table@Unique@D, 8n<D<,
ReplaceList@list,
Append@Flatten@8___, Pattern@#, _D< & •ž vD, ___D -> vDD;
betaList = Table@bi @tD, 8i, neq<D;
list2 = 8<;
Scan@
If@NullSpace@Outer@Coefficient, #, betaListDD === 8<, list2 = -#; Return@DD &,
newLists@list1, neqDD;
If@list2 === 8<, Print@
"Autonomous system is not possible!!!\nThe phase list is:"D; Return@list1D
D;
betaRule1 = Solve@list3 = Table@gi @tD, 8i, neq<D == list2 •• Thread, betaListD@@1DD;
betaRule2 =
Table@bi -> Function@t •• Evaluate, bi @tD •. betaRule1 •• EvaluateD, 8i, neq<D;
Do@eq0@iD = eqModPolar@iD •. betaRule2 •. Exp@a_D :> Exp@Expand@aDD;
eq@iD =
Solve@ComplexExpand@Im@eq0@iDDD == 0 •. realRule, ai ¢@tDD@@1, 1DD •. Rule -> Equal;
eqRe@iD = ComplexExpand@Re@eq0@iDDD == 0 •. realRule, 8i, neq<
D;
gammapList = Table@gi ¢ @tD, 8i, neq<D;
eqList = Solve@Array@eqRe, neqD, gammapListD@@1DD •. Rule -> Equal •• ExpandAll;
Do@eq@i + neqD = eqList@@iDD, 8i, neq<D;
8Array@eq, 2 * neqD, list3<
E
Chapter7.nb 213

CartesianForm@eqs_List, S_Symbol: AD :=
ModuleA8neq = Length@eqsD, t<,
t = Cases@eqs, Si_@a_D -> a, ¥D@@1DD;
cartRule = 9S k_ -> H1 • 2 Hp k @#D - I q k @#DL Exp@I l k @#DD &L,
S k_ -> H1 • 2 Hp k@#D + I q k@#DL Exp@-I l k @#DD &L=;
•

realRule = 8Im@x_D -> 0, Re@x_D -> x, l k_ ¢ @tD -> n k <;

eqModCart@k_D := Exp@-I l k @tDD Subtract žž eqs@@kDD •. cartRule •• Expand;
eqs1 = 8ComplexExpand@Im@#DD == 0, ComplexExpand@Re@#DD == 0< & •ž
HArray@eqModCart, neqD •. E^ _ -> 1L ••. realRule •• Flatten;
eqs2 = Solve@eqs1, Table@8pk ¢ @tD, q k¢ @tD<, 8k, neq<D •• FlattenD@@1DD •.
Rule -> Equal;
list1 = I Cases@Array@eqModCart, neqD, Exp@a_D -> a, ¥D •• Expand •• Union;
newLists@list_, n_D :=
With@8v = Table@Unique@D, 8n<D<,
ReplaceList@list,
Append@Flatten@8___, Pattern@#, _D< & •ž vD, ___D -> vDD;
lambdaList = Table@li @tD, 8i, neq<D;
list2 = 8<;
Scan@
If@NullSpace@Outer@Coefficient, #, lambdaListDD === 8<, list2 = -#; Return@DD &,
newLists@list1, neqDD;
If@list2 === 8<, Print@
"Autonomous system is not possible!!!\nThe phase list is:"D; Return@list1D
D;
lambdaRule = Solve@list2 + 2 Table@mi, 8i, neq<D p == 0 •• Thread, lambdaListD@@1DD;
iterList = Table@8mi , 0, 10<, 8i, neq<D;
phase =
Flatten@Table@Mod@Select@#, ! FreeQ@#, pD &D & •ž HlambdaList •. lambdaRule •• ExpandL,
2 pD •• Evaluate, Sequence žž iterList •• EvaluateD, neq - 1D •• Union;
rotMat@q_D := 88Cos@qD, -Sin@qD<, 8Sin@qD, Cos@qD<<;
symmetryList = Flatten@Table@rotMat@#@@kDDD.8p k, q k <, 8k, 2<DD & •ž phase;
8eqs2, D@lambdaRule, tD •. realRule, symmetryList<
E

Ÿ Principal Parametric Resonance of the Second Mode

In this case, W » w2 . To describe the nearness of the internal and principal parametric resonances, we introduce the two
detuning parameters s1 and s2 defined by

ResonanceConds = 8w2 == 3 w1 + e s1 , W == 2 w2 + e s2 <;

and define the following rules:

214 Chapter7.nb

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

98w2 ® e s1 + 3 w1, W ® 2 e s1 + e s2 + 6 w1 <, 9w1 ® €€€€ H-e s1 + w2 L, W ® e s2 + 2 w2==

1
3

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand side of order1Eq and obtain the source of secular terms as

ST = Table@Coefficient@order1Eq@@2DD •. expRule@iD, Exp@I wi T0DD, 8i, 2<D;

ST •. displayRule

i 1 ¢ y 2• ²
:-2 I HD1A1 L w1 f1 @xD - 2 I A1 w1 m@xD f1 @xD + 3 a j jà f1 @xD2 â xz z A1 A1 f1@xD +
k 0 {
i y i y
2 EI T1 s1 a j
jà f¢1 @xD f¢2 @xD â xz z A2 A1 f²1 @xD + 2 a j
jà f¢2 @xD2 â xz z A1 A2 A2 f²1 @xD +
k 0 { k 0 {
1 • 2
1 •

i y i y
EI T1 s1 a j
jà f¢1 @xD2 â xz z A2 A1 f²2 @xD + 4 a jjà f¢1@xD f¢2 @xD â xz z A1 A2 A2 f²2 @xD,
• •
k 0 { k 0 {
1 2 1

i
jà f1@xD2 â xy z A31 f²1@xD +
-2 I HD1A2 L w2 f2 @xD - 2 I A2 w2 m@xD f2 @xD + E aj z
k 0 {
1
-I T1 s1 ¢

i
jà f¢1@xD f¢2 @xD â xy z A1 A2 • i
jà f¢1 @xD2 â xy z A1 A2 •
4aj z A1 f²1 @xD + 2 a j z A1 f²2 @xD +
k 0 { k 0 {
1 1

i 1 ¢ y 2• ²
2 EI T1 s2 F A2 f²2 @xD + 3 a j
jà f2 @xD2 â xz z A2 A2 f2@xD>
k 0 {
•

order1BC demand that ST be orthogonal to f1 @xD and f2@xD, respectively. Imposing these conditions, we have
As discussed in the case of principal parametric resonance of the first mode, the solvability conditions of order1Eq and
Chapter7.nb 215

SCond =
Table@int@fj@xD ST@@jDD, 8x, 0, 1<D •. intRule1 ••. intRule2 •. int -> Integrate, 8j, 2<D ==
0 •• Thread;
SCond •. displayRule

i 1 y
:-2 I jjà f1 @xD2 â xz z HD1A1 L w1 -
k 0 {
i y i 1 ¢ yi y 2•
2Ij jà m@xD f1 @xD2 â xz z A1 w1 + 3 a jjà f1@xD2 â xz zjjà f1 @xD f²1@xD â xz z A1 A1 +
k 0 { k 0 {k 0 {
1 1

i y i y i y
2 EI T1 s1 a j
jà f¢1 @xD f¢2 @xD â xz zjjà f1@xD f²1 @xD â xz z A2 A1 + EI T1 s1 a j
jà f¢1 @xD2 â xzz
•
k 0 {k 0 { k 0 {
1 1 2 1

i
jà f1@xD f2 @xD â xz y
z A2 A1 + 2 a ji
jà f2@xD â xz y
zji
jà f1 @xD f1@xD â xz y
z A1 A2 A2 +
j
• •
k 0 { k 0 {k 0 {
1 1 1
² 2 ¢ 2 ²

i
jà f¢1 @xD f¢2@xD â xy zijà f1 @xD f²2 @xD â xy z A1 A2 •
4aj z j z
k 0 {k 0 {
1 1
A2 == 0,

i 1 ¢ yji 1 y 3 i 1 y
E-I T1 s1 a j
jà f1 @xD2 â xz z jà f2@xD f²1 @xD â xz z A1 - 2 I j
jà f2 @xD2 â xz z HD1A2 L w2 -
k 0 {k 0 { k 0 {
i y i y i y
2Ij jà m@xD f2 @xD2 â xz z A2 w2 + 4 a jjà f¢1@xD f¢2 @xD â xz zjjà f2 @xD f²1@xD â xz z A1 A2 A1 +
•
k 0 { k 0 {k 0 {
1 1 1

i y i y i y•
2aj jà f¢1 @xD2 â xzzjjà f2 @xD f²2@xD â xz z A1 A2 A1 + 2 EI T1 s2 F j
jà f2 @xD f²2@xD â xz z A2 +
•
k 0 {k 0 { k 0 {
1 1 1

i y i y
3aj jà f¢2 @xD2 â xzzjjà f2 @xD f²2@xD â xz z A22 A2 == 0>
•
k 0 {k 0 {
1 1

Using the notations introduced in the preceding section, we can express SCond in a more concise form as

eqMod = ExpandA- €€€€€€ HHSCond@@#, 1DD •. notationRule •. Thread@secularTerms@@#DD -> 0DL +

1
w#
symbolList@@#DD.secularTerms@@#DDLE == 0 & •ž 81, 2<

:2 I m1 A1@T1 D + 8 g11 A1 @T1 D2 A1 @T1D +

•

8 EI T1 s1 d1 A2 @T1D A1 @T1D + 8 g12 A1@T1 D A2@T1 D A2@T1 D + 2 I A¢1 @T1D == 0,

• 2 •

8 E-I T1 s1 d2 A1 @T1D3 + 2 I m2 A2@T1 D + 8 g21 A1 @T1 D A2 @T1 D A1 @T1 D +

•

2 EI T1 s2 f s2,2 A2@T1 D + 8 g22 A2@T1 D2 A2 @T1D + 2 I A¢2@T1 D == 0>

• •

where

-symbolList@@1DD -> H-Coefficient@SCond@@1, 1DD, #D & •ž secularTerms@@1DDL •• Thread ••

TableForm

8 g11 w1 ® -3 a IÙ0 f¢1 @xD2 â xM Ù0 f1@xD f²1 @xD â x

1 1

8 d1 w1 ® -2 a IÙ01 f¢1 @xD f¢2@xD â xM Ù01 f1@xD f²1 @xD â x - a IÙ01 f¢1 @xD2 â xM Ù01 f1@xD f²2 @xD â x

8 g12 w1 ® -2 a IÙ0 f¢2 @xD2 â xM Ù0 f1@xD f²1 @xD â x - 4 a IÙ0 f¢1 @xD f¢2 @xD â xM Ù0 f1@xD f²2 @xD â x
1 1 1 1
216 Chapter7.nb

-symbolList@@2DD -> H-Coefficient@SCond@@2, 1DD, #D & •ž secularTerms@@2DDL •• Thread ••

TableForm

8 d2 w2 ® -a IÙ0 f¢1 @xD2 â xM Ù0 f2 @xD f²1 @xD â x

1 1

8 g21 w2 ® -4 a IÙ0 f¢1 @xD f¢2@xD â xM Ù0 f2@xD f²1 @xD â x - 2 a IÙ0 f¢1@xD2 â xM Ù0 f2@xD f²2 @xD â x
1 1 1 1

8 g22 w2 ® -3 a IÙ01 f¢2 @xD2 â xM Ù01 f2@xD f²2 @xD â x

Modulation Equations in Polar Form

Using eqMod and the function PolarForm defined in the preceding section, we obtain the modulation equations in polar form
and the definitions for gi as

PolarForm@eqModD

::a¢1 @T1D == -m1 a1 @T1 D - Sin@g1@T1 DD d1 a1 @T1D2 a2 @T1 D,

a¢2 @T1D == Sin@g1 @T1DD d2 a1@T1 D3 - m2 a2 @T1D - f Sin@g2@T1 DD s2,2 a2 @T1D,

g1¢ @T1D == s1 + f Cos@g2 @T1DD s2,2 - 3 g11 a1 @T1D2 + g21 a1 @T1D2 +
Cos@g1@T1 DD d2 a1 @T1D3
€€€€€€€€€€€€€€€€€€€€€€ - 3 Cos@g1@T1 DD d1 a1 @T1D a2 @T1D - 3 g12 a2 @T1D2 + g22 a2@T1 D2,
a2@T1 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

2 Cos@g1@T1 DD d2 a1 @T1D3
g2¢ @T1D == s2 - 2 f Cos@g2 @T1DD s2,2 - 2 g21 a1 @T1D2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 g22 a2@T1 D2>,
a2 @T1D
8g1 @T1D == T1 s1 - 3 b1@T1 D + b2 @T1D, g2@T1 D == T1 s2 - 2 b2 @T1D<>

Modulation Equations in Cartesian Form and Symmetry Property

Using eqMod and the function CartesianForm defined in the preceding section, we obtain the modulation equations in
Cartesian form, the definitions for ni, and the symmetry property as
Chapter7.nb 217

CartesianForm@eqModD

:8p¢1 @T1D == -m1 p1 @T1 D - n1 q1@T1 D + g11 p1@T1 D2 q1 @T1D - 2 d1 p1 @T1D p2 @T1D q1 @T1D +

g12 p2@T1 D2 q1 @T1D + g11 q1 @T1D3 + d1 p1@T1 D2 q2 @T1D - d1 q1 @T1 D2 q2 @T1D + g12 q1 @T1D q2 @T1D2 ,
q¢1 @T1D == n1 p1@T1 D - g11 p1@T1 D3 - d1 p1 @T1 D2 p2 @T1D - g12 p1 @T1D p2 @T1D2 - m1 q1@T1 D -
g11 p1@T1 D q1@T1 D2 + d1 p2 @T1D q1 @T1D2 - 2 d1 p1@T1 D q1@T1 D q2@T1 D - g12 p1@T1 D q2@T1 D2,
p¢2 @T1D == -m2 p2 @T1 D + 3 d2 p1 @T1 D2 q1 @T1D - d2 q1 @T1D3 - n2 q2@T1 D - f s2,2 q2@T1 D +
g21 p1@T1 D2 q2 @T1D + g22 p2 @T1D2 q2@T1 D + g21 q1@T1 D2 q2 @T1D + g22 q2 @T1D3 ,
q¢2 @T1D == -d2 p1 @T1 D3 + n2 p2 @T1D - f s2,2 p2 @T1 D - g21 p1 @T1D2 p2@T1 D - g22 p2@T1 D3 +
3 d2 p1@T1 D q1@T1 D2 - g21 p2 @T1D q1 @T1D2 - m2 q2@T1 D - g22 p2@T1 D q2@T1 D2<,
• !!! • !!!
9n1 ® €€€€ H2 s1 + s2L, n2 ® €€€€€€ =, :8p1, q1, p2 , q2 <, : €€€€1€€ - €€€€€€€€€€€€€€€1€ , €€€€€€€€€€€€€1€€€ + €€€€€€ , -p2, -q2 >,
1 s2 p 3 q 3 p q1

• !!! • !!!
6 2 2 2 2 2

:- €€€€€€ - €€€€€€€€€€€€€€€€ , €€€€€€€€€€€€€€€€ - €€€€€€ , p2 , q2 >, 8-p1 , -q1, -p2 , -q2<,

p1 3 q1 3 p1 q1

• !!! •!!
2 2 2 2
1 •!! 1 • !!!
:- €€€€€€ + €€€€€€€€€€€€€€€€ , - €€€€ 3 p1 - €€€€€€ , p2 , q2>, : €€€€€€ + €€€€€€€€€€€€€€€€ , - €€€€ 3 p1 + €€€€€€ , -p2 , -q2>>>
p1 3 q1 q1 p1 3 q1 q1
2 2 2 2 2 2 2 2

Ÿ Combination Parametric Resonance of the Two Modes

In this case, W » w1 + w2. To describe the nearness of the internal and combination parametric resonances, we introduce the
two detuning parameters s1 and s2 defined by

ResonanceConds = 8w2 == 3 w1 + e s1 , W == w1 + w2 + e s2 <;

and define the following rules:

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

98w2 ® e s1 + 3 w1, W ® e s1 + e s2 + 4 w1 <, 9w1 ® €€€€ H-e s1 + w2 L, W ® €€€€ H-e s1 + 3 e s2 + 4 w2L==

1 1
3 3

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand side of order1Eq and obtain the source of secular terms as
218 Chapter7.nb

ST = Table@Coefficient@order1Eq@@2DD •. expRule@iD, Exp@I wi T0DD, 8i, 2<D;

ST •. displayRule

i 1 ¢ y 2• ²
:-2 I HD1A1 L w1 f1 @xD - 2 I A1 w1 m@xD f1 @xD + 3 a j jà f1 @xD2 â xz z A1 A1 f1@xD +
k 0 {
i
jà f¢ @xD f¢ @xD â xy z A2 • i
jà f¢ @xD2 â xy z
2 EI T1 s1 a j z A1 f²1 @xD + 2 a j z A1 A2 A2 f²1 @xD +
•
k 0 { k 0 {
1 2 1
1 2 2

i
jà f¢1 @xD2 â xy z A2 • i
jà f¢1@xD f¢2 @xD â xy z
EI T1 s1 a j z A1 f²2 @xD + 2 EI T1 s2 F A2 f²2 @xD + 4 a j z A1 A2 A2 f²2 @xD,
• •
k 0 { k 0 {
1 2 1

i y 3 ²
-2 I HD1A2 L w2 f2 @xD - 2 I A2 w2 m@xD f2 @xD + E-I T1 s1 a j jà f¢1@xD2 â xz z A1 f1@xD +
k 0 {
1

i y
2 EI T1 s2 F A1 f²1 @xD + 4 a j
jà f¢1 @xD f¢2@xD â xz z A1 A2 A1 f²1 @xD +
• •
k 0 {
1

i y i y 2• ²
2aj jà f¢1@xD2 â xz z A1 A2 A1 f²2 @xD + 3 a jjà f¢2@xD2 â xz z A2 A2 f2@xD>
•
k 0 { k 0 {
1 1

The solvability conditions demand that ST be orthogonal to solutions of the adjoint homogeneous problem. Since the
problem is self-adjoint, we have

SCond =
Table@int@fj@xD ST@@jDD, 8x, 0, 1<D •. intRule1 ••. intRule2 •. int -> Integrate, 8j, 2<D ==
0 •• Thread;
SCond •. displayRule

i 1 y i 1 y
:-2 I jjà f1 @xD2 â xz z HD1A1 L w1 - 2 I j jà m@xD f1@xD2 â xz z A1 w1 +
k 0 { k 0 {
i yji y 2 i 1 ¢ y
3aj jà f¢1 @xD2 â xzz jà f1 @xD f²1@xD â xz z A1 A1 + 2 EI T1 s1 a jjà f1@xD f¢2 @xD â xz z
•
k 0 {k 0 { k 0 {
1 1

i
j y i yji y
jà f1@xD f²1 @xD â xz z A2 A1 + EI T1 s1 a j
jà f¢1 @xD2 â xz z jà f1 @xD f²2 @xD â xz z A2 A1 +
• •2
k 0 { k 0 {k 0 {
1 2 1 1

i y i y i y
2 EI T1 s2 F j
jà f1 @xD f²2 @xD â xz z A2 + 2 a j
jà f¢2 @xD2 â xz zjjà f1 @xD f²1@xD â xz z A1 A2 A2 +
• •
k 0 { k 0 {k 0 {
1 1 1

i y i y
4aj jà f¢1 @xD f¢2@xD â xz zjjà f1 @xD f²2 @xD â xz z A1 A2 A2 == 0,
•
k 0 {k 0 {
1 1

i
jà f¢1 @xD2 â xzy
zji
jà f2@xD f²1 @xD â xz y
z A31 - 2 I i
jà f2 @xD2 â xy z HD1A2 L w2 -
E-I T1 s1 a j j z
k 0 {k 0 { k 0 {
1 1 1

i
jà m@xD f2 @xD â xz y
z A2 w2 + 2 E 1 2 F j i
jà f2 @xD f1 @xD â xz y
z A1 +
2Ij
•
k 0 { k 0 {
1 1
2 I T s ²

i
jà f¢1 @xD f¢2@xD â xy zijà f2 @xD f²1 @xD â xy z A1 A2 A1 + 2 a i jà f¢1@xD2 â xy z
4aj z j z j z
•
k 0 {k 0 { k 0 {
1 1 1

i
j y i yji y 2•
jà f2@xD f²2 @xD â xz z A1 A2 A1 + 3 a jjà f¢2 @xD2 â xz z jà f2@xD f²2 @xD â xz z A2 A2 == 0>
•
k 0 { k 0 {k 0 {
1 1 1

Using the notations introduced in the preceding section, we can express SCond in a more concise form as
Chapter7.nb 219

eqMod = ExpandA- €€€€€€ HHSCond@@#, 1DD •. notationRule •. Thread@secularTerms@@#DD -> 0DL +

1
w#
symbolList@@#DD.secularTerms@@#DDLE == 0 & •ž 81, 2<

:2 I m1 A1@T1 D + 8 g11 A1 @T1 D2 A1 @T1D + 8 EIT1 s1 d1 A2@T1 D A1@T1 D +

• • 2

2 EI T1 s2 f s1,2 A2@T1 D + 8 g12 A1@T1 D A2@T1 D A2@T1 D + 2 I A¢1 @T1D == 0,

• •

8 E-I T1 s1 d2 A1 @T1D3 + 2 I m2 A2@T1 D + 2 EI T1 s2 f s2,1 A1@T1 D +

•

8 g21 A1 @T1 D A2 @T1 D A1 @T1 D + 8 g22 A2 @T1 D2 A2 @T1D + 2 I A¢2 @T1 D == 0>
• •

where

-symbolList@@1DD -> H-Coefficient@SCond@@1, 1DD, #D & •ž secularTerms@@1DDL •• Thread ••

TableForm

8 g11 w1 ® -3 a IÙ0 f¢1 @xD2 â xM Ù0 f1@xD f²1 @xD â x

1 1

8 d1 w1 ® -2 a IÙ0 f¢1 @xD f¢2@xD â xM Ù0 f1@xD f²1 @xD â x - a IÙ0 f¢1 @xD2 â xM Ù0 f1@xD f²2 @xD â x
1 1 1 1

8 g12 w1 ® -2 a IÙ01 f¢2 @xD2 â xM Ù01 f1@xD f²1 @xD â x - 4 a IÙ01 f¢1 @xD f¢2 @xD â xM Ù01 f1@xD f²2 @xD â x

-symbolList@@2DD -> H-Coefficient@SCond@@2, 1DD, #D & •ž secularTerms@@2DDL •• Thread ••

TableForm

8 d2 w2 ® -a IÙ0 f¢1 @xD2 â xM Ù0 f2 @xD f²1 @xD â x

1 1

8 g21 w2 ® -4 a IÙ01 f¢1 @xD f¢2@xD â xM Ù01 f2@xD f²1 @xD â x - 2 a IÙ01 f¢1@xD2 â xM Ù01 f2@xD f²2 @xD â x

8 g22 w2 ® -3 a IÙ0 f¢2 @xD2 â xM Ù0 f2@xD f²2 @xD â x

1 1

Modulation Equations in Polar Form

Using eqMod and the function PolarForm, we obtain the modulation equations in polar form and the definitions for gi as

PolarForm@eqModD

::a¢1 @T1D == -m1 a1 @T1 D - f Sin@g2@T1 DD s1,2 a2 @T1D - Sin@g1@T1 DD d1 a1 @T1D2 a2@T1 D,
a¢2 @T1D == -f Sin@g2@T1 DD s2,1 a1 @T1D + Sin@g1@T1 DD d2 a1 @T1D3 - m2 a2@T1 D,
f Cos@g2@T1 DD s2,1 a1 @T1D Cos@g1@T1 DD d2 a1 @T1D3
g1¢ @T1D == s1 - 3 g11 a1@T1 D2 + g21 a1 @T1D2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
a2 @T1D a2 @T1 D
€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ -

3 f Cos@g2 @T1 DD s1,2 a2@T1 D

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 3 Cos@g1@T1 DD d1 a1 @T1D a2 @T1D - 3 g12 a2 @T1D2 + g22 a2@T1 D2,
a1@T1 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

f Cos@g2 @T1 DD s2,1 a1 @T1D Cos@g1 @T1 DD d2 a1 @T1 D3

g2¢ @T1D == s2 - g11 a1@T1 D2 - g21 a1 @T1D2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
a2 @T1D a2 @T1D
€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ -

f Cos@g2@T1 DD s1,2 a2@T1 D

€€€€€€€€€€€€€€€€€€€€€€€€€€€ - Cos@g1 @T1 DD d1 a1 @T1 D a2 @T1 D - g12 a2 @T1 D2 - g22 a2 @T1D2 >,
a1@T1 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

8g1 @T1D == T1 s1 - 3 b1@T1 D + b2 @T1D, g2@T1 D == T1 s2 - b1 @T1D - b2@T1 D<>

220 Chapter7.nb

Modulation Equations in Cartesian Form and Symmetry Property

Using eqMod and the function CartesianForm, we obtain the modulation equations in Cartesian form, the definitions for ni ,
and the symmetry property as

CartesianForm@eqModD

98p¢1 @T1D == -m1 p1 @T1 D - n1 q1@T1 D + g11 p1@T1 D2 q1 @T1D - 2 d1 p1 @T1D p2 @T1D q1 @T1D + g12 p2 @T1D2 q1@T1 D +
g11 q1@T1 D3 - f s1,2 q2 @T1 D + d1 p1 @T1D2 q2@T1 D - d1 q1@T1 D2 q2 @T1D + g12 q1 @T1D q2 @T1D2 ,
q¢1 @T1D == n1 p1@T1 D - g11 p1@T1 D3 - f s1,2 p2 @T1D - d1 p1 @T1D2 p2@T1 D - g12 p1@T1 D p2@T1 D2 -
m1 q1@T1 D - g11 p1@T1 D q1@T1 D2 + d1 p2 @T1D q1 @T1D2 - 2 d1 p1@T1 D q1@T1 D q2@T1 D - g12 p1 @T1 D q2 @T1 D2 ,
p2 @T1D == -m2 p2 @T1 D - f s2,1 q1 @T1D + 3 d2 p1 @T1D2 q1@T1 D - d2 q1 @T1D3 - n2 q2@T1 D +
¢

g21 p1@T1 D2 q2 @T1D + g22 p2 @T1D2 q2@T1 D + g21 q1@T1 D2 q2 @T1D + g22 q2 @T1D3 ,
q¢2 @T1D == -f s2,1 p1 @T1D - d2 p1 @T1D3 + n2 p2 @T1 D - g21 p1 @T1D2 p2@T1 D - g22 p2@T1 D3 +
3 d2 p1@T1 D q1@T1 D2 - g21 p2 @T1D q1 @T1D2 - m2 q2@T1 D - g22 p2@T1 D q2@T1 D2<,

9n1 ® €€€€ Hs1 + s2L, n2 ® €€€€ H-s1 + 3 s2 L=, 88p1, q1 , p2 , q2<, 8-q1 , p1, q2 , -p2<,
1 1

8-p1, -q1 , -p2, -q2 <, 8q1, -p1 , -q2, p2 <<=

4 4

Ÿ 7.2.3 Discretization of the Continuous Problem

discretized form, we expand w@x, tD in terms of the linear mode shapes fm @xD as
As an alternative, we apply the method of multiple scales to the discretized system of EOM and BC. To determine the

wRule1 = w -> FunctionA8x, t<, â u m @tD f m @xDE;

¥

m=1

To simplify the computation, we temporarily drop the Ú sign in wRule1, as long as we know the repeated m represents a
summation index, and rewrite it as

wRule2 = w -> Function@8x, t<, u m @tD f m @xDD;

where the u m @tD are the generalized coordinates. Substituting wRule2 into EOM, multiplying the result with fn @xD, and
integrating the outcome from x = 0 to x = 1 using intRule1 and intRule2 , we obtain the following discretized form of the
equations describing the response of the beam:

eq723a = int@f n @xD H# •. Integrate -> int •. wRule2L, 8x, 0, 1<D •. intRule1 ••. intRule2 •.
int -> Integrate & •ž EOM

i 1 y i 1 y i y
-2 P j
jà fn @xD f²m@xD â xz
z u m @tD + j
jà fn@xD fH4L z j z ²
m @xD â xz u m@tD + jà f m @xD fn@xD â xz u m @tD ==
k 0 { k 0 { k 0 {
1

i
jà fn @xD f² @xD â xz y
z u m @tD +
4 F e Cos@t WD j
k 0 {
1
m

i
jà fm @xD â xz y
zji
jà fn @xD f m@xD â xy z u m @tD3 - 2 e i
jà m@xD f m @xD fn@xD â xy
z u¢m @tD
aej z j z
k 0 {k 0 { k 0 {
1 1 1
¢ 2 ²
Chapter7.nb 221

where the damping is assumed to be modal. The nonlinear term, in general, should be rewritten as

NT = ICoefficientAeq723a@@2DD, u m @tD3 E •. f¢m @xD2 -> f k¢ @xD f{¢ @xDM u m @tD u k @tD u{@tD

i 1 ¢ yi y
aej
jà fk @xD f¢{ @xD â xz
zjjà fn @xD f²m@xD â xz
z uk @tD u m@tD u{ @tD
k 0 {k 0 {
1

Using the orthonormality of the mode shapes fi @xD and the results from the corresponding eigenvalue problem, we define the
following rules:

notationRule1 =

9à f m @xD f n @xD âx -> d m,n, à f n @xD fH4L

m @xD âx -> 2 P à f n @xD f²m @xD â x + w2m d m,n=;
1 1 1

0 0 0

notationRule2 = 9à m@xD f m @xD f n@xD â x -> m m d m,n, à f n @xD f²m @xD âx -> -g n,m ,
1 1

i yi y
0 0

j
jà f¢k@xD f¢{ @xD â xz
zjjà f n @xD f²m @xD âxz
z -> -g k,{ g n,m , a_ d m,n :> Ha •. m -> nL=;
1 1

k 0 {k 0 {

and then rewrite eq723a as

EOM1 =
Expand@eq723a@@1DD •. notationRule1D == Ieq723a@@2DD •. u m @tD3 -> 0M + NT ••. notationRule2

w2n un@tD + u²n@tD == -4 F e Cos@t WD gn,m u m@tD - a e gk,{ gn,m uk @tD u m@tD u{ @tD - 2 e mn u¢n@tD

where m, k, and { represent summation indicies.

Using the method of multiple scales, we seek a first-order uniform expansion in the form

i1 j
j y
z
solRule = u n_ -> j
j z;
jâ e u n,j@#1, #2D &z
z
kj=0 {

Transforming the total time derivatives in EOM1 into partial derivatives in terms of T0 and T1 , substituting the solRule into
EOM1, expanding the result for small e, and discarding terms of order higher than e, we obtain

eq723b = HEOM1 •. 8us_@tD -> us @T0, T1D, Derivative@n_D@us_D@tD -> dt@nD@us @T0 , T1 DD, t -> T0< •.
solRule •• ExpandAllL •. e n_•;n>1 -> 0

w2n un,0@T0 , T1D + e w2n un,1 @T0, T1 D + 2 e un,0 @T0, T1 D + un,0 @T0 , T1 D + e un,1 @T0 , T1 D ==
H1,1L H2,0L H2,0L

-4 F e Cos@T0 WD gn,m u m,0 @T0, T1 D -

a e gk,{ gn,m uk,0@T0 , T1 D u m,0 @T0, T1 D u{,0@T0 , T1 D - 2 e mn un,0 @T0, T1 D
H1,0L

Equating coefficients of like powers of e in eq723b, we obtain

222 Chapter7.nb

eqEps = CoefficientList@#, eD & •ž eq723b •• Thread

9w2n un,0@T0 , T1D + un,0 @T0 , T1D == 0,

H2,0L

w2n un,1@T0 , T1D + 2 un,0 @T0 , T1D + un,1 @T0 , T1D == -4 F Cos@T0 WD gn,m u m,0 @T0, T1 D -
H1,1L H2,0L

a gk,{ gn,m uk,0@T0 , T1 D u m,0 @T0, T1D u{,0@T0 , T1 D - 2 mn uH1,0L

n,0 @T0, T1D=

Because in the presence of damping, all modes that are not directly or indirectly excited decay with time, all of the u n,0 decay
with time except u 1,0 and u 2,0 . Then it follows from eqEps[[1]] that

sol0 = ui_,0 -> FunctionA8T0 , T1<, Ai@T1 D Exp@I wi T0D + Ai@T1 D Exp@-I wi T0 DE;
•

Taking account of the summation indicies, we define

sumRule = 9g k,{ g n,m u k,0@a__D u m,0@a__D u{,0 @a__D -> â â â g k,{ g n,m u k,0 @aD u m,0 @aD u{,0@aD,
2 2 2

k=1 m=1 {=1

g n,m um,0 @a__D -> â g n,m u m,0 @aD=;

2

m=1

Hence, we rewrite eqEps[[2]] as

Clear@order1EqD

order1Eq@n_D = HeqEps@@1, 1DD •. u n,0 -> u n,1 L ==

HeqEps@@1, 1DD •. u n,0 -> u n,1 L - HSubtract žž eqEps@@2DDL •. sumRule

w2n un,1@T0 , T1D + un,1 @T0 , T1D == -4 F Cos@T0 WD Hgn,1 u1,0 @T0, T1D + gn,2 u2,0@T0 , T1 DL -
H2,0L

a Hg1,1 gn,1 u1,0@T0 , T1D3 + g1,2 gn,1 u1,0 @T0, T1 D2 u2,0 @T0, T1 D +
g2,1 gn,1 u1,0@T0 , T1D2 u2,0@T0 , T1 D + g1,1 gn,2 u1,0@T0 , T1D2 u2,0@T0 , T1 D +
g2,2 gn,1 u1,0@T0 , T1D u2,0 @T0, T1 D2 + g1,2 gn,2 u1,0@T0 , T1D u2,0 @T0, T1 D2 +
g2,1 gn,2 u1,0@T0 , T1D u2,0 @T0, T1 D2 + g2,2 gn,2 u2,0@T0 , T1D3 L - 2 mn un,0 @T0 , T1D - 2 un,0 @T0 , T1D
H1,0L H1,1L

Substituting sol0 into order1Eq[1] and order1Eq[2] yields

eqOrder1 = Array@order1Eq, 2D •. sol0 •• TrigToExp •• ExpandAll;

à Principal Parametric Resonance of the First Mode

ResonanceConds = 8w2 == 3 w1 + e s1 , W == 2 w1 + e s2 <;

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

98w2 ® e s1 + 3 w1, W ® e s2 + 2 w1 <, 9w1 ® €€€€ H-e s1 + w2 L, W ® €€€€ H-2 e s1 + 3 e s2 + 2 w2L==

1 1
3 3

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand sides of eqOrder1 and obtain the solvability condition as
Chapter7.nb 223

SCond = Table@Coefficient@-eqOrder1@@i, 2DD •. expRule@iD, Exp@I wi T0 DD == 0, 8i, 2<D;

SCond •. displayRule

:2 I HD1 A1L w1 + 2 I A1 m1 w1 + 2 EI T1 s2 F A1 g1,1 +

•

• 2 •2 •
3 a A21 A1 g1,1 + 2 EI T1 s1 -I T1 s2 F A2 g1,2 + 2 EI T1 s1 a A2 A1 g1,1 g1,2 + 2 a A1 A2 A2 g21,2 +
•2 • •
EI T1 s1 a A2 A1 g1,1 g2,1 + 2 a A1 A2 A2 g1,2 g2,1 + 2 a A1 A2 A2 g1,1 g2,2 == 0,
2 I HD1 A2L w2 + 2 I A2 m2 w2 + 2 E-I T1 s1 +I T1 s2 F A1 g2,1 + E-IT1 s1 a A31 g1,1 g2,1 +
• • 2 • • 2
2 a A1 A2 A1 g1,2 g2,1 + 2 a A1 A2 A1 g2,1 + 2 a A1 A2 A1 g1,1 g2,2 + 3 a A22 A2 g2,2 == 0>

which is in agreement with that obtained by direct approach.

à Principal Parametric Resonance of the Second Mode

ResonanceConds = 8w2 == 3 w1 + e s1 , W == 2 w2 + e s2 <;

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

98w2 ® e s1 + 3 w1, W ® 2 e s1 + e s2 + 6 w1 <, 9w1 ® €€€€ H-e s1 + w2 L, W ® e s2 + 2 w2==

1
3

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand sides of eqOrder1 and obtain the solvability condition as

SCond = Table@Coefficient@-eqOrder1@@i, 2DD •. expRule@iD, Exp@I wi T0 DD == 0, 8i, 2<D;

SCond •. displayRule

:2 I HD1 A1L w1 + 2 I A1 m1 w1 + 3 a A21 A1 g1,1

• 2 •2
+ 2 EI T1 s1 a A2 A1 g1,1 g1,2 +
• 2 •2 • •
2 a A1 A2 A2 g1,2 + EI T1 s1 a A2 A1 g1,1 g2,1 + 2 a A1 A2 A2 g1,2 g2,1 + 2 a A1 A2 A2 g1,1 g2,2 == 0,
2 I HD1 A2L w2 + 2 I A2 m2 w2 + E-I T1 s1 a A31 g1,1 g2,1 + 2 a A1 A2 A1 g1,2 g2,1 +
•
• 2 • • • 2
2 a A1 A2 A1 g2,1 + 2 EI T1 s2 F A2 g2,2 + 2 a A1 A2 A1 g1,1 g2,2 + 3 a A22 A2 g2,2 == 0>

which is in agreement with that obtained by direct approach.

à Combination Parametric Resonance of the Two Modes

ResonanceConds = 8w2 == 3 w1 + e s1 , W == w1 + w2 + e s2 <;

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

98w2 ® e s1 + 3 w1, W ® e s1 + e s2 + 4 w1 <, 9w1 ® €€€€ H-e s1 + w2 L, W ® €€€€ H-e s1 + 3 e s2 + 4 w2L==

1 1
3 3

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand sides of eqOrder1 and obtain the solvability condition as
224 Chapter7.nb

SCond = Table@Coefficient@-eqOrder1@@i, 2DD •. expRule@iD, Exp@I wi T0 DD == 0, 8i, 2<D;

SCond •. displayRule

:2 I HD1 A1L w1 + 2 I A1 m1 w1 + 3 a A21 A1 g1,1

• 2 • •2
+ 2 EI T1 s2 F A2 g1,2 + 2 EIT1 s1 a A2 A1 g1,1 g1,2 +
• 2 •2 • •
2 a A1 A2 A2 g1,2 + EI T1 s1 a A2 A1 g1,1 g2,1 + 2 a A1 A2 A2 g1,2 g2,1 + 2 a A1 A2 A2 g1,1 g2,2 == 0,
2 I HD1 A2L w2 + 2 I A2 m2 w2 + 2 EI T1 s2 F A1 g2,1 + E-I T1 s1 a A31 g1,1 g2,1 +
•
• • 2 • • 2
2 a A1 A2 A1 g1,2 g2,1 + 2 a A1 A2 A1 g2,1 + 2 a A1 A2 A1 g1,1 g2,2 + 3 a A22 A2 g2,2 == 0>

which is in agreement with that obtained by direct approach.

Ÿ 7.2.4 Method of Time-Averaged Lagrangian

As a second alternative, we derive the modulation equations by using the method of time-averaged Lagrangian. The nondi-
mensional Lagrangian of the beam can be expressed as (Nayfeh, 1998)

Lagr1 = €€€€ à H¶t w@x, tDL2 â x - €€€€ à H¶x,x w@x, tDL2 âx -

1 1 1 1

2 0 2 0
i 1 y
HP + 2 e F Cos@W tDL à H¶x w@x, tDL2 âx - €€€€ e a j
jà H¶x w@x, tDL2 âxz
z + C@tD;
1 2
1
0 4 k 0 {

where C@tD is independent of w. Transforming the total time derivatives into partial derivatives in terms of T0 and T1 , we
modify Lagr1 as

Lagr2 = Lagr1 •. Integrate -> int •. 8w@x_, tD -> w@x, T0, T1 D,

Derivative@m_, n_D@wD@x_, tD -> dt@nD@D@w@x, T0, T1 D, 8x, m<DD, t -> T0 <

C@T0 D + €€€€ intBHe wH0,0,1L@x, T0 , T1 D + wH0,1,0L@x, T0 , T1 DL , 8x, 0, 1<F -

1 2

HP + 2 F e Cos@T0 WDL intBwH1,0,0L @x, T0, T1D , 8x, 0, 1<F -

2
2

€€€€ a e intBwH1,0,0L@x, T0 , T1 D , 8x, 0, 1<F - €€€€ intBwH2,0,0L @x, T0, T1 D , 8x, 0, 1<F
1 2 2 1 2

4 2

To this end, we let

solRule = w -> IEvaluateASumAAi@#3D fi @#1D EI wi + Ai@#3D fi @#1D E-I wi , 8i, 2<EE &M
#2 • #2

w ® HEI #2 w1 A1 @#3D f1 @#1D + EI #2 w2 A2 @#3D f2@#1D + E-I #2 w1 f1 @#1D A1@#3D + E-I#2 w2 f2 @#1D A2@#3D &L
• •

where the first independent variable (#1) stands for x and the last two independent variables (#2 and #3) stand for the two
time scales T0 and T1. Substituting solRule into Lagr2, using intRule1 and intRule2 to simplify the expansions of the
integrands, and collecting the coefficients of e, we obtain

order1Lagr =
Lagr2 •. solRule •. intRule1 ••. intRule2 •• TrigToExp •• Expand •• Coefficient@#, eD &;
Chapter7.nb 225

à Principal Parametric Resonance of the First Mode

ResonanceConds = 8w2 == 3 w1 + e s1 , W == 2 w1 + e s2 < •. Equal -> Rule

8w2 ® e s1 + 3 w1, W ® e s2 + 2 w1 <

Using ResonanceConds, we obtain the slowly varying terms from order1Lagr as

TAL = order1Lagr •. Exp@a_D :> Exp@Expand@a •. ResonanceCondsD •. e T0 -> T1 D •.

Exp@_ T0 + _.D -> 0 •. int -> Integrate

i
jà f¢1@xD2 â xy
z i 1 ¢ y
-E-I T1 s2 F j z A1 @T1D2 - 2 EI T1 s1 -I T1 s2 F j
jà f1@xD f¢2 @xD â xz
z A2@T1 D A1@T1 D -
•
k 0 { k 0 {
1

i 1 ¢ y• 3 i 1 ¢ y
EI T1 s2 F jjà f1 @xD2 â xz z A1 @T1D - €€€€ a j
jà f1 @xD2 â xz z A1@T1 D2 A1 @T1 D -
2
•
k 0 { 2 k 0 {
2 2

i
jà f¢1 @xD2 â xz y
zji
jà f¢1@xD f¢2 @xD â xz y
z A2@T1 D A1@T1 D3 -
EI T1 s1 a j
•
k 0 {k 0 {
1 1

i
jà f @xD f @xD â xz y
z A1@T1 D •
Fj A2@T1 D -
k 0 {
1
-I T1 s1 +I T1 s2 ¢ ¢
2E 1 2

i
jà f¢1 @xD2 â xyzijà f¢1@xD f¢2 @xD â xy z A1@T1 D3 •
E-I T1 s1 a j z j z A2 @T1 D -
k 0 {k 0 {
1 1

i 1 ¢ y
4aj
jà f1@xD f¢2 @xD â xzz A1 @T1D A2 @T1D A1 @T1D A2 @T1D -
2
• •
k 0 {
i
jà f1@xD â xz y i y
2aj zjjà f¢2 @xD2 â xzz A1@T1 D A2@T1 D A1@T1 D A2@T1 D -
k 0 {k 0 {
1
¢ 2
1 • •

3 i 1 ¢ y
€€€€ a j
jà f2@xD2 â xz
z A2@T1 D2 A2 @T1D -
2
•
2 k 0 {
2

i y i 1 y •
Ij jà f1@xD2 â xz
z w1 A1 @T1 D A¢1 @T1 D - I j
jà f2 @xD2 â xz
z w2 A2 @T1D A¢2 @T1D +
•
k 0 { k 0 {
1

i y i y
Ij jà f1@xD2 â xz
z w1 A1 @T1 D A1 @T1 D + I jjà f2 @xD2 â xz
z w2 A2 @T1D A2 @T1D
• •¢
k 0 { k 0 {
1 ¢ 1

The Euler-Lagrange equations corresponding to the TAL can be written as

eqMod1 = DADATAL, A1 @T1DE, T1 E - DATAL, A1 @T1DE == 0

•¢ •

i 1 ¢ y i 1 ¢ y•
2 EI T1 s1 -I T1 s2 F j
jà f1@xD f¢2 @xD â xz
z A2 @T1 D + 2 EI T1 s2 F j
jà f1 @xD2 â xz
z A1@T1 D +
k 0 { k 0 {
i 1 ¢ y i 1 ¢ yi y
3aj
jà f1@xD2 â xz
z A1 @T1D2 A1@T1 D + 3 EI T1 s1 a j
jà f1 @xD2 â xz
zjjà f¢1 @xD f¢2 @xD â xz
z A2 @T1D A1 @T1D +
2
• •
k 0 { k 0 {k 0 {
1 2

i 1 ¢ y
4aj
jà f1@xD f¢2 @xD â xzz A1 @T1D A2 @T1D A2 @T1D +
2
•
k 0 {
i y i y i 1 y
2aj
jà f1@xD â xz zjjà f¢2 @xD2 â xzz A1 @T1 D A2 @T1 D A2 @T1 D + 2 I j
jà f1 @xD2 â xz
z w1 A¢1 @T1D == 0
•
k 0 {k 0 { k 0 {
1 1
¢ 2

and
226 Chapter7.nb

eqMod2 = DADATAL, A2 @T1DE, T1 E - DATAL, A2 @T1DE == 0

•¢ •

i 1 ¢ y i 1 ¢ yji 1 y
2 E-I T1 s1 +I T1 s2 F j
jà f1@xD f¢2 @xD â xz
z A1@T1 D + E-I T1 s1 a j
jà f1 @xD2 â xz
z jà f¢1 @xD f¢2 @xD â xz
z A1 @T1D3 +
k 0 { k 0 {k 0 {
i 1 ¢ y
4aj
jà f1@xD f¢2 @xD â xz
z A1 @T1D A2 @T1D A1 @T1D +
2
•
k 0 {
i
jà f¢ @xD2 â xy
z i 1 ¢ y
2aj zjjà f2 @xD2 â xzz A1 @T1 D A2 @T1 D A1 @T1 D +
•
k 0 {k 0 {
1
1

i 1 ¢ y i 1 y
3aj
jà f2@xD2 â xz
z A2 @T1D2 A2@T1 D + 2 I j
jà f2@xD2 â xz
z w2 A¢2@T1 D == 0
2
•
k 0 { k 0 {

Adding linear viscous damping to eqMod1 and eqMod2 and performing integration by parts yields the same modulation
equations as obtained by directly attacking the partial-differential system.

Ÿ Principal Parametric Resonance of the Second Mode

ResonanceConds = Solve@8w2 == 3 w1 + e s1 , W == 2 w2 + e s2<, 8w2 , W<D@@1DD

8w2 ® e s1 + 3 w1, W ® 2 e s1 + e s2 + 6 w1 <

Using ResonanceConds, we obtain the slowly varying terms from order1Lagr as

TAL = order1Lagr •. Exp@a_D :> Exp@Expand@a •. ResonanceCondsD •. e T0 -> T1 D •.

Exp@_ T0 + _.D -> 0 •. int -> Integrate

i 1 ¢ y 3 i 1 ¢ y
-E-I T1 s2 F j
jà f2@xD2 â xz z A2 @T1D2 - €€€€ a j
jà f1@xD2 â xzz A1@T1 D2 A1 @T1D -
2
•
k 0 { 2 k 0 {
2

i yji y
EI T1 s1 a j
jà f¢1 @xD2 â xz
z jà f¢1@xD f¢2 @xD â xzz A2@T1 D A1@T1 D -
•
k 0 {k 0 {
1 1 3

i y i y
E-I T1 s1 a j
jà f¢1 @xD2 â xz
zjjà f¢1@xD f¢2 @xD â xzz A1@T1 D3 A2 @T1 D -
•
k 0 {k 0 {
1 1

i 1 ¢ y
4aj
jà f1@xD f¢2 @xD â xzz A1 @T1D A2 @T1D A1 @T1D A2 @T1D -
2
• •
k 0 {
i yji y
2aj
jà f¢1@xD2 â xz
z jà f¢2 @xD2 â xzz A1@T1 D A2@T1 D A1@T1 D A2@T1 D -
• •
k 0 {k 0 {
1 1

i
jà f¢2 @xD2 â xyz 3 i 1 ¢ y
EI T1 s2 F j z A2 @T1D - €€€€ a j jà f2 @xD2 â xzz A2@T1 D2 A2 @T1 D -
2

k 0 { 2 k 0 {
1 • 2 • 2

i
jà f1@xD2 â xy z w1 • i 1 y •
Ij z A1 @T1 D A¢1 @T1 D - I j
jà f2 @xD2 â xz
z w2 A2 @T1D A¢2 @T1D +
k 0 { k 0 {
1

i y i y
Ijjà f1@xD2 â xz z w1 A1 @T1 D A1 @T1 D + I jjà f2 @xD2 â xz
z w2 A2 @T1D A2 @T1D
•¢ •¢
k 0 { k 0 {
1 1

The Euler-Lagrange equations corresponding to the TAL can be written as

Chapter7.nb 227

eqMod1 = DADATAL, A1 @T1DE, T1 E - DATAL, A1 @T1DE == 0

•¢ •

i 1 ¢ y i 1 ¢ yi y
3aj
jà f1@xD2 â xz
z A1 @T1D2 A1 @T1 D + 3 EI T1 s1 a j
jà f1 @xD2 â xz
zjjà f¢1 @xD f¢2@xD â xz
z A2 @T1D A1 @T1D +
2
• •
k 0 { k 0 {k 0 {
1 2

i 1 ¢ y
4aj
jà f1@xD f¢2 @xD â xzz A1 @T1D A2 @T1D A2 @T1D +
2
•
k 0 {
i y i y i 1 y
2aj
jà f¢1@xD2 â xz
zjjà f¢2 @xD2 â xzz A1 @T1 D A2 @T1 D A2 @T1 D + 2 I j
jà f1 @xD2 â xz
z w1 A¢1 @T1D == 0
1 1 •
k 0 {k 0 { k 0 {

and

eqMod2 = DADATAL, A2 @T1DE, T1 E - DATAL, A2 @T1DE == 0

•¢ •

i
jà f¢1 @xD2 â xy
z i 1 ¢ y i 1 ¢ y
E-I T1 s1 a j zjjà f1@xD f¢2 @xD â xzz A1@T1 D3 + 4 a j jà f1 @xD f¢2@xD â xz z A1 @T1D A2 @T1D A1 @T1D +
2
•
k 0 {k 0 { k 0 {
1

i
jà f¢1@xD2 â xyzijà f¢2 @xD2 â xy
z A1 @T1 D A2 @T1 D • i 1 ¢ y•
2aj z j z A1 @T1 D + 2 EI T1 s2 F j
jà f2@xD2 â xz
z A2 @T1D +
k 0 {k 0 { k 0 {
1 1

i 1 ¢ y i 1 y
3aj
jà f2@xD2 â xz
z A2 @T1D2 A2@T1 D + 2 I j
jà f2@xD2 â xz
z w2 A¢2@T1 D == 0
2
•
k 0 { k 0 {

Adding linear viscous damping to eqMod1 and eqMod2 and performing integration by parts yields the same modulation
equations as obtained by directly attacking the partial-differential system.

Ÿ Combination Parametric Resonance of the Two Modes

ResonanceConds = Solve@8w2 == 3 w1 + e s1 , W == w1 + w2 + e s2 <, 8w2, W<D@@1DD

8w2 ® e s1 + 3 w1, W ® e s1 + e s2 + 4 w1 <

Using ResonanceConds, we obtain the slowly varying terms from order1Lagr as

228 Chapter7.nb

TAL = order1Lagr •. Exp@a_D :> Exp@Expand@a •. ResonanceCondsD •. e T0 -> T1 D •.

Exp@_ T0 + _.D -> 0 •. int -> Integrate

i 1 ¢ y 3 i 1 ¢ y
-2 E-I T1 s2 F j
jà f1@xD f¢2 @xD â xz
z A1@T1 D A2@T1 D - €€€€ a j
jà f1@xD2 â xz
z A1@T1 D2 A1 @T1D -
2
•
k 0 { 2 k 0 {
2

i y i y
EI T1 s1 a j
jà f¢1 @xD2 â xz
zjjà f¢1@xD f¢2 @xD â xz
z A2@T1 D A1@T1 D -
•
k 0 {k 0 {
1 1 3

i y i y
E-I T1 s1 a j
jà f¢1 @xD2 â xz
zjjà f¢1@xD f¢2 @xD â xz
z A1@T1 D3 A2 @T1 D -
•
k 0 {k 0 {
1 1

i 1 ¢ y• i 1 ¢ y
2 EI T1 s2 F j
jà f1 @xD f¢2 @xD â xz
z A1 @T1D A2 @T1D - 4 a j
jà f1 @xD f¢2 @xD â xz
z A1@T1 D A2@T1 D A1@T1 D A2@T1 D -
2
• • •
k 0 { k 0 {
i yji y
2aj jà f¢1@xD2 â xz z jà f¢2 @xD2 â xz z A1@T1 D A2@T1 D A1@T1 D A2@T1 D -
• •
k 0 {k 0 {
1 1

3 i 1 ¢ y
€€€€ a j
jà f2@xD2 â xz
z A2@T1 D2 A2 @T1D -
2
•
2 k 0 {
2

i
jà f1@xD2 â xy
z i 1 y •
Ij z w1 A1 @T1 D A¢1 @T1 D - I j
jà f2 @xD2 â xz
z w2 A2 @T1D A¢2 @T1D +
•
k 0 { k 0 {
1

i
jà f1@xD2 â xy
z w1 A1 @T1 D A¢1 @T1 D + I i
j y
Ij z jà f2 @xD2 â xz
z w2 A2 @T1D A2 @T1D
• •¢
k 0 { k 0 {
1 1

The Euler-Lagrange equations corresponding to the TAL can be written as

eqMod1 = DADATAL, A1 @T1DE, T1 E - DATAL, A1 @T1DE == 0

•¢ •

i 1 ¢ y i 1 ¢ yi y
3aj
jà f1@xD2 â xz
z A1 @T1D2 A1 @T1 D + 3 EI T1 s1 a j
jà f1 @xD2 â xz
zjjà f¢1 @xD f¢2@xD â xz
z A2 @T1D A1 @T1D +
2
• •
k 0 { k 0 {k 0 {
1 2

i 1 ¢ y• i 1 ¢ y
2 EI T1 s2 F j
jà f1 @xD f¢2 @xD â xz
z A2 @T1D + 4 a jjà f1 @xD f¢2 @xD â xz z A1@T1 D A2@T1 D A2@T1 D +
2
•
k 0 { k 0 {
i yji y i 1 y
2aj jà f¢1@xD2 â xz z jà f¢2 @xD2 â xz z A1 @T1 D A2 @T1 D A2 @T1 D + 2 I j
jà f1 @xD2 â xz
z w1 A¢1 @T1D == 0
•
k 0 {k 0 { k 0 {
1 1

and

eqMod2 = DADATAL, A2 @T1DE, T1 E - DATAL, A2 @T1DE == 0

•¢ •

i 1 ¢ yi y
E-I T1 s1 a j
jà f1 @xD2 â xz
zjjà f¢1@xD f¢2 @xD â xz
z A1@T1 D3 +
k 0 {k 0 {
1

i
jà f¢1 @xD f¢2 @xD â xy
z i 1 ¢ y
2 EI T1 s2 F j z A1 @T1D + 4 a jjà f1 @xD f¢2 @xD â xz
z A1@T1 D A2@T1 D A1@T1 D +
2
• •
k 0 { k 0 {
1

i
jà f¢1@xD2 â xy zijà f¢2 @xD2 â xy z A1 @T1 D A2 @T1 D •
2aj z j z A1 @T1 D +
k 0 {k 0 {
1 1

i 1 ¢ y i 1 y
3aj
jà f2@xD2 â xz
z A2 @T1D2 A2@T1 D + 2 I j
jà f2@xD2 â xz
z w2 A¢2@T1 D == 0
2
•
k 0 { k 0 {

Adding linear viscous damping to eqMod1 and eqMod2 and performing integration by parts yields the same modulation
equations as obtained by directly attacking the partial-differential system.
Chapter7.nb 229

à 7.3 Cantilever Beam

Ÿ 7.3.1 EOM and BC's

We consider the nonlinear nonplanar response of an inextensional cantilever beam to an external excitation of one of its
flexural modes. The lowest torsional frequencies of the beam considered are much higher than the frequencies of the excited
flexural modes so that the torsional inertia can be neglected. We assume that the beam is uniform and homogeneous. The
transverse deflections v@x, tD and w@x, tD of the beam at the position x and time t are governed by the nondimensional integral-
partial-differential equation (Crespo da Silva and Glynn, 1978)

EOM1 = 9¶t,t v@x, tD + b y ¶x,x,x,x v@x, tD == -2 e m@xD ¶t v@x, tD + e H1 - b yL

¶x J¶x,x w@x, tD à ¶x,x v@x, tD ¶x,x w@x, tD âx - ¶x,x,x w@x, tD à ¶x,x v@x, tD ¶x w@x, tD â xN -
x x

H1 - b y L
1 0

e €€€€€€€€€€€€€€€€€€€€€ ¶x,x J¶x,x w@x, tD à à ¶x,x v@x, tD ¶x,x w@x, tD âx â xN -

2 x x

bg
e b y ¶x H¶x v@x, tD ¶x H¶x v@x, tD ¶x,x v@x, tD + ¶x w@x, tD ¶x,x w@x, tDLL -
0 1

e €€€€ ¶x J¶x v@x, tD à ¶t,t Jà IH¶x v@x, tDL2 + H¶x w@x, tDL2M â xN âxN - e F@xD Cos@W tD,
1 x x

¶t,t w@x, tD + ¶x,x,x,x w@x, tD == -2 e m@xD ¶t w@x, tD - e H1 - b y L

2 1 0

¶x J¶x,x v@x, tD à ¶x,x v@x, tD ¶x,x w@x, tD âx - ¶x,x,x v@x, tD à ¶x,x w@x, tD ¶x v@x, tD â xN -
x x

H1 - b y L
1 0

e €€€€€€€€€€€€€€€€€€€€€ ¶x,x J¶x,x v@x, tD à à ¶x,x v@x, tD ¶x,x w@x, tD âx â xN -

2 x x

bg
e ¶x H¶x w@x, tD ¶x H¶x v@x, tD ¶x,x v@x, tD + ¶x w@x, tD ¶x,x w@x, tDLL -
0 1

e €€€€ ¶x J¶x w@x, tD à ¶t,t Jà IH¶x v@x, tDL2 + H¶x w@x, tDL2M â xN â xN=;
1 x x

2 1 0

along with the homogeneous boundary conditions

BC1 = 8v@0, tD == 0, ¶x v@x, tD == 0 •. x -> 0,

¶x,x v@x, tD == 0 •. x -> 1, ¶x,x,x v@x, tD == 0 •. x -> 1<

8v@0, tD == 0, vH1,0L @0, tD == 0, vH2,0L @1, tD == 0, vH3,0L@1, tD == 0<

BC = Join@BC1, BC1 •. v -> wD

8v@0, tD == 0, vH1,0L @0, tD == 0, vH2,0L @1, tD == 0, vH3,0L@1, tD == 0,

w@0, tD == 0, wH1,0L @0, tD == 0, wH2,0L @1, tD == 0, wH3,0L@1, tD == 0<

The spatial derivatives outside the integrals on the right-hand sides of EOM1 result in lengthy expressions. To simplify the
symbolic computation, we first define some operators as
230 Chapter7.nb

op1@v_, w_D = ¶x,x w@x, tD à ¶x,x v@x, tD ¶x,x w@x, tD âx -

x

¶x,x,x w@x, tD à ¶x,x v@x, tD ¶x w@x, tD âx •. Integrate -> int@1D

x

int@1D@vH2,0L@x, tD wH2,0L @x, tD, 8x, 1, x<D wH2,0L @x, tD -

int@1D@wH1,0L@x, tD vH2,0L@x, tD, 8x, 0, x<D wH3,0L @x, tD

op2@v_, w_D =

HoldA¶x,x w@x, tD à à ¶x,x v@x, tD ¶x,x w@x, tD â x âxE •. Integrate -> int@2D •• ReleaseHold
x x

0 1

@x, tD w @x, tD, 8x, 0, x<, 8x, 1, x<D wH2,0L@x, tD

H2,0L H2,0L
int@2D@v

op3@v_, w_D = ¶x v@x, tD à ¶t,t Jà IH¶x v@x, tDL2 + H¶x w@x, tDL2M â xN âx •. Integrate -> int@1D
x x

1 0

int@1DB

int@1DB2 vH1,1L@x, tD + 2 wH1,1L@x, tD + 2 vH1,0L @x, tD vH1,2L @x, tD + 2 wH1,0L @x, tD wH1,2L @x, tD,
2 2

8x, 0, x<F, 8x, 1, x<F vH1,0L @x, tD

where int@iD indicates the whole expression, including the integrand and the term multiplied by int@iD @exprD, will later be
wrapped by the ith spatial derivative in the solvability condition.

Then we rewrite EOM1 as

EOM = 9¶t,t v@x, tD + b y ¶x,x,x,x v@x, tD ==

H1 - b y L2
-2 e m@xD ¶t v@x, tD + e H1 - b y L op1@v, wD - e €€€€€€€€€€€€€€€€€€€€€ op2@v, wD -
bg
e b y ¶x H¶x v@x, tD ¶x H¶x v@x, tD ¶x,x v@x, tD + ¶x w@x, tD ¶x,x w@x, tDLL -
1
e €€€€ op3@v, wD - e F@xD Cos@W tD, ¶t,t w@x, tD + ¶x,x,x,x w@x, tD ==

H1 - b y L2
2

-2 e m@xD ¶t w@x, tD - e H1 - b y L op1@w, vD - e €€€€€€€€€€€€€€€€€€€€€ op2@w, vD -

bg

e ¶x H¶x w@x, tD ¶x H¶x v@x, tD ¶x,x v@x, tD + ¶x w@x, tD ¶x,x w@x, tDLL - e €€€€ op3@w, vD=;
1
2

It follows from EOM1 and BC that the linear undamped natural frequencies and their corresponding mode shapes are given
by

9ẁ1,m == l2m " ##

b y## , w2,n == l2n,
Cos@ liD + Cosh@ liD
fi @xD == Cosh@ li xD - Cos@ li xD + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ HSin@ li xD - Sinh@ li xDL H* i=m,n *L=;
Sin@ liD + Sinh@ liD

`
where w 1, m and w2, n are the natural frequencies in the y and z directions and the l i are the roots of

1 + Cos@ li D Cosh@ liD == 0

Chapter7.nb 231

We consider the case of one-to-one internal resonance between the mth mode in the y or v direction and the nth mode in the
`
z or w direction; that is, w 1, m » w2, n . To express the nearness of these frequencies quantitatively, we let

betayRule = 8b y -> 1 + d0 + e d1 <;

• !!!!!!!!!!!
so that w1, m = l2m 1 + d0 = l2n = w2, n . The beam has a near-square cross-section when d0 = 0. We assume neither of these
two modes is involved in an internal resonance with any other mode. Moreover, we consider a primary resonance of the
flexural mode in the y direction.

To manipulate some complicated integrals, we define the following rules:

intRule3 = 8int@i_D@fun_, arg2__D :> int@iD@Expand@funD, arg2D<;

intRule4 = 8int@i_D@a_ + b_, arg2__D :> int@iD@a, arg2D + int@iD@b, arg2D,
int@i_D@e n_. fun_, arg2__D :> e n int@iD@fun, arg2D,
int@i_D@a_ fun_, a1_, b1___D :> a int@iD@fun, a1, b1D •; FreeQ@a, First@a1DD,
int@i_D@int@i_D@a1_, a2_D fun_., a3__D :> int@iD@a1, a2D int@iD@fun, a3D •;
FreeQ@Rest@a2D, First@a2DD, int@i_D@int@i_D@a1_, a2_D, a3__D :> int@iD@a1, a3, a2D<;

Ÿ 7.3.2 Direct Attack of the Continuous Problem

In this section, we directly attack the integral-partial-differential equation EOM and associated boundary conditions BC and
seek a first-order uniform expansion of their solution in the form

solRule = 9v -> IEvaluateASumAej vj @#1, #2, #3D, 8j, 0, 1<EE &M,

w -> IEvaluateASumAej wj @#1, #2, #3D, 8j, 0, 1<EE &M=

8v ® Hv0 @#1, #2, #3D + e v1 @#1, #2, #3D &L, w ® Hw0 @#1, #2, #3D + e w1@#1, #2, #3D &L<

where the first independent variable stands for x and the last two independent variables stand for the two time scales T0 and
T1 . Transforming the total time derivatives in EOM into partial derivatives in terms of T0 and T1 , substituting the solRule and
betayRule into EOM and BC, expanding the result for small e, discarding terms of order higher than e, and using intRule3
and intRule4 to simplify the expansions of the integrands, we obtain

Ieq732a =
HHJoin@EOM, BCD •. 8v@x_, tD -> v@x, T0, T1D, w@x_, tD -> w@x, T0 , T1 D, Derivative@m_, n_D@
w_D@x_, tD -> dt@nD@D@w@x, T0 , T1 D, 8x, m<DD, t -> T0 < •.
solRule •. betayRule •• ExpandAllL •. intRule3 ••.
intRule4 •• ExpandAllL •. en_•;n>1 -> 0;M •• Timing

812.629 Second, Null<

Equating coefficients of like powers of e in eq732a, we obtain

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq732a •• Transpose;

232 Chapter7.nb

eqEps@@1DD •. displayRule

9v'''' + D20 v0 + Hv'''' L d0 == 0, w'''' + D20 w0 == 0, v0@0, T0 , T1 D == 0,

@0, T0 , T1 D == 0, v0 @1, T0, T1D == 0, v0 @1, T0 , T1 D == 0,
H1,0,0L H2,0,0L H3,0,0L
0 0 0

w0 @0, T0, T1D == 0, w0 @0, T0 , T1 D == 0, w0 @1, T0, T1D == 0, w0 @1, T0 , T1 D == 0=

H1,0,0L H2,0,0L H3,0,0L
v0

In the presence of damping, all modes that are not directly excited by the forcing or indirectly excited by the internal reso -
nance will decay with time. Hence, the solution of eqEps[[1]] can be expressed in terms of the two excited linear free-vibration
modes; that is,

sol0 = 9v0 -> FunctionA8x, T0 , T1 <, A1@T1 D f m @xD Exp@I w1,m T0D + A1@T1 D f m @xD Exp@-I w1,m T0 DE,
•

w0 -> FunctionA8x, T0, T1 <, A2@T1 D f n@xD Exp@I w2,n T0 D + A2@T1 D fn @xD Exp@-I w2,n T0DE=;
•

One can easily show that the mode shapes fi@xD are orthogonal. We assume that these modes are normalized so that

Ù0 fi@xD fj @xD â x -> dij

1

where dij is the Kronecker delta function.

Substituting sol0 into the first-order equations, eqEps[[2,1]] and eqEps[[2,2]], and using intRule3 and intRule4, we obtain

Clear@order1EqD

order1Eq@1D = HeqEps@@1, 1, 1DD •. v0 -> v1 L ==

HHeqEps@@1, 1, 1DD •. v0 -> v1L - HSubtract žž eqEps@@2, 1DD •. sol0 •• TrigToExp ••
ExpandL •. intRule3 ••. intRule4 •• ExpandL;

order1Eq@2D = HeqEps@@1, 2, 1DD •. w0 -> w1 L ==

HHeqEps@@1, 2, 1DD •. w0 -> w1L - HSubtract žž eqEps@@2, 2DD •. sol0 •• TrigToExp ••
ExpandL •. intRule3 ••. intRule4 •• ExpandL;

It follows from eqEps[[2]] that the first-order boundary conditions are

order1BC = Drop@eqEps@@2DD, 2D

9v1 @0, T0, T1D == 0, v1 @0, T0 , T1 D == 0, v1 @1, T0, T1D == 0, v1 @1, T0 , T1 D == 0,

H1,0,0L H2,0,0L H3,0,0L

w1 @0, T0, T1D == 0, @0, T0 , T1 D == 0, @1, T0, T1D == 0, @1, T0 , T1 D == 0=

H1,0,0L H2,0,0L H3,0,0L
w1 w1 w1

We consider the case of primary resonance of the flexural mode in the y direction, W » w1, m . To describe the nearness of the
primary resonance, we introduce the detuning parameter s by

omgList = 8w1,m , w2,n <;

ResonanceConds = 8w2,n == w1,m , W == w1,m + e s<;

and define the following rules:

Chapter7.nb 233

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

88w2,n ® w1,m , W ® e s + w1,m <, 8w1,m ® w2,n , W ® e s + w2,n <<

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand side of order1Eq[1] and order1Eq[2] and obtain the source of secular
terms as

ST = Table@Coefficient@order1Eq@iD@@2DD •. expRule@iD, Exp@I omgList@@iDD T0DD, 8i, 2<D;

The solvability conditions of order1Eq[i] and order1BC demand that ST[[i]] be orthogonal to fi @xD. Imposing these condi-
tions, we have

SCond1 = int@f m @xD ST@@1DD, 8x, 0, 1<D == 0 •. intRule1 ••. intRule2;

SCond2 = int@f n @xD ST@@2DD, 8x, 0, 1<D == 0 •. intRule1 ••. intRule2;

Recall that the ith spatial derivative should be recovered back from int@iD as stated in the preceding section. Hence, we define

intRule5@mode_D =
int@mode a_ int@i_D@b_, c__D, d_D :> int@mode HoldForm@D@a int@b, cD, 8x, i<DD, dD;

and rewrite SCond1 and SCond2 as

SCond = 8SCond1 •. intRule5@f m @xDD, SCond2 •. intRule5@f n@xDD<;

To simplify the notation in the solvability conditions, we use the orthonormality of the mode shapes and define the following
parameters:

H4L
notationRule = 9intAfi_@xD2 , 8x, 0, 1<E -> 1, intAfi_ @xD fi_ @xD, 8x, 0, 1<E -> l4i ,
intAm@xD fi_@xD2 , 8x, 0, 1<E -> mi , int@F@xD fi_@xD, 8x, 0, 1<D -> w1,i f=;

To identify the forms of the nonlinear terms in the solvability conditions, we first identify all of the possible forms of the
nonhomogeneous terms in the first-order problem. To this end, we let

basicTerms = TableA9Ai@T1 D EI omgList@@iDD T0 , Ai @T1 D E-I omgList@@iDD T0 =, 8i, 2<E •• Flatten

•

8EI T0 w1,m A1@T1 D, E-I T0 w1,m A1 @T1D, EI T0 w2,n A2@T1 D, E-I T0 w2,n A2 @T1D<
• •

Then, all of the possible forms of the nonlinear terms in the first-order problem are given by

cubicTerms = Nest@Outer@Times, basicTerms, #D &, basicTerms, 2D •• Flatten •• Union;

cubicTerms •. displayRule

:E3 I T0 w1,m A31 , EI T0 H2 w1,m +w2,n L A21 A2 , EI T0 Hw1,m+2 w2,n L A1 A22, E3 I T0 w2,n A32, EI T0 w1,m A21 A1 ,
•

EI T0 w2,n A1 A2 A1, EI T0 H-w1,m +2 w2,n L A22 A1 , E-I T0 w1,m A1 A1, EI T0 H-2 w1,m +w2,n L A2 A1 , E-3 IT0 w1,m A1 ,
• • •2 •2 •3

EI T0 H2 w1,m -w2,n L A21 A2 , EI T0 w1,m A1 A2 A2 , EI T0 w2,n A22 A2, E-I T0 w2,n A1 A1 A2, E-I T0 w1,m A2 A1 A2,
• • • • • • •

EI T0 H-2 w1,m -w2,n L A1 A2 , EI T0 Hw1,m-2 w2,n L A1 A2, E-I T0 w2,n A2 A2 , EI T0 H-w1,m-2 w2,n L A1 A2, E-3 I T0 w2,n A2 >
•2 • •2 •2 • •2 •3
234 Chapter7.nb

Out of these terms, only the terms that may lead to secular terms appear in the solvability conditions, which can be identified
according to

secularTerms =
IE-I omgList@@#DD T0 cubicTerms •. expRule@#D •. Exp@_ T0 + _.D -> 0 •• Union •• RestM & •ž 81, 2<

88A1 @T1D2 A1@T1 D, A1 @T1 D A2 @T1 D A1 @T1 D, A2 @T1D2 A1@T1 D, A1@T1 D2 A2 @T1 D,
• • • •

A1 @T1D A2 @T1D A2 @T1D, A2 @T1D2 A2 @T1 D<, 8A1 @T1D2 A1@T1 D, A1@T1 D A2@T1 D A1@T1 D,
• • • •

A2 @T1D2 A1@T1 D, A1 @T1 D2 A2@T1 D, A1@T1 D A2@T1 D A2@T1 D, A2@T1 D2 A2 @T1 D<<
• • • •

Next, we define the following parameters:

coef = Table@Coefficient@SCond@@i, 1DD, secularTerms@@iDDD, 8i, 2<D •• Collect@#, d0 D &;

Clear@symbolListD

symbolList@1D = MapIndexed@If@#1 =!= 0, 8 w1,m a1,#2@@1DD , 0D &, coef@@1DDD

88 a1,1 w1,m , 0, 8 a1,3 w1,m , 0, 8 a1,5 w1,m , 0<

symbolList@2D = MapIndexed@If@#1 =!= 0, 8 w2,n a2,#2@@1DD , 0D &, coef@@2DDD

80, 8 a2,2 w2,n , 0, 8 a2,4 w2,n , 0, 8 a2,6 w2,n <

and express SCond in a more concise form as

eqMod =

ExpandA- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ HHSCond@@#, 1DD •. notationRule •. Thread@secularTerms@@#DD -> 0DL +

1
omgList@@#DD
symbolList@#D.secularTerms@@#DDLE == 0 & •ž 81, 2<

d1 l4m A1 @T1D
: €€€€ EI T1 s f + 2 I m m A1 @T1D + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€ - 8 a1,1 A1@T1 D2 A1 @T1D -
1 •
w1,m
8 a1,3 A2 @T1 D A1 @T1D - 8 a1,5 A1@T1 D A2@T1 D A2@T1 D + 2 I A¢1 @T1D == 0, 2 I mn A2@T1 D -
2
2• •

8 a2,2 A1 @T1 D A2 @T1 D A1 @T1 D - 8 a2,4 A1 @T1D2 A2@T1 D - 8 a2,6 A2@T1 D2 A2 @T1D + 2 I A¢2@T1 D == 0>
• • •

where
Chapter7.nb 235

HsymbolList@1D -> coef@@1DD •• Thread •• Union •• RestL •. int -> Integrate •• Timing

:29.683 Second, :8 a1,1 w1,m ®

-3 à f m @xD f²m@xD3 â x - 12 à f m @xD f¢m @xD f²m@xD fH3L

m @xD â x - 3 à f m@xD f¢m @xD2 fH4L
m @xD â x +
1 1 1

i
j y
z
0 0 0

j-3 à f m @xD f²m@xD3 â x - 12 à f m@xD f¢m @xD f²m@xD fH3L

m @xD â x - 3 à f m@xD f m @xD f m @xD â xz d0 +
2 H4L
k {
1 1 1
¢

i
jà ¶8x,1< Jf m @xD à à f m@xD â x â xN f m @xD â xz y
0 0 0

2j z w1,m,
k 0 {
1 x x
¢ ¢ 2 2
1 0

8 a1,3 w1,m ® - à f m @xD f²m @xD f²n@xD2 â x - à f m @xD f¢n@xD f²m @xD fH3L
n @xD â x -
1 1

0 0

3à f m@xD f¢m @xD f²n @xD fH3L

n @xD â x - à f m @xD f¢m@xD f¢n @xD fH4L
n @xD â x +
1 1

i
j
0 0

jà ¶8x,1< JfH3L
n @xD à fn @xD f m @xD â xN f m @xD â x - à ¶8x,1< Jfn @xD à f m @xD fn@xD â xN f m @xD â x -
k 0
1 x 1 x
¢ ² ² ² ²
0 0 1

à f m@xD f²m @xD f²n @xD2 â x - à f m@xD f¢n @xD f²m@xD fH3L
n @xD â x -
1 1

y
z
0 0

3à f m @xD f¢m@xD f²n @xD fH3L

n @xD â x - à f m @xD f¢m@xD f¢n @xD fH4L
n @xD â xz d0 -
{
1 1

IÙ0 ¶8x,2< Hf²n@xD Ù0 Ù1 f²m @xD f²n@xD â x â xL f m@xD â xM d20

0 0
1 x x

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
bg
i 1 y 2
2j
jà ¶8x,1< Jf¢m @xD à à f¢n@xD2 â x â xN f m @xD â xz
z w2,n,
k 0 {
x x

1 0

8 a1,5 w1,m ® -2 à f m @xD f²m @xD f²n@xD2 â x - 2 à f m @xD f¢n@xD f²m@xD fH3L
n @xD â x -
1 1

0 0

6à f m@xD f¢m @xD f²n @xD fH3L

n @xD â x - 2 à f m @xD f¢m@xD f¢n @xD fH4L
n @xD â x +
1 1

i
j
0 0

j2 à ¶8x,1< JfH3L
n @xD à fn@xD f m @xD â xN f m @xD â x - 2 à ¶8x,1< Jfn @xD à f m @xD fn@xD â xN f m @xD
k 0
1 x 1 x
¢ ² ² ² ²
0 0 1

âx - 2 à f m @xD f²m@xD f²n @xD2 â x - 2 à f m @xD f¢n@xD f²m @xD fH3L

n @xD â x -
1 1

y
z
0 0

6à f m @xD f¢m@xD f²n @xD fH3L

n @xD â x - 2 à n @xD â xz d0 -
f m @xD f¢m@xD f¢n @xD fH4L
{
1 1

2 IÙ0 ¶8x,2< Hf²n@xD Ù0 Ù1 f²m @xD f²n@xD â x â xL f m @xD â xM d20

0 0

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >>
1 x x

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
bg
236 Chapter7.nb

HsymbolList@2D -> coef@@2DD •• Thread •• Union •• RestL •. int -> Integrate •• Timing

:22.282 Second, :8 a2,2 w2,n ® -2 à fn@xD f²m @xD2 f²n @xD â x -

1

6à fn@xD f¢n @xD f²m @xD fH3L

m @xD â x - 2 à fn @xD f¢m@xD f²n @xD fH3L
m @xD â x -
1 1

i
âx + j
0 0

2à fn@xD f¢m @xD f¢n @xD m @xD

fH4L j-2 à ¶8x,1< JfH3Lm @xD à f m @xD fn@xD â xN fn @xD â x +
k
1 1 x
¢ ²

y
z d0 -
0 0 0

2à ¶8x,1< Jf²m @xD à f²m @xD f²n @xD â xN fn @xD â xz

{
1 x

IÙ0 ¶8x,2< Hf²m@xD Ù0 Ù1 f²m @xD f²n @xD â x â xL fn @xD â xM d20

0 1
1 x x
2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , 8 a2,4 w2,n ®
bg

-à fn @xD f²m@xD2 f²n@xD â x - 3 à fn@xD f¢n @xD f²m @xD fH3L

m @xD â x - à fn @xD f¢m@xD f²n @xD fH3L
m @xD â x -
1 1 1

i
j
0 0 0

m @xD â x + j- à ¶8x,1< Jf m @xD à f m@xD fn @xD â xN fn@xD â x +

fn@xD f¢m @xD f¢n @xD fH4L H3L
à
k 0
1 1 x
¢ ²

y
à ¶8x,1< Jf m @xD à f m @xD fn @xD â xN fn @xD â xz
0 0

z d0 -
{
1 x
² ² ²

IÙ01 ¶8x,2< Hf²m@xD Ù0x Ù1x f²m @xD f²n@xD â x â xL fn@xD â xM d20
0 1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
bg
i 1 y 2
2j
jà ¶8x,1< Jf¢n @xD à à f¢m@xD2 â x â xN fn @xD â xz
z w1,m,
x x

k 0 1 0 {

8 a2,6 w2,n ® -3 à fn @xD f²n @xD3 â x - 12 à fn@xD f¢n @xD f²n@xD fH3L
n @xD â x -
1 1

i
j y
z 2
0 0

3à n @xD â x + 2 jà ¶8x,1< Jfn @xD à à fn@xD â x â xN fn @xD â xz w2,n>>

fn@xD f¢n @xD2 fH4L
1 1 x x

k 0 {
¢ ¢ 2
0 1 0

Modulation Equations in Polar Form

Using eqMod and the function PolarForm defined in the preceding section, we obtain the modulation equations in polar form
and the definitions for gi as

PolarForm@eqModD

::a¢1 @T1D == €€€€ H-f Sin@g1@T1 DD - 2 m m a1@T1 D + 2 Sin@g2 @T1DD a1,3 a1@T1 D a2@T1 D2L,
1

a¢2 @T1D == -mn a2 @T1 D - Sin@g2@T1 DD a2,4 a1 @T1D2 a2@T1 D,

2

f Cos@g1 @T1DD
g2¢ @T1D == - €€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 a1,1 a1@T1 D2 - 2 a2,2 a1 @T1D2 -
d1 l4m
w1,m a1@T1 D
2 Cos@g2@T1 DD a2,4 a1 @T1D2 + 2 Cos@g2 @T1DD a1,3 a2@T1 D2 + 2 a1,5 a2 @T1D2 - 2 a2,6 a2 @T1 D2,
f Cos@g1 @T1DD
g1¢ @T1D == s - €€€€€€€€€€€€m€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + a1,1 a1@T1 D2 + Cos@g2@T1 DD a1,3 a2 @T1D2 + a1,5 a2@T1 D2>,
d1 l4
2 w1,m 2 a1@T1 D
8g1 @T1D == T1 s - b1 @T1D, g2 @T1 D == -2 b1 @T1D + 2 b2@T1 D<>
Chapter7.nb 237

Modulation Equations in Cartesian Form and Symmetry Property

Using eqMod and the function CartesianForm defined in the preceding section, we obtain the modulation equations in
Cartesian form, the definitions for ni, and the symmetry property as

CartesianForm@eqModD

d1 l4m q1@T1 D
::p¢1 @T1D == -m m p1 @T1 D - n1 q1@T1 D + €€€€€€€€€€€€€€€€
€€€€€€€€€€€€ -
2 w1,m
a1,1 p1 @T1 D2 q1 @T1D + a1,3 p2 @T1 D2 q1@T1 D - a1,5 p2 @T1 D2 q1 @T1D - a1,1 q1 @T1D3 -
2 a1,3 p1 @T1D p2 @T1D q2 @T1D - a1,3 q1 @T1D q2 @T1D2 - a1,5 q1@T1 D q2@T1 D2, q¢1@T1 D ==
d1 l4m p1 @T1D
- €€€€ + n1 p1 @T1D - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + a1,1 p1@T1 D3 + a1,3 p1 @T1D p2 @T1D2 + a1,5 p1@T1 D p2@T1 D2 - m m q1@T1 D +
f
2 w1,m
a1,1 p1 @T1 D q1 @T1 D2 + 2 a1,3 p2@T1 D q1@T1 D q2@T1 D - a1,3 p1@T1 D q2@T1 D2 + a1,5 p1 @T1D q2 @T1D2 ,
2

p¢2 @T1D == -mn p2 @T1 D - 2 a2,4 p1 @T1D p2 @T1D q1 @T1D - n2 q2 @T1D - a2,2 p1 @T1D2 q2@T1 D +
a2,4 p1 @T1 D2 q2 @T1D - a2,6 p2 @T1 D2 q2@T1 D - a2,2 q1 @T1 D2 q2 @T1D - a2,4 q1 @T1D2 q2@T1 D - a2,6 q2@T1 D3,
q¢2 @T1D == n2 p2@T1 D + a2,2 p1@T1 D2 p2 @T1D + a2,4 p1 @T1D2 p2@T1 D + a2,6 p2@T1 D3 + a2,2 p2 @T1D q1 @T1D2 -
a2,4 p2 @T1 D q1 @T1 D2 - mn q2 @T1D + 2 a2,4 p1 @T1D q1 @T1D q2 @T1D + a2,6 p2 @T1 D q2 @T1 D2 >,

8n1 ® s, n2 ® s<, 88p1 , q1 , p2, q2<, 8p1 , q1, -p2 , -q2<<>

Ÿ 7.3.3 Discretization of the Continuous Problem

discretized form, we express v@x, tD and w@x, tD in terms of the linear mode shapes fm @xD and fn @xD, respectively, as
As an alternative, we apply the method of multiple scales to the discretized system of EOM and BC. To determine the

discretRule = 8v -> Function@8x, t<, q m @tD f m @xDD, w -> Function@8x, t<, q n@tD f n @xDD<;

modes = 8f m @xD, f n@xD<;

where the q i @tD are the generalized coordinates. We consider only q m @tD and q n @tD because in the presence of damping, all

ing the results with fm @xD and fn @xD, respectively, and integrating the outcome from x = 0 to x = 1 using intRule3, intRule4,
other modes that are not directly or indirectly excited decay with time. Substituting discretRule into each of EOM, multiply-

intRule1, intRule2, and intRule5, we obtain the following discretized form of the equations describing the response of the
beam:

eq733a =
Table@int@modes@@kDD # •. discretRule, 8x, 0, 1<D •. intRule3 ••. intRule4 •. intRule1 ••.
intRule2 •. intRule5@modes@@kDDD ••.
int@a_, b__D -> HoldForm@Integrate@a, bDD & •ž EOM@@kDD, 8k, 2<D

i 1 y i y
:j
jà f m@xD fH4L z j z ²
m @xD â xz by q m @tD + jà f m@xD â xz q m@tD ==
k 0 { k 0 {
1
2

i
jà f m@xD f²m @xD3 â xy
z
-e Cos@t WD à F@xD f m@xD â x - e j z by q m@tD3 -
k 0 {
1 1

0
238 Chapter7.nb

i 1 y i y
4ejjà fm @xD f¢m@xD f²m @xD fH3L z j z
m @xD â xz by q m@tD - e jà f m@xD f m @xD f m @xD â xz by q m @tD -
2 H4L
k 0 { k 0 {
1
3 ¢ 3

i 1 y
ej
jà ¶8x,1< JfH3L z
n @xD à fn@xD f m @xD â xN f m @xD â xz q m@tD qn @tD +
k 0 {
x
¢ ² 2

i
jà ¶8x,1< Jf²n @xD à f²m @xD f²n @xD â xN f m @xD â xy z q m@tD qn @tD2 +
0

ej z
k 0 {
1 x

i
jà ¶8x,1< JfH3L y
z
1

ej n @xD à fn@xD f m @xD â xN f m @xD â xz by q m @tD qn@tD -

k 0 {
1 x
¢ ² 2

i
jà ¶8x,1< Jf²n @xD à f²m @xD f²n @xD â xN f m @xD â xy z by q m @tD qn@tD2 -
0

ej z
k 0 {
1 x

i y
ej
jà f m@xD f²m @xD f²n @xD2 â xz
1

z by q m @tD qn@tD2 -
k 0 {
1

i y i y
ej
jà f m@xD f¢n @xD f²m @xD fH3L z j
n @xD â xz by q m@tD qn @tD - 3 e jà f m @xD f m@xD fn @xD fn @xD â xz
H3L z
k 0 { k 0 {
1 1
2 ¢ ²

i y
by q m@tD qn @tD2 - e j
jà f m @xD f¢m @xD f¢n@xD fH4L z
n @xD â xz by q m @tD qn@tD -
k 0 {
1
2

e IÙ0 ¶8x,2< Hf²n@xD Ù0 Ù1 f²m @xD f²n@xD â x â xL fm @xD â xM q m @tD qn@tD2

1 x x

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ +
bg

2 e IÙ0 ¶8x,2< Hf²n@xD Ù0 Ù1 f²m @xD f²n @xD â x â xL f m @xD â xM by q m@tD qn @tD2
1 x x

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
bg

e IÙ0 ¶8x,2< Hf²n@xD Ù0 Ù1 f²m @xD f²n@xD â x â xL fm @xD â xM b2y q m@tD qn @tD2
1 x x

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
bg
i 1 y ¢ i 1 y
2ej jà m@xD f m @xD2 â xzz q m@tD - e j jà ¶8x,1< Jf¢m@xD à à f¢m @xD2 â x â xN f m@xD â xz z q m @tD q¢m@tD2 -
k 0 { k 0 {
x x

i y
ej jà ¶8x,1< Jf¢m @xD à à f¢n@xD2 â x â xN f m @xD â xz
1 0

z q m@tD q¢n @tD2 -

k 0 {
1 x x

i y
ej jà ¶8x,1< Jf¢m @xD à à f¢m@xD2 â x â xN f m @xD â xz
1 0

z q m@tD2 q²m@tD -
k 0 {
1 x x

i
jà ¶8x,1< Jf @xD à à f @xD â x â xN f m @xD â xz y
z q m@tD qn @tD q²@tD,
1 0

ej
k 0 {
1 x x
¢ ¢ 2
m n n

i
j y
z i
j y
z ²
1 0

jà fn@xD fH4L
n @xD â xz qn@tD + jà fn @xD â xz qn @tD ==
k 0 { k 0 {
1 1
2

i y
ejjà ¶8x,1< JfH3L z
m @xD à f m@xD fn @xD â xN fn @xD â xz q m@tD qn@tD -
k 0 {
1 x
¢ ² 2

i 1 y
ej jà ¶8x,1< Jf²m @xD à f²m @xD f²n @xD â xN fn @xD â xz
0

z q m@tD2 qn@tD -
k 0 {
x

i y i y
ej jà fn@xD f²m @xD2 f²n @xD â xz z q m@tD2 qn@tD - 3 e j z
1

jà fn @xD f¢n @xD f²m@xD fH3L

m @xD â xz q m@tD qn@tD -
k 0 { k 0 {
1 1
2

i 1 y
ej jà fn@xD f¢m @xD f²n @xD fH3L z
m @xD â xz q m @tD qn @tD -
k 0 {
2

i
jà fn@xD f¢ @xD f¢ @xD fH4L@xD â xy z q m @tD2 qn @tD -
ej z
k 0 {
1
m n m

i
jà ¶8x,1< Jf m @xD à f m@xD fn @xD â xN fn @xD â xy z by q m @tD2 qn @tD +
ej H3L
z
k 0 {
1 x
¢ ²
0
Chapter7.nb 239

i 1 y
ej
jà ¶8x,1< Jf²m @xD à f²m @xD f²n @xD â xN fn @xD â xz
z by q m @tD2 qn @tD -
k 0 {
x

e IÙ01 ¶8x,2< Hf²m@xD Ù0x Ù1x f²m @xD f²n@xD â x â xL fn @xD â xM q m @tD2 qn @tD
1

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ +
bg

2 e IÙ01 ¶8x,2< Hf²m@xD Ù0x Ù1x f²m @xD f²n @xD â x â xL fn @xD â xM by q m@tD2 qn@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
bg

e IÙ01 ¶8x,2< Hf²m@xD Ù0x Ù1x f²m @xD f²n@xD â x â xL fn @xD â xM b2y q m@tD2 qn@tD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
bg
i 1 y i 1 y
ej
jà fn@xD f²n @xD3 â xzz qn @tD3 - 4 e j
jà fn @xD f¢n@xD f²n @xD fH3L z
n @xD â xz qn @tD -
k 0 { k 0 {
3

i 1 y
ej
jà fn@xD f¢n @xD2 fH4L z
n @xD â xz qn @tD -
k 0 {
3

i 1 y
ej
jà ¶8x,1< Jf¢n @xD à à f¢m@xD2 â x â xN fn @xD â xz z qn@tD q¢m @tD2 -
k 0 {
x x

i y i y
2ejjà m@xD fn @xD2 â xzz q¢n@tD - e j
jà ¶8x,1< Jf¢n@xD à à f¢n @xD2 â x â xN fn@xD â xz
1 0

z qn @tD q¢n@tD2 -
k 0 { k 0 {
1 1 x x

i
jà ¶8x,1< Jf @xD à à f @xD â x â xN fn @xD â xz y
z q m@tD qn @tD q @tD -
1 0

ej
k 0 {
1 x x
¢ ¢ 2 ²
n m m

i
jà ¶8x,1< Jf¢n @xD à à f¢n@xD2 â x â xN fn @xD â xy z qn@tD2 q²n@tD>
1 0

ej z
k 0 {
1 x x

1 0

where the damping is assumed to be modal. Using the orthonormality of the mode shapes fi @xD, we define

H4L
notationRule = 9intAfi_@xD2 , 8x, 0, 1<E -> 1, intAfi_ @xD fi_ @xD, 8x, 0, 1<E -> l4i ,
intAm@xD fi_@xD2 , 8x, 0, 1<E -> mi , int@F@xD fi_@xD, 8x, 0, 1<D -> w1,i f= •.
int@a_, b__D -> HoldForm@Integrate@a, bDD

:à fi_@xD2 â x ® 1, à fi_@xD fi_ @xD â x ® l4i, à m@xD fi_@xD2 â x ® mi , à F@xD fi_ @xD â x ® f w1,i >
H4L
1 1 1 1

0 0 0 0

To simplify the expression for eq733a, we first find possible nonlinear terms as follows:

basicTerms = Table@D@q m @tD, 8t, i<D, 8i, 0, 2<D

8q m @tD, q¢m@tD, q²m @tD<

linearTerms = Join@basicTerms, basicTerms •. m -> nD

8q m @tD, q¢m@tD, q²m @tD, qn @tD, q¢n@tD, q²n @tD<

cubicTerms = e Nest@Outer@Times, linearTerms, #D &, linearTerms, 2D •• Flatten •• Union;

The coefficients corresponding to cubicTerms in eq733a are obtained as

coefList = Coefficient@eq733a@@#, 2DD, cubicTermsD & •ž 81, 2<;

We replace these coefficients by ai, j as

240 Chapter7.nb

symbolList = MapIndexed@If@#1 =!= 0, aSequencežž #2, 0D &, coefList, 82<D;

Using notationRule, symbolList, and cubicTerms, we can rewrite eq733a as

EOM1 = Table@eq733a@@k, 1DD == Heq733a@@k, 2DD •. Thread@cubicTerms -> 0DL +

symbolList@@kDD.cubicTerms •. notationRule, 8k, 2<D

8by l4m q m @tD + q²m @tD == -f e Cos@t WD w1,m + e a1,1 q m @tD3 + e a1,3 q m @tD qn @tD2 - 2 e m m q¢m @tD +
e a1,8 q m @tD q¢m@tD2 + e a1,17 q m@tD q¢n @tD2 + e a1,21 q m@tD2 q²m@tD + e a1,37 q m @tD qn @tD q²n@tD,
l4n qn@tD + q²n@tD == e a2,2 q m @tD2 qn @tD + e a2,4 qn@tD3 + e a2,9 qn@tD q¢m @tD2 -
2 e mn q¢n @tD + e a2,18 qn@tD q¢n @tD2 + e a2,22 q m@tD qn @tD q²m@tD + e a2,38 qn @tD2 q²n @tD<

Using the method of multiple scales, we seek a first-order uniform expansion in the form

i1
j y
z
solRule = q n_ -> j
j z;
jâ e q n,j@#1, #2D &z
j
z
kj=0 {

Transforming the total time derivatives in EOM1 into partial derivatives in terms of T0 and T1, substituting the solRule and
betayRule into EOM1, expanding the result for small e, and discarding terms of order higher than e, we obtain

eq733b = HEOM1 •. 8qs_@tD -> qs @T0, T1D, Derivative@n_D@qs_D@tD -> dt@nD@qs @T0 , T1 DD, t -> T0< •.
solRule •. betayRule •• ExpandAllL •. e n_•;n>1 -> 0;

Equating coefficients of like powers of e in eq733b, we obtain

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq733b •• Transpose

:9l4m q m,0 @T0 , T1D + d0 l4m q m,0@T0 , T1D + q m,0 @T0 , T1D == 0, l4n qn,0@T0 , T1 D + qn,0 @T0 , T1 D == 0=,
H2,0L H2,0L

:f Cos@T0 WD w1,m + d1 l4m q m,0 @T0, T1 D - a1,1 q m,0@T0 , T1D3 + l4m q m,1@T0 , T1 D + d0 l4m q m,1 @T0 , T1 D -

a1,3 q m,0 @T0, T1D qn,0@T0 , T1 D2 + 2 m m q m,0 @T0 , T1 D - a1,8 q m,0 @T0, T1D q m,0 @T0 , T1D -
H1,0L H1,0L 2

a1,17 q m,0 @T0, T1 D qH1,0L

n,0 @T0 , T1D + 2 q m,0 @T0, T1 D - a1,21 q m,0 @T0 , T1D q m,0 @T0 , T1D +
2H1,1L 2 H2,0L

q m,1 @T0 , T1 D - a1,37 q m,0 @T0, T1 D qn,0@T0 , T1 D qn,0 @T0, T1 D == 0,

H2,0L H2,0L

-a2,2 q m,0 @T0, T1 D2 qn,0 @T0, T1 D - a2,4 qn,0@T0 , T1D3 + l4n qn,1@T0 , T1 D -
a2,9 qn,0 @T0, T1D q m,0 @T0 , T1D + 2 mn qn,0 @T0 , T1D - a2,18 qn,0 @T0, T1 D qn,0 @T0 , T1D +
H1,0L 2 H1,0L H1,0L 2

2 qn,0 @T0 , T1 D - a2,22 q m,0 @T0, T1 D qn,0@T0 , T1D q m,0 @T0, T1 D -

H1,1L H2,0L

a2,38 qn,0 @T0, T1 D2 qn,0 @T0, T1 D + qn,1 @T0, T1 D == 0>>

H2,0L H2,0L

Then it follows from eqEps[[1]] that

sol0 = 9q m,0 -> FunctionA8T0 , T1 <, A1@T1 D Exp@I w1,m T0D + A1@T1 D Exp@-I w1,m T0DE,
•

q n,0 -> FunctionA8T0 , T1 <, A2@T1 D Exp@I w2,n T0D + A2@T1 D Exp@-I w2,n T0DE=;
•

Substituting sol0 into eqEps[[2]] yields

order1Eq = H#@@1DD & •ž eqEps@@1DD •. q k_,0 -> q k,1L == H#@@1DD & •ž eqEps@@1DD •. qk_,0 -> q k,1L -
H#@@1DD & •ž eqEps@@2DDL •. sol0 •• TrigToExp •• ExpandAll •• Thread;
Chapter7.nb 241

We consider the case of primary resonance of the flexural mode in the y direction, W » w1, m . To describe the nearness of the
primary resonance, we introduce the detuning parameter s and define

omgList = 8w1,m , w2,n <;

ResonanceConds = 8w2,n == w1,m , W == w1,m + e s<;

OmgRule = Solve@ResonanceConds, Drop@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž 81, 2<

88w2,n ® w1,m , W ® e s + w1,m <, 8w1,m ® w2,n , W ® e s + w2,n <<

expRule@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

We substitute ResonanceConds into the right-hand sides of order1Eq and obtain the solvability condition as

SCond =
Table@Coefficient@-order1Eq@@i, 2DD •. expRule@iD, Exp@I omgList@@iDD T0 DD == 0, 8i, 2<D;
SCond •. displayRule

9A1 d1 l4m - 3 A21 A1 a1,1 - A22 A1 a1,3 - 2 A1 A2 A2 a1,3 +

• • •

€€€€ EI T1 s f w1,m + 2 I HD1A1 L w1,m + 2 I A1 m m w1,m - A21 A1 a1,8 w21,m + 3 A21 A1 a1,21 w21,m +
1 • •
2
• • • •
A22 A1 a1,17 w22,n - 2 A1 A2 A2 a1,17 w22,n + A22 A1 a1,37 w22,n + 2 A1 A2 A2 a1,37 w22,n == 0,
• • • • • •
-2 A1 A2 A1 a2,2 - A21 A2 a2,2 - 3 A22 A2 a2,4 - 2 A1 A2 A1 a2,9 w21,m + A21 A2 a2,9 w21,m + 2 A1 A2 A1 a2,22 w21,m +
A21 A2 a2,22 w21,m + 2 I HD1 A2L w2,n + 2 I A2 mn w2,n - A22 A2 a2,18 w22,n + 3 A22 A2 a2,38 w22,n == 0=
• • •

where
242 Chapter7.nb

# -> H# •. ai_,j_ :> coefList@@i, jDD •. betayRule •. e -> 0 •• Expand •• Collect@#, d0D &L & •ž
HsymbolList •• Flatten •• Union •• RestL

:a1,1 ® -à f m@xD f²m @xD3 â x - 4 à f m@xD f¢m@xD f²m @xD fH3L

m @xD â x - à f m @xD f¢m@xD2 fH4L
m @xD â x +
1 1 1

i
j y
z
0 0 0

j- à f m @xD f²m@xD3 â x - 4 à f m @xD f¢m @xD f²m@xD fH3L

m @xD â x - à f m@xD f m @xD f m @xD â xz d0,
2 H4L
k 0 {
1 1 1
¢
0 0

a1,3 ® -à f m@xD f²m @xD f²n@xD2 â x - à f m @xD f¢n @xD f²m@xD fH3L
n @xD â x -
1 1

0 0

3à f m@xD f¢m @xD f²n @xD fH3L

n @xD â x - à fm @xD f¢m@xD f¢n @xD fH4L
n @xD â x +
1 1

i
j
0 0

jà ¶8x,1< JfH3L
n @xD à fn @xD f m @xD â xN f m @xD â x - à ¶8x,1< Jfn @xD à f m @xD fn@xD â xN f m @xD â x -
k 0
1 x 1 x
¢ ² ² ² ²
0 0 1

à f m@xD f²m @xD f²n @xD2 â x - à f m@xD f¢n @xD f²m@xD fH3L
n @xD â x -
1 1

y
z
0 0

3à fm @xD f¢m@xD f²n @xD fH3L

n @xD â x - à f m @xD f¢m@xD f¢n @xD fH4L
n @xD â xz d0 -
{
1 1

IÙ01 ¶8x,2< Hf²n@xD Ù0x Ù1x f²m @xD f²n@xD

â x â xL f m@xD â xM d20
0 0

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
bg

a1,8 ® -à ¶8x,1< Jf¢m @xD à à f¢m@xD2 â x â xN f m @xD â x,

1 x x

0 1 0

a1,17 ® -à ¶8x,1< Jf¢m @xD à à f¢n@xD2 â x â xN f m @xD â x,

1 x x

0 1 0

a1,21 ® -à ¶8x,1< Jf¢m @xD à à f¢m@xD2 â x â xN f m @xD â x,

1 x x

0 1 0

a1,37 ® -à ¶8x,1< Jf¢m @xD à à f¢n@xD2 â x â xN f m @xD â x,

1 x x

0 1 0

a2,2 ® -à fn@xD f²m @xD2 f²n @xD â x - 3 à fn @xD f¢n @xD f²m@xD fH3L
m @xD â x -
1 1

0 0

à fn@xD f¢m @xD f²n @xD fH3L

m @xD â x - à fn@xD f¢m @xD f¢n @xD fH4L
m @xD â x +
1 1

i
j y
z
0 0

j- à ¶8x,1< JfH3L
m @xD à f m@xD fn @xD â xN fn@xD â x + à ¶8x,1< Jf m @xD à f m @xD fn @xD â xN fn @xD â xz
k 0 {
1 x 1 x
¢ ² ² ² ²

IÙ0 ¶8x,2< Hf²m@xD Ù0 Ù1 f²m @xD f²n@xD â x â xL fn@xD â xM d20

0 0 1
1 x x

d0 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
bg

a2,4 ® -à fn@xD f²n @xD3 â x - 4 à fn@xD f¢n @xD f²n @xD fH3L
n @xD â x - à fn @xD f¢n@xD2 fH4L
n @xD â x,
1 1 1

0 0 0

a2,9 ® -à ¶8x,1< Jf¢n @xD à à f¢m@xD2 â x â xN fn @xD â x,

1 x x

0 1 0

a2,18 ® -à ¶8x,1< Jf¢n @xD à à f¢n@xD2 â x â xN fn @xD â x,

1 x x

0 1 0

a2,22 ® -à ¶8x,1< Jf¢n @xD à à f¢m@xD2 â x â xN fn @xD â x,

1 x x

0 1 0

a2,38 ® -à ¶8x,1< Jf¢n @xD à à f¢n@xD2 â x â xN fn @xD â x>

1 x x

0 1 0
Chapter7.nb 243

which is in agreement with that obtained by direct approach.

Ÿ 7.3.4 Method of Time-Averaged Lagrangian

As a second alternative, we derive the modulation equations by using the method of time-averaged Lagrangian. The nondi-
mensional Lagrangian of the beam can be expressed as (Crespo da Silva and Glynn, 1978; Arafat, Nayfeh, and Chin, 1998)

Lagr1 =
1i 1 i y
HoldAà jj j i y 2z
j €€€€ je j¶t à €€€€ IH¶x v@x, tDL + H¶x w@x, tDL M âxz + H¶t v@x, tDL + H¶t w@x, tDL z -
x 1 2

0 k2 k k { {
2 2 2
0 2

e H1 - b y L J¶x,x v@x, tD ¶x,x w@x, tD à ¶x,x v@x, tD ¶x w@x, tD â xN -

x

€€€€ IH¶x,x w@x, tDL2 + e H¶x w@x, tDL2 H¶x,x w@x, tDL2 M - €€€€ b y IH¶x,x v@x, tDL2 +
1 1

e H¶x v@x, tDL2 H¶x,x v@x, tDL2 + e 2 ¶x v@x, tD ¶x,x v@x, tD ¶x w@x, tD ¶x,x w@x, tDM -
2 2

H1 - b y L2 i
e €€€€€€€€€€€€€€€€€€€€€ j
jJà ¶x,x v@x, tD ¶x,x w@x, tD âxN + 2 ¶x,x v@x, tD ¶x,x w@x, tD
x 2

2 bg k 1
y
z y
z âxE •.
à à ¶x,x v@x, tD ¶x,x w@x, tD â x âxz - e F@xD Cos@W tD v@x, tDz
z
x x

0 1 { {
Integrate -> int •. HoldPattern@¶t int@a_, b_DD :> int@¶t a, bD •• ReleaseHold;

where Hold and int are used to keep Integrate from being evaluated. Transforming the total time derivatives into partial
derivatives in terms of T0 and T1 , we modify Lagr1 as

Lagr2 = Lagr1 •. 8v@x_, tD -> v@x, T0 , T1D, w@x_, tD -> w@x, T0 , T1 D,

Derivative@m_, n_D@w_D@x_, tD -> dt@nD@D@w@x, T0 , T1 D, 8x, m<DD, t -> T0 <;

To this end, we let

solRule = 9v -> FunctionA8x, T0, T1<, A1 @T1D f m @xD Exp@I w1,m T0D + A1 @T1 D f m @xD Exp@-I w1,m T0 DE,
•

w -> FunctionA8x, T0, T1<, A2 @T1D f n @xD Exp@I w2,n T0D + A2@T1 D f n@xD Exp@-I w2,n T0 DE=;
•

and define

intRule6 = 8int@c1_. int@a1__D^n_., b1_D :>

c1 int@int@a1D^n •. intRule1 ••. intRule2 •. e -> 0, b1D •; FreeQ@c1, First@b1DD,
int@c1_ int@a1__D^n_., b1_D :> int@c1 int@a1D^ n •. intRule1 ••. intRule2 •. e -> 0, b1D<;

Substituting solRule and betayRule into Lagr2, using intRule1 , intRule2, and intRule6 to simplify the expansions of the
integrands, and collecting the coefficients of e, we obtain
244 Chapter7.nb

Horder1Lagr =
Lagr2 •. solRule •. betayRule •. intRule1 ••. intRule2 ••. intRule6 •. intRule1 ••.
intRule2 ••. intRule6 •. intRule1 ••. intRule2 ••
TrigToExp •• Expand •• Coefficient@#, eD &;L •• Timing

814.951 Second, Null<

ResonanceConds = 8w2,n == w1,m , W == w1,m + e s< •. Equal -> Rule

8w2,n ® w1,m , W ® e s + w1,m <

notationRule1 = 9intAfi_@xD2 , 8x, 0, 1<E -> 1,

int@F@xD fi_@xD, 8x, 0, 1<D -> f w1,i , intAf²i_@xD2 , 8x, 0, 1<E -> l4i =;

Using ResonanceConds and notationRule1, we obtain the slowly varying terms from order1Lagr as

TAL = order1Lagr •. Exp@a_D :> Exp@Expand@a •. ResonanceCondsD •. e T0 -> T1 D •.

Exp@_ T0 + _.D -> 0 •. notationRule1;

The Euler-Lagrange equations corresponding to the TAL can be written as

Chapter7.nb 245

eqMod1 =
DADATAL, A1 @T1DE, T1 E - DATAL, A1 @T1DE == 0 ••. int@a_, b__D -> HoldForm@Integrate@a, bDD
•¢ •

i 1 ¢ y
€€€€ EI T1 s f w1,m + d1 l4m A1 @T1D + 6 j jà f m @xD2 f²m @xD2 â xz z A1 @T1D2 A1@T1 D +
•
k 0 {
1

i y i y 2
z d0 A1 @T1D2 A1@T1 D - 2 j jà Jà f¢m @xD2 â xN â xz
2

6j jà f¢m@xD2 f²m@xD2 â xz z w1,m A1@T1 D2 A1 @T1D +

• •
k 0 { k 0 {
1 1 x 2

i
jà f¢m@xD f¢n @xD f²m @xD f²n@xD â xz y
z A2 @T1 D2 A1 @T1D -
0

2j
•
k 0 {
1

i
jà Jà fn @xD fm @xD â xN f m @xD f²n@xD â xy z d0 A2@T1 D2 •
2j z A1 @T1D +
k 0 {
1 x
¢ ² ²

i
jà f¢m@xD f¢n @xD f²m @xD f²n@xD â xy z d0 A2 @T1D2 •
0

2j z A1@T1 D +
k 0 {
1

IÙ0 HÙ1 f²m@xD f²n @xD â xL â xM d20 A2 @T1D2 A1@T1 D

1 x 2 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ +
bg

2 IÙ0 HÙ0 Ù1 f²m @xD f²n@xD â x â xL f²m @xD f²n@xD â xM d20 A2 @T1D2 A1@T1 D
1 x x •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ -
bg
i 1 y
2j
jà Jà f¢m @xD2 â xN à f¢n@xD2 â x â xz z w1,m w2,n A2 @T1D2 A1@T1 D +
•
k 0 {
x x

i
jà y
z A1 @T1 D A2 @T1 D •
0 0

4j f¢m@xD f¢n @xD f²m @xD f²n@xD â xz A2 @T1 D -

k 0 {
1

i
jà y
z d0 A1@T1 D A2@T1 D •
4j Jà f¢n @xD f²m @xD â xN f²m @xD f²n@xD â xz A2@T1 D +
k 0 {
1 x

i
jà y
z d0 A1 @T1D A2 @T1D •
0

4j f¢m@xD f¢n @xD f²m @xD f²n@xD â xz A2 @T1D +

k 0 {
1

2 IÙ01 HÙ1x f²m@xD f²n @xD â xL â xM d20 A1 @T1D A2 @T1D A2 @T1D

2 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
bg

4 IÙ01 HÙ0x Ù1x f²m @xD f²n@xD â x â xL f²m @xD f²n@xD â xM d20 A1 @T1D A2 @T1D A2 @T1D
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + 2 I w1,m A¢1 @T1 D == 0
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
bg

and
246 Chapter7.nb

eqMod2 =
DADATAL, A2 @T1DE, T1 E - DATAL, A2 @T1DE == 0 ••. int@a_, b__D -> HoldForm@Integrate@a, bDD
•¢ •

i 1 ¢ y
4j
jà f m@xD f¢n @xD f²m @xD f²n@xD â xz z A1@T1 D A2@T1 D A1@T1 D -
•
k 0 {
i y
4j jà Jà fn @xD fm @xD â xN f m @xD f²n@xD â xz z d0 A1@T1 D A2@T1 D A1@T1 D +
•
k 0 {
1 x
¢ ² ²

i
jà f m@xD fn @xD f m @xD fn@xD â xz y
z d0 A1 @T1D A2 @T1D •
0

4j A1 @T1D +
k 0 {
1
¢ ¢ ² ²

2 IÙ0 HÙ1 f²m@xD f²n @xD â xL â xM d20 A1 @T1D A2 @T1D A1 @T1D

1 x 2 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
bg

4 IÙ0 HÙ0 Ù1 f²m @xD f²n@xD â x â xL f²m @xD f²n@xD â xM d20 A1 @T1D A2 @T1D A1 @T1D
1 x x •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
bg
i 1 ¢ y
2j jà f m@xD f¢n @xD f²m @xD f²n@xD â xz z A1 @T1 D2 A2 @T1D -
•
k 0 {
i y
2j jà Jà f¢n @xD f²m @xD â xN f²m @xD f²n@xD â xz z d0 A1@T1 D2 A2 @T1D +
•
k 0 {
1 x

i y
2j jà f¢m@xD f¢n @xD f²m @xD f²n@xD â xz
0

z d0 A1 @T1D2 A2@T1 D +
•
k 0 {
1

IÙ0 HÙ1 f m@xD fn @xD â xL â xM d0 A1 @T1D2 A2@T1 D

1 x ² ² 2 2 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ +
bg

2 IÙ0 HÙ0 Ù1 f²m @xD f²n@xD â x â xL f²m @xD f²n@xD â xM d20 A1 @T1D2 A2@T1 D
1 x x •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ -
bg
i 1 y
2j
jà Jà f¢n@xD2 â x â xz
f¢m @xD2 â xN à z w1,m w2,n A1 @T1D2 A2@T1 D +
•
k 0 {
x x

i 1 y
6j f¢n@xD2 f²n@xD2 â xz
0 0

jà z A2@T1 D2 A2 @T1 D -
•
k 0 {
i 1 y 2
2j
jà Jà f¢n @xD2 â xN â xz z w2,n A2 @T1 D2 A2 @T1D + 2 I w2,n A¢2@T1 D == 0
2 •
k 0 {
x

Adding linear viscous damping to eqMod1 and eqMod2 and performing integration by parts yields the same modulation
equations as obtained by directly attacking the partial-differential system.
Chapter 8
Continuous Systems with Quadratic and
Cubic Nonlinearities
In this chapter, we use the method of multiple scales to determine second-order uniform asymptotic solutions of continuous
or distributed-parameter systems with quadratic and cubic nonlinearities. We consider some of the internal resonances
arising from the cubic nonlinearities, namely, one-to-one and three-to-one internal resonances. Two-to-one internal reso-
nances are treated in Chapter 9. We directly attack these continuous systems because treatment of reduced-order models
obtained with the Galerkin procedure might lead to erroneous results. We consider buckled beams in Section 8.1, circular
cylindrical shells in Section 8.2, and near-square plates in Section 8.3.

à Preliminaries
Off@General::spell1, Integrate::generD

Needs@"Utilities`Notation`"D

To determine second-order uniform asymptotic expansions of the solutions of buckled beams, circular cylindrical shells, and
near-square plates by using the method of multiple scales, we introduce the three time scales T0 = t, T1 = e t, and T2 = e2 t,
where e is a small nondimensional parameter and the order of the amplitude of oscillations. Moreover, we symbolize these
scales according to

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

timeScales = 8T0 , T1, T2 <;

In terms of the time scales T0, T1 , and T2, the time derivatives can be expressed as

dt@0D@expr_D := expr; dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, 2<E;

dt@2D@expr_D := Hdt@1D@dt@1D@exprDD •• ExpandL •. ei_•;i>3 -> 0;

In the course of the analysis, we need the complex conjugates of A and other variables. We define them by using the
following rule:

conjugateRule = 9A -> A, A -> A, Complex@m_, n_D -> Complex@m, -nD=;

• •

To speed up manipulations of the integrals involved in the governing equations with Mathematica, we introduce the rules
248 Chapter8.nb

intRule1 = 8int@fun_, arg2__D :> int@Expand@funD, arg2D<;

intRule2 = 8int@a_ + b_, arg2__D :> int@a, arg2D + int@b, arg2D,
int@e n_. fun_, arg2__D :> e n int@fun, arg2D,
int@a_ fun_, a1_, b1___D :> a int@fun, a1, b1D •; FreeQ@a, First@a1DD,
int@int@a1_, a2_D fun_., a3__D :> int@a1, a2D int@fun, a3D •; FreeQ@Rest@a2D, First@a2DD<;

To display the outputs in easily read expressions, we introduce the rules

displayRule = 9Derivative@a_, b__DAw_i_ E@_, T0 , __D :>

#2@@1DD-1 &, 8b<E; arg1 =!= 1, arg1, ""E,

SequenceFormAIfAarg1 = Times žž MapIndexedAD #1
wi SequenceForm žž Table@"'",8a<DE,
Derivative@a__D@Ai__D@__D :> SequenceFormATimes žž MapIndexedAD#1
#2@@1DD &, 8a<E, Ai E,

Derivative@a__D@Ai__D@__D :> SequenceFormATimes žž MapIndexedAD#1

#2@@1DD &, 8a<E, Ai E,
• •

w_i_ @_, T0, __D -> wi , Ai__@__D -> Ai , Ai__@__D -> Ai ,

• •

Exp@a_. + b_. Complex@0, m_D T0 + c_. Complex@0, n_D T0D -> Exp@a + Hm * b + n * cL I T0 D=;

à 8.1 Buckled Beams

In this section, we consider the nonlinear transverse response of a buckled beam possessing either a one-to-one or a three-
to-one internal resonance to a principal parametric resonance of the higher mode. The analysis assumes a unimodal buckled
deflection.

The equation governing nonlinear transverse vibrations of a parametrically excited homogeneous beam is (Nayfeh and
Mook, 1979)

eq81a =
i y
m wt,t + c wt + Es Is wx,x,x,x + j
jP - f p Cos@W tD - €€€€€€€€€€ à Hwx L2 âxz
z wx,x == 0 •. w m__ :> ¶ m w@x, tD;
Es A {

k 2{ 0 {

where P is a static axial load, fp Cos@W tD is an axial harmonic excitation, w@x, tD is the transverse deflection of the beam, x is
the distance along the undeflected beam, and t is time. Other dimensional parameters are the damping coefficient c and the

• !!!!!!
beam length {, mass per unit length m, cross-sectional area A, modulus of elasticity Es , and cross-sectional area moment of
inertia Is . We note that the subscript s is used with I and E so that Mathematica will not confuse them with -1 and the
exponential function . Equation eq81a is supplemented with the four linear homogeneous boundary conditions

bc81a = 8B1 @w@0, tDD == 0, B2@w@0, tDD == 0, B3 @w@{, tDD == 0, B4@w@{, tDD == 0<;

where the Bi are linear homogeneous operators.

It is good practice to cast the governing equations and boundary conditions in nondimensional form. Thus, we introduce the
following nondimensional variables:

T = $ %%€€€€€€€€
%%%%
m {4%%%%
€€€€€€ ;
Es Is
Chapter8.nb 249

nondimRule = 9w -> H{ w @#1 • {, #2 • TD &L, x -> { x, t -> T t, W -> W • T,

Ž Ž Ž Ž

Ž • !! !!! !!!! !
P -> P Es Is • {2 , f p -> f Es Is • {2, c -> c Es Is m ’ {2, A -> 2 a Is • {2 =;
Ž Ž

where the tilde indicates a nondimensional variable. Substituting nondimRule into eq81a and dropping the tildes, we obtain

i {3
eq81b = j
j €€€€€€€€€€€€ eq81a@@1DD •. Integrate -> int •. nondimRule •. s_ -> s •.
Ž
k Es Is
y
int@a_, 8x {, __<D -> { int@a, 8x, 0, 1<D •• Expand •• PowerExpandz
z == 0;
{
eq81b •.
int ->
Integrate

c wH0,1L@x, tD + wH0,2L@x, tD + P wH2,0L @x, tD -

i
jà wH1,0L@x, tD2 â xy
z
f Cos@t WD wH2,0L @x, tD - a j z wH2,0L@x, tD + wH4,0L@x, tD == 0
k 0 {
1

Ÿ 8.1.1 Postbuckling Deflection

Pk and their corresponding mode shapes fk @xD. We symbolize the Pk as

The first step in analyzing the nonlinear vibrations of a buckled beam is the determination of the critical Euler buckling loads

Symbolize@P kD;

Dropping the time derivative and neglecting the forcing and nonlinear terms in eq81b, we define the linear buckling problem
as

buckEq = eq81b •. int -> H0 &L •. 8P -> P k , f -> 0, w -> Hf k@#1D &L<

Pk f²k @xD + fk @xD == 0

H4L

Solving buckEq for fk @xD yields

H4L

H4L
buckRule = SolveAbuckEq, f k @xDE@@1DD

9fk @xD ® -Pk f²k @xD=

H4L

Next, we increase the axial load P beyond Pk and assume that the postbuckling displacement is b fk , where b is a nondimen-
sional measure of the buckling level. Substituting this displacement into eq81b, dropping the forcing term and the time
derivative, and using intRule2 and buckRule, we obtain

bEq = Factor@eq81b@@1DD •. 8f -> 0, w -> Hb f k@#1D &L< •. intRule2 •. buckRuleD == 0

-b H-P + Pk + b2 a int@f¢k @xD2 , 8x, 0, 1<DL f²k @xD == 0

It follows from bEq that either b = 0 (unbuckled case) or b is given by

250 Chapter8.nb

IbRule = b2 -> Ib2 •. Solve@bEq, bD@@2DDMM •. int -> Integrate

P - Pk
a Ù0 fk @xD2 â x
b2 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
1 ¢

And solving bEq for P yields

HPRule = Solve@bEq, PD@@1DDL •. int -> Integrate

:P ® Pk + b2 a à f¢k @xD2 â x>

1

Because P is greater than Pk , PRule yields the amplitude of the kth buckling mode.

b fk @xD, corresponding to the kth buckled mode, and a time-dependent relative deflection u@x, tD; that is,
Having solved the postbuckling problem, we assume that the beam deflection is the sum of the static buckled displacement

wRule = 8w -> Hb f k@#1D + u@#1, #2D &L<;

Substituting wRule into eq81b and using intRule1, intRule2, buckRule, and PRule, we find that the nonlinear transverse
vibrations of the beam around its buckled configuration are given by

EOM = eq81b •. wRule •. intRule1 ••. intRule2 •. buckRule •. PRule •• ExpandAll;

EOM •. int -> Integrate

i 1 ¢ y ² i 1 H1,0L y ²
-b f Cos@t WD f²k @xD - 2 b2 a j
jà fk @xD uH1,0L@x, tD â xzz fk @xD - b a j
jà u @x, tD â xz
z fk @xD +
k 0 { k 0 {
2

c uH0,1L@x, tD + uH0,2L@x, tD + Pk uH2,0L @x, tD - f Cos@t WD uH2,0L @x, tD -

i 1 ¢ y H2,0L i 1 H1,0L y H2,0L
2baj
jà fk @xD uH1,0L@x, tD â xz
zu @x, tD - a j
jà u @x, tD â xz
zu @x, tD + uH4,0L @x, tD == 0
k 0 { k 0 {
2

BC = bc81a •. 8w -> u, { -> 1<

8B1 @u@0, tDD == 0, B2@u@0, tDD == 0, B3 @u@1, tDD == 0, B4 @u@1, tDD == 0<

static deflection fk @xD can be written as

In the following analysis, we focus on hinged-hinged , first mode-buckled beams . Then, the boundary conditions for the

bBC = 8f k @0D == 0, f²k@0D == 0, f k @1D == 0, f²k @1D == 0<;

and the boundary conditions for the dynamic deflection u@x, tD can be written as

BC1 = 8u@0, tD == 0, ¶x,x u@x, tD == 0 •. x -> 0, u@1, tD == 0, ¶x,x u@x, tD == 0 •. x -> 1<

8u@0, tD == 0, uH2,0L @0, tD == 0, u@1, tD == 0, uH2,0L@1, tD == 0<

For the case of first buckled mode, it follows from buckEq and bBC that the first critical buckling load and associated mode-
shape are
Chapter8.nb 251

bucklingLoad = P k -> p 2 ;
modeshape = 8f k -> HSin@p #D &L<;

where the buckling modeshape was normalized so that fk @1 • 2D = 1.

Ÿ Linear Vibrations

The next step in analyzing the nonlinear vibrations of a buckled beam is the calculation of its linear natural frequencies and
corresponding modeshapes. To this end, we linearize EOM in u@x, tD, drop the forcing, damping, and nonlinear terms, and
obtain

linearEq = EOM •. 8c -> 0, f -> 0< •. u -> He u@#1, #2D &L •. intRule2 •. e n_•;n>1 -> 0 •. e -> 1;
linearEq •. int -> Integrate

i 1 ¢ y ²
-2 b2 a j
jà fk @xD uH1,0L @x, tD â xz
z fk @xD + uH0,2L@x, tD + Pk uH2,0L @x, tD + uH4,0L@x, tD == 0
k 0 {

The boundary conditions are the same as BC1, which we list below as

linearBC = BC1

8u@0, tD == 0, uH2,0L @0, tD == 0, u@1, tD == 0, uH2,0L@1, tD == 0<

Then, using separation of variables, we seek the solution of linearEq and linearBC in the form

uSol = u -> Function@8x, t<, F@xD Exp@I w tDD;

where w is the nondimensional linear natural frequency and F@xD is the corresponding modeshape. Substituting uSol into
linearEq and linearBC and using intRule2 and Thread, we obtain the equation governing F@xD as

FEq = linearEq •. uSol •. intRule2 •• ThreadAE-I w t #, EqualE & •• ExpandAll;

FEq •. int -> Integrate

i 1 ¢ y ²
-w2 F@xD + Pk F² @xD - 2 b2 a j
jà F @xD f¢k @xD â xz
z fk @xD + FH4L @xD == 0
k 0 {

and the boundary conditions as

FBC = linearBC •. u -> HF@#1D &L

8F@0D == 0, F² @0D == 0, F@1D == 0, F²@1D == 0<

To find the general solution of FEq and FBC, we follow Nayfeh, Kreider, and Anderson (1995) and treat the definite integral

expressed as a linear combination of homogeneous and particular solutions. Because fk @xD is a solution of buckEq, the
as a constant B; thus, we have a nonhomogeneous linear ordinary-differential equation. Hence, its general solution can be

particular solution is proportional to fk ² @xD. Consequently, we express the general solution of FEq and FBC as

cList = Table@ci , 8i, 5<D;

shapeFunc = 8Cos@ l1 xD, Sin@ l1 xD, Cosh@ l2 xD, Sinh@ l2 xD, f k² @xD<;
252 Chapter8.nb

FSol = F -> Function@x, cList.shapeFunc •. modeshape •• EvaluateD

F ® Function@x, Cos@x l1D c1 + Sin@x l1 D c2 + Cosh@x l2D c3 + Sinh@x l2 D c4 - p 2 Sin@p xD c5D

where the li satisfy the characteristic equation

charEq =
FEq •. b -> 0 •. bucklingLoad •. F -> HExp@ l #D &L •• ThreadAE-l x #, EqualE & •• ExpandAll

p 2 l2 + l4 - w2 == 0

lRule = 9- l21, l22 = -> I l2 •. Solve@charEq, lD •• UnionM •• Thread

1 j 2 " # #### #### #### z 1 j 2 " ## 4#### #### ##

#z
:-l21 ® €€€€ i-p - p 4 + 4 w2 y , l22 ® €€€€ i-p + p + 4 w2 y >
2 k { 2 k {

Substituting bucklingLoad, modeshape, and FSol into the left-hand sides of FBC and FEq and using intRule1 and int-
Rule2, we obtain

expr1 = Append@#@@1DD & •ž FBC •. FSol,

FEq@@1DD •. bucklingLoad •. modeshape •. FSol •. intRule1 ••. intRule2 •. int -> IntegrateD

:c1 + c3 , -c1 l21 + c3 l22, Cos@l1 D c1 + Sin@l1D c2 + Cosh@l2 D c3 + Sinh@l2D c4 ,

-Cos@l1D c1 l21 - Sin@l1D c2 l21 + Cosh@l2D c3 l22 + Sinh@l2D c4 l22,
-p 6 Sin@p xD c5 - w2 HCos@x l1 D c1 + Sin@x l1D c2 + Cosh@x l2 D c3 + Sinh@x l2D c4 - p 2 Sin@p xD c5 L +
Cos@x l1 D c1 l41 + Sin@x l1D c2 l41 + Cosh@x l2D c3 l42 + Sinh@x l2D c4 l42 +
p 2 Hp 4 Sin@p xD c5 - Cos@x l1D c1 l21 - Sin@x l1D c2 l21 + Cosh@x l2 D c3 l22 + Sinh@x l2 D c4 l22 L +
i 1 4 p Sin@l1 D c2 l21 i Cos@l1D l1 y
2 b2 p 2 a Sin@p xD j
j- €€€€ p c5 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - p c1 l1 j- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z -
j
Hp L Hp L Hp L Hp L Hp L Hp L
l1

k 2 - l + l k - l + l - l + l {
p Sinh@l2D c4 l22 Cosh@l2D l2 yy
1 1 1 1 1 1

i
€€€€€€€€€€€€€€ + p c3 l2 j
j- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ zz
€€€€€€€€€€€€€€ zz
z>
l2
Hp - I l2 L Hp + I l2 L k Hp - I l2 L Hp + I l2 L Hp - I l2L Hp + I l2L {{
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

The coefficient matrix of cList in expr1 can be obtained as

Chapter8.nb 253

coefMat = Outer@Coefficient, expr1, cListD

:81, 0, 1, 0, 0<, 8-l21 , 0, l22 , 0, 0<, 8Cos@l1 D, Sin@l1D, Cosh@l2 D, Sinh@l2D, 0<,

8-Cos@l1D l21 , -Sin@l1D l21 , Cosh@l2D l22 , Sinh@l2D l22 , 0<, :-w2 Cos@x l1 D -

2 b2 p 3 a Cos@l1D Sin@p xD l21

p 2 Cos@x l1 D l21 + Cos@x l1D l41 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 b2 p 3 a Sin@p xD l21
Hp - l1 L Hp + l1 L Hp - l1 L Hp + l1 L
€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

2 b2 p 3 a Sin@p xD Sin@l1D l21

-w2 Sin@x l1 D - p 2 Sin@x l1D l21 + Sin@x l1 D l41 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , -w2 Cosh@x l2 D +
Hp - l1 L Hp + l1 L
2 b2 p 3 a Cosh@l2 D Sin@p xD l22
p 2 Cosh@x l2 D l22 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + Cosh@x l2D l42 ,
2 b2 p 3 a Sin@p xD l22
Hp - I l2L Hp + I l2L Hp - I l2 L Hp + I l2 L
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

2 b2 p 3 a Sin@p xD Sinh@l2 D l22

-w2 Sinh@x l2 D + p 2 Sinh@x l2D l22 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + Sinh@x l2 D l42,
Hp - I l2 L Hp + I l2 L
-b2 p 6 a Sin@p xD + p 2 w2 Sin@p xD>>

Substituting charEq into coefMat[[5]] and eliminating the common factor Sin@p xD, we have

coefMat@@5DD = Coefficient@coefMat@@5DD, Sin@p xDD

2 b2 p 3 a Cos@l1 D l21 2 b2 p 3 a Sin@l1 D l21

: €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 b2 p 3 a l21
Hp - l1 L Hp + l1 L Hp - l1L Hp + l1L Hp - l1 L Hp + l1 L
€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ , €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ ,

2 b2 p3 a Cosh@l2D l22 2 b2 p 3 a Sinh@l2D l22

€€€€€€€€€€€€€€€€ , -b2 p 6 a + p 2 w2 >
2 b2 p 3 a l22
Hp - I l2L Hp + I l2L Hp - I l2 L Hp + I l2L Hp - I l2L Hp + I l2L
- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ , - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

Setting the determinant of coefMat equal to zero yields the following characteristic equation for the natural frequencies:

Factor@Det@coefMatDD == 0

p 2 Hb2 p 4 a - w2 L Sin@l1D Sinh@l2D Hl21 + l22 L == 0

2

There are two possibilities: b 2 p4 a - w2 = 0 and Sin@l1D = 0. Hence, either

wSol1 = SolveAb2 p 4 a - w2 == 0, wE@@2DD

• !!!
9w ® b p 2 a=

or

wSol2 = SolveA- l21 + Hn pL2 == 0 •. lRule, wE@@2DD

" ## #### ####

#
:w ® n -1 + n2 p 2 >

for n > 1. The lowest five natural frequencies are

frequencies = 8w1 -> Hw •. wSol1L, w2 -> Hw •. wSol2 •. n -> 2L,

w3 -> Hw •. wSol2 •. n -> 3L, w4 -> Hw •. wSol2 •. n -> 4L, w5 -> Hw •. wSol2 •. n -> 5L<
• !!! •!! • !!! • !!!!! •!!
9w1 ® b p 2 a , w2 ® 2 3 p 2, w3 ® 6 2 p 2, w4 ® 4 15 p 2 , w5 ® 10 6 p 2=
254 Chapter8.nb

The first two frequencies vs. the buckling level are plotted as follows:

• !!!
PlotAw •. 9wSol1 •. b a -> s, wSol2 •. n -> 2= •• Evaluate,
8s, 0, 15<, PlotStyle -> 8RGBColor@1, 0, 0D, RGBColor@0, 0, 1D<,
• !!!
Frame -> True, FrameLabel -> 9"b a ", "w"=, RotateLabel -> False,
• !!! • !! • !!
Epilog -> 9TextA"C", 92 3 , 2 3 p 2 + 5=E, TextA"b1 : 3", 92 ’ 3 , 3=E,
• !! • !!!
TextA"b1 : 1", 92 3 , 3=E, TextA"b3 : 1", 96 3 , 3=E=E;

140
120
100

w 80
60
40 C
20
b1:3 b1:1 b3: 1

• !8!
0
0 2 4 6 10 12 14
b a

where the low-frequency mode (red line) is symmetric and the second mode (blue line) is antisymmetric. As the buckling level

• !!!!
increases from zero, the frequency w1 of the first mode increases from zero and crosses the frequency w2 of the second mode

• !!!!
at point C. Hence, there are two possible three-to-one internal resonances: w2 = 3 w1 when b a = b1:3 and w1 = 3 w2 when
b 3:1. There is also a possible one-to-one internal resonance w2 = w1 when b a = b 1:1 . Whether or not these candidates for
internal resonances will be activated depends on the corresponding modeshapes. In this section, we consider the case of
one-to-one internal resonance between the first and second modes. In Section 8.1.4, we consider three-to-one internal
resonances between the first two modes. In Sections 8.1.5 and 8.1.6, we consider one-to-one and three-to-one internal
resonances between the first and third modes and the first and fouth modes, respectively.

In the next section, we consider a combination of a one-to-one internal resonance between the lowest two modes and a
principal parametric resonance of the second mode; that is,

omgList = 8w1 , w2<;

ResonanceConds = 9w2 == w1 + e2 s1 , W == 2 w2 + e2 s2 =;

where the si are detuning parameters that describe the nearness of the resonances. The buckling level for which the natural
frequencies of the first and second modes are equal (i.e., w2 = w1 ) is given by

values1 = Solve@w1 == w2 •. frequencies, bD@@1DD

• !!!
:b ® €€€€€€€€€€€€€€ >
• !!!
2 3
a

The corresponding lm are

lRuleN = Solve@ lRule •. w -> w1 •. frequencies •. values1 •. Rule -> Equal, 8 l1, l2 <D@@4DD
•!!
9l1 ® 2 p, l2 ® 3 p=
Chapter8.nb 255

In order that the influence of the damping and nonlinearity balance the influence of the forcing, we scale the damping and
forcing terms as

scaling = 9c -> 2 e2 m , f -> e2 f=;

Substituting for the natural frequency w1 back into coefMat, we determine the modeshape as

NullSpace@coefMat •. w -> w1 •. frequenciesD@@1DD.shapeFunc •. modeshape

-p 2 Sin@p xD

which we normalize as

shape1 = c Sin@p xD •. SolveAà c2 Sin@p xD2 âx == 1, cE@@2DD

1

• !!!
0

2 Sin@p xD

Substituting w2 into coefMat, we determine the modeshape as

NullSpace@coefMat •. lRuleN •. w -> w2 •. frequenciesD@@1DD.shapeFunc •. lRuleN

Sin@2 p xD

which we also normalize as

shape2 = c Sin@2 p xD •. SolveAà c2 Sin@2 p xD2 âx == 1, cE@@2DD

1

• !!!
0

2 Sin@2 p xD

In a similar fashion, we find that the next three normalized modeshapes are

• !!!
shape3 = 2 Sin@3 p xD;

• !!!
shape4 = 2 Sin@4 p xD;

• !!!
shape5 = 2 Sin@5 p xD;

To treat a one-to-one internal resonance between the first and second modes, we define fk and the first two eigenmodes of
the buckled beam as

modeshapes = Join@modeshape,
8F1 -> Function@x, shape1 •• EvaluateD, F2 -> Function@x, shape2 •• EvaluateD<D;

Ÿ 8.1.2 Perturbation Analysis

We use the method of multiple scales to directly attack EOM and BC1. To transform the time derivatives in EOM in terms of
the scales T0, T1 , and T2, we let
256 Chapter8.nb

multiScales = 8u@x_, tD -> u@x, T0 , T1, T2 D,

Derivative@m_, n_D@u_D@x_, tD -> dt@nD@D@u@x, T0, T1 , T2 D, 8x, m<DD, t -> T0<;

Then, we seek a second-order approximate solution in the form

solRule = u -> IEvaluateASumAej uj @#1, #2, #3, #4D, 8j, 3<EE &M

u ® He u1 @#1, #2, #3, #4D + e2 u2@#1, #2, #3, #4D + e3 u3@#1, #2, #3, #4D &L

where #1 stands for x and #2, #3, and #4 stand for T0 , T1, and T2 , respectively.

Substituting multiScales , solRule, and scaling into EOM and BC1, expanding the result for small e, and discarding terms of
order higher than e3, we obtain

eq81c = HJoin@8EOM<, BC1D •. multiScales •. solRule •. scaling •. intRule1 ••. intRule2 ••

TrigToExp •• ExpandAllL •. en_•;n>3 -> 0;

Equating coefficients of like powers of e, we obtain

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq81c •• Transpose •• Rest;

eqEps •. displayRule

98Pk Hu''
1 L + u1 + D20u1 - 2 b2 a int@Hu'1 L f¢k @xD, 8x, 0, 1<D f²k @xD == 0,
''''

u1 == 0, u'' ''
1 == 0, u1 == 0, u1 == 0<,

9-2 b a int@Hu'1L f¢k @xD, 8x, 0, 1<D Hu'' 1 L + Pk Hu2 L + u2 + D20 u2 + 2 HD0 D1 u1L - €€€€ b E-I T0 W f f²k @xD -
'' '''' 1
2
f fk @xD - b a intBHu1L , 8x, 0, 1<F fk @xD -
1 I T0 W ² ' 2 ²
€€€€ b E

2 b2 a int@Hu'2 L f¢k @xD, 8x, 0, 1<D f²k @xD == 0, u2 == 0, u''

2
''
2 == 0, u2 == 0, u2 == 0=,

9- €€€€ E-I T0 W f Hu''

1 L - €€€€ E f Hu''
1 L - a intBHu1 L , 8x, 0, 1<F Hu1 L -
1 1 I T0 W ' 2 ''

2 b a int@Hu'2L f¢k @xD, 8x, 0, 1<D Hu'' 1 L - 2 b a int@Hu1L fk @xD, 8x, 0, 1<D Hu2 L + Pk Hu3 L +
2 2
' ¢ '' ''

u'''' + 2 m HD0u1 L + D20 u3 + 2 HD0 D1 u2 L + D21 u1 + 2 HD0 D2 u1L - 2 b a int@Hu'1 L Hu'2L, 8x, 0, 1<D f²k @xD -
2 b a int@Hu'3 L f¢k @xD, 8x, 0, 1<D f²k @xD == 0, u3 == 0, u''
3
2 ''
3 == 0, u3 == 0, u3 == 0==

Ÿ First-Order Solution

Because in the presence of damping all of the modes that are not directly or indirectly excited decay with time, the solution of
eqEps[[1]] is taken to consist of the two modes that might be involved in the internal resonance; that is,

sol1 = 9u1 -> FunctionA8x, T0 , T1 , T2<,

SumAAi @T1 , T2 D Fi@xD Exp@I wi T0 D + Ai @T1 , T2D Fi @xD Exp@-I wi T0D, 8i, 2<E •• EvaluateE=;
•

where the Fi are the eigenmodes, which satisfy the orthonormality condition Ù0 Fn @xD Fm @xDâ x = dn,m .
1
Chapter8.nb 257

Ÿ Second-Order Solution

Substituting sol1 into eqEps[[2,1]] and using intRule1 and intRule2 yields

order2Eq =
HeqEps@@1, 1, 1DD •. u1 -> u2L == HeqEps@@1, 1, 1DD •. u1 -> u2L - eqEps@@2, 1, 1DD •. sol1 •.
intRule1 ••. intRule2 •• ExpandAll; order2Eq •. displayRule

Pk Hu'' 2 L + u2
''''
+ D20u2 - 2 b2 a int@Hu'2 L f¢k @xD, 8x, 0, 1<D f²k @xD ==
-2 I EI T0 w1 HD1 A1L w1 F1@xD + 2 I E-IT0 w1 HD1 A1L w1 F1 @xD -
•

2 I EI T0 w2 HD1 A2 L w2 F2 @xD + 2 I E-I T0 w2 HD1 A2L w2 F2@xD + €€€€ b E-I T0 W f f²k @xD +
• 1
2

€€€€ b EI T0 W f f²k @xD + b E2 I T0 w1 a int@F¢1 @xD2, 8x, 0, 1<D A21 f²k @xD +
1

2 b EI T0 Hw1 +w2L a int@F¢1@xD F¢2 @xD, 8x, 0, 1<D A1 A2 f²k @xD +

2

b E2 I T0 w2 a int@F¢2 @xD2, 8x, 0, 1<D A22 f²k @xD + 2 b a int@F¢1@xD2 , 8x, 0, 1<D A1 A1 f²k @xD +
•

2 b EI T0 H-w1 +w2L a int@F¢1 @xD F¢2@xD, 8x, 0, 1<D A2 A1 f²k @xD +

•

b E-2 I T0 w1 a int@F¢1 @xD2, 8x, 0, 1<D A1 f²k @xD +

•2

2 b EI T0 Hw1 -w2L a int@F¢1@xD F¢2 @xD, 8x, 0, 1<D A1 A2 f²k @xD + 2 b a int@F¢2 @xD2 , 8x, 0, 1<D A2 A2 f²k @xD +
• •

2 b EI T0 H-w1 -w2L a int@F¢1 @xD F¢2@xD, 8x, 0, 1<D A1 A2 f²k @xD +

• •

b E-2 I T0 w2 a int@F¢2 @xD2, 8x, 0, 1<D A2 f²k @xD + 2 b E2 I T0 w1 a int@f¢k @xD F¢1@xD, 8x, 0, 1<D A21 F²1@xD +
•2

2 b EI T0 Hw1 +w2L a int@f¢k @xD F¢2 @xD, 8x, 0, 1<D A1 A2 F²1@xD +

4 b a int@f¢k @xD F¢1 @xD, 8x, 0, 1<D A1 A1 F²1 @xD +
•

2 b EI T0 H-w1 +w2L a int@f¢k @xD F¢2 @xD, 8x, 0, 1<D A2 A1 F²1@xD +

•

2 b E-2 I T0 w1 a int@f¢k @xD F¢1@xD, 8x, 0, 1<D A1 F²1@xD +

•2

2 b EI T0 Hw1 -w2L a int@f¢k @xD F¢2 @xD, 8x, 0, 1<D A1 A2 F²1@xD +

•

2 b EI T0 H-w1 -w2L a int@f¢k @xD F¢2 @xD, 8x, 0, 1<D A1 A2 F²1@xD +

• •

2 b EI T0 Hw1 +w2L a int@f¢k @xD F¢1 @xD, 8x, 0, 1<D A1 A2 F²2@xD +

2 b E2 I T0 w2 a int@f¢k @xD F¢2@xD, 8x, 0, 1<D A22 F²2@xD +
2 b EI T0 H-w1 +w2L a int@f¢k @xD F¢1 @xD, 8x, 0, 1<D A2 A1 F²2@xD +
•

2 b EI T0 Hw1 -w2L a int@f¢k @xD F¢1 @xD, 8x, 0, 1<D A1 A2 F²2@xD +

•

4 b a int@f¢k @xD F¢2 @xD, 8x, 0, 1<D A2 A2 F²2 @xD +

•

2 b EI T0 H-w1 -w2L a int@f¢k @xD F¢1 @xD, 8x, 0, 1<D A1 A2 F²2@xD +

• •

2 b E-2 I T0 w2 a int@f¢k @xD F¢2@xD, 8x, 0, 1<D A2 F²2@xD

•2

To collect the terms that might lead to secular terms from the right-hand side of order2Eq, we define the rules:

OmgRule = Solve@ResonanceConds, Complement@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž omgList

88w2 ® e2 s1 + w1 , W ® 2 e2 s1 + e2 s2 + 2 w1 <, 8w1 ® -e2 s1 + w2, W ® e2 s2 + 2 w2<<

expRule1@i_D := Exp@arg_D :> ExpAExpand@arg •. OmgRule@@iDDD •. e2 T0 -> T2E

Collecting the terms that may lead to secular terms, the terms proportional to EI wi T0 , we have
258 Chapter8.nb

ST1 = CoefficientAorder2Eq@@2DD •. expRule1@#D, EI w# T0 E & •ž 81, 2<;

ST1 •. displayRule

8-2 I HD1A1 L w1 F1 @xD - 2 I EI T2 s1 HD1A2 L w2 F2 @xD, -2 I E-I T2 s1 HD1A1 L w1 F1 @xD - 2 I HD1 A2L w2 F2@xD<

Because the homogeneous parts of order2Eq and corresponding boundary conditions have a nontrivial solution, the
corresponding nonhomogeneous problem has a solution only if solvability conditions are satisfied. These conditions

adjoint and hence solutions of the adjoint problem are given by the Fi @xD. Demanding that ST1[[i]] be orthogonal to the Fi@xD,
demand that ST1 be orthogonal to every solution of the adjoint homogeneous problem. In this case, the problem is self-

we obtain the solvability conditions

SCond1 = TableAà ST1@@kDD F k @xD âx == 0, 8k, 2<E •. modeshapes

1

9-2 I @T1 , T2 D == 0, -2 I w2 A2 @T1 , T2D == 0=

H1,0L H1,0L
w1 A1

H1,0L H1,0L
SCond1Rule = SolveASCond1, 9A1 @T1 , T2D, A2 @T1, T2 D=E@@1DD

9AH1,0L
1 @T1 , T2 D ® 0, AH1,0L
2 @T1 , T2D ® 0=

whose complex conjugate is

ccSCond1Rule = SCond1Rule •. conjugateRule

•H1,0L •H1,0L
:A1 @T1 , T2 D ® 0, A2 @T1 , T2D ® 0>

Substituting SCond1Rule and ccSCond1Rule into order2Eq and using modeshapes, we have

order2Eqm = order2Eq •. SCond1Rule •. ccSCond1Rule •. modeshapes •. int -> Integrate •.

Integrate -> int;
order2Eqm •. displayRule

Pk Hu'' 2 L + u2
''''
+ D20u2 + 2 b2 p 3 a int@Cos@p xD Hu'2L, 8x, 0, 1<D Sin@p xD ==
1 1
- €€€€ b E-I T0 W f p 2 Sin@p xD - €€€€ b EIT0 W f p 2 Sin@p xD - 3 b E2 I T0 w1 p 4 a Sin@p xD A21 -

8 b EI T0 Hw1 +w2L p 4 a Sin@2 p xD A1 A2 - 4 b E2 I T0 w2 p 4 a Sin@p xD A22 -

2 2

6 b p 4 a Sin@p xD A1 A1 - 8 b EIT0 H-w1+w2 L p 4 a Sin@2 p xD A2 A1 -

• •

3 b E-2 I T0 w1 p 4 a Sin@p xD A1 - 8 b EI T0 Hw1-w2 L p 4 a Sin@2 p xD A1 A2 -

•2 •

8 b p 4 a Sin@p xD A2 A2 - 8 b EIT0 H-w1-w2 L p 4 a Sin@2 p xD A1 A2 - 4 b E-2 IT0 w2 p 4 a Sin@p xD A2

• • • •2

The associated boundary conditions are

order2BC = eqEps@@2DD •• Rest

9u2 @0, T0, T1, T2 D == 0, u2 @0, T0 , T1, T2 D == 0,

H2,0,0,0L

u2 @1, T0, T1, T2 D == 0, @1, T0 , T1, T2 D == 0=

H2,0,0,0L
u2

We use the method of undetermined coefficients to determine a particular solution of order2Eqm and order2BC. To accom-
plish this, we first determine the forms of the nonhomogeneous terms (i.e., the terms on the right-hand side of order2Eqm) as
Chapter8.nb 259

rhsTerms = Cases@order2Eqm@@2DD, #D & •ž 9a_ f_@b_ xD :> f@b xD •; FreeQ@a, T0D,

_ Ea_ T0+b_. f_@c_ xD -> Ea T0 +b f@c xD= •• Flatten •• Union

8Sin@p xD, E-I T0 W Sin@p xD, EI T0 W Sin@p xD, E-2 I T0 w1 Sin@p xD,
E2 I T0 w1 Sin@p xD, E-2 I T0 w2 Sin@p xD, E2 IT0 w2 Sin@p xD, E-I T0 w1 -I T0 w2 Sin@2 p xD,
EI T0 w1 -I T0 w2 Sin@2 p xD, E-I T0 w1+I T0 w2 Sin@2 p xD, EI T0 w1+I T0 w2 Sin@2 p xD<

which all satisfy the order2BC. Since we have only even spatial derivatives on the left-hand side of order2Eqm, we seek a
particular solution as a linear combination of the rhsTerms as follows:

symbolList = Table@ai, 8i, Length@rhsTermsD<D

8a1 , a2, a3, a4 , a5 , a6, a7 , a8 , a9, a10, a11 <

sol2Form = u2 -> Function@8x, T0 , T1 , T2<, symbolList.rhsTerms •• EvaluateD;

Substituting sol2Form into order2Eqm, equating coefficients of like terms, and solving for the symbolList, we have

symbolRule =
Solve@Coefficient@Subtract žž order2Eqm •. sol2Form •. intRule1 ••. intRule2 •. int ->
Integrate, rhsTermsD == 0 •. Exp@_ T0 + _.D -> 0 •• Thread, symbolListD@@1DD

2 H3 b p 2 a A1 @T1, T2 D A1@T1 , T2 D + 4 b p 2 a A2 @T1 , T2D A2 @T1, T2 DL

:a1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€ ,
p 2 - Pk + b2 p 2 a
b f p2 b f p2
2 Hp 4 - p 2 Pk + b2 p 4 a - W2 L 2 Hp 4 - p 2 Pk + b2 p4 a - W2L
a2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a3 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

3 b p 4 a A1 @T1, T2D 3 b p 4 a A1@T1 , T2 D2

• 2
a4 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ , a5 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ ,
p 4 - p 2 Pk + b2 p 4 a - 4 w21 p 4 - p2 Pk + b2 p 4 a - 4 w21

4 b p 4 a A2 @T1, T2D 4 b p 4 a A2@T1 , T2 D2

• 2
a6 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ , a7 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ ,
p 4 - p 2 Pk + b2 p 4 a - 4 w22 p 4 - p2 Pk + b2 p 4 a - 4 w22
8 b p 4 a A1@T1 , T2 D A2@T1 , T2D 8 b p 4 a A1 @T1, T2 D A2@T1 , T2 D
• • •
a8 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a9 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
16 p 4 - 4 p 2 Pk - w21 - 2 w1 w2 - w22 16 p 4 - 4 p 2 Pk - w21 + 2 w1 w2 - w22
8 b p 4 a A2@T1 , T2 D A1 @T1 , T2D 8 b p 4 a A1 @T1, T2 D A2@T1 , T2 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
•
a10 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a11 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
16 p 4 - 4 p 2 Pk - w21 + 2 w1 w2 - w22 16 p 4 - 4 p 2 Pk - w21 - 2 w1 w2 - w22

Substituting for the parameter values in symbolRule yields

260 Chapter8.nb

symbolRuleN = symbolRule •. W -> 2 w2 •. frequencies •. bucklingLoad •. values1

• !!! • !!! •!! •!!
6 3 p2 a A1 @T1 , T2D A1 @T1, T2 D + 8 3 p 2 a A2 @T1 , T2 D A2@T1 , T2D
:a1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• •

• !!! €€€€€€€€ •€€€€!!!

f
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€ , a2 ® €€€€€€€€€€€€€€€€ €€€€ ,
6 p2
• !!! •!a! A @T , T D2 2 • !!!
12 3 p 2 a

a A1 @T1, T2 D a A2@T1 , T2D

• 2 • 2

•!! •!! •!! • !!! • !!!

f 1 1 2
a3 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a4 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a5 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€ ,

•!! • !!! • •! !
12 3 p 2 a 2 3 2 3 3 3
2 a A2@T1 , T2 D2 4 a A1 @T1 , T2D A2 @T1, T2 D 4 a A1 @T1, T2 D A2@T1 , T2 D
• •

• !!! • !!! • !!!

a7 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a8 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a9 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

• !!! • !!!
3 3 3 3 3
4 a A2 @T1 , T2D A1 @T1, T2 D 4 a A1 @T1, T2 D A2@T1 , T2 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
•

• !!! •!!
a10 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , a11 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 3 3

Substituting symbolRuleN into sol2Form yields

sol2 = sol2Form •. symbolRuleN

u2 ® FunctionB8x, T0, T1 , T2 <,

Sin@p xD I-I6 • !!! 3 p 2 • !!! a A1@T1 , T2D A1 @T1, T2 D + 8 •!3! p 2 •!a! A2 @T1, T2 D A2@T1 , T2 DMM
• •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

•!! •
6 p2
E-2 I T0 w1 Sin@p xD J a A1@T1 , T2 D N
2

• !!! • !!! •!! •!! • !!!

E-I T0 W Sin@p xD f EI T0 W Sin@p xD f
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ +

•!! •
12 3 p 2 a 12 3 p 2 a
• !!!
2 3

E2 I T0 w1 Sin@p xD I a A1 @T1, T2D2 M E-2 I T0 w2 Sin@p xD J2 a A2@T1 , T2 D N

2

2 • !!! 3 • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ +

E2 I T0 w2 Sin@p xD I2 • !!! E-I T0 w1-I T0 w2 Sin@2 p xD I4 • !!!

3 3

a A2 @T1, T2 D2 M a A1@T1 , T2D A2 @T1, T2 DM

• •

• !!! • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

•!!
3 3 3 3
EI T0 w1 -IT0 w2 Sin@2 p xD I-4 a A1@T1 , T2 D A2 @T1 , T2DM
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

• !!!
3
E-I T0 w1 +I T0 w2 Sin@2 p xD I-4 a A2@T1 , T2D A1 @T1, T2DM
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

• !!!
3

EI T0 w1 +IT0 w2 Sin@2 p xD I4 a A1 @T1, T2 D A2@T1 , T2 DM

€€€€€€€€€€€€€€€€€€€€€€€€€ F
• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 3

Ÿ Solvability Conditions

Substituting sol1 and sol2 into eqEps[[3,1]] and using intRule1 and intRule2 yields

order3Eq = HeqEps@@1, 1, 1DD •. u1 -> u3L ==

HHeqEps@@1, 1, 1DD •. u1 -> u3L - eqEps@@3, 1, 1DD •. sol1 •. sol2 •. intRule1 ••. intRule2 •.
modeshapes •. int -> Integrate •• ExpandL; •• Timing

810.475 Second, Null<

Chapter8.nb 261

Collecting the terms that may lead to secular terms, the terms proportional to EI wi T0 , we have
262 Chapter8.nb

ST2 = CoefficientAorder3Eq@@2DD •. expRule1@#D, EI w# T0 E & •ž 81, 2<;

ST2 •. displayRule

:- • !!!
2 HD21 A1 L Sin@p xD - • !!! 2 EI T2 s1 HD21A2 L Sin@2 p xD -

2 I • !!! 2 HD2 A1 L Sin@p xD w1 - 2 I • !!! 2 m Sin@p xD A1 w1 - 2 I • !!! 2 EI T2 s1 HD2A2 L Sin@2 p xD w2 -

• !!!
•

• !!!
IT2 s1 E2 I T2 s1 +I T2 s2 f p 2 Sin@p xD A1
2I 2 E m Sin@2 p xD A2 w2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

•!!
2

€€€€€€€€€€€€€€€€€ - 3 • !!!
2 p 4 a Sin@p xD A21 A1 + 5 $ %€€€%%% € % b p 4 a3•2 Sin@p xD A21 A1 -
•

• !!!
b E2 IT2 s1+I T2 s2 f p 2 a Sin@p xD A1 • 3 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 6 2

• !!! IT2 s1 4
p a Sin@2 p xD A1 A2 A1 + €€€€€€€ $ %€€€%%%
€ % b EI T2 s1 p 4 a3•2 Sin@2 p xD A1 A2 A1 -
• 68 2 •
8 2 E
3 3

• !!! 2I T2 s1 4
p a Sin@p xD A22 A1 + 14 $ %%%%
€€€€ % b E2 I T2 s1 p 4 a3•2 Sin@p xD A22 A1 -
• 2 •
4 2 E
3

2 • !!!
2 EIT2 s1+I T2 s2 f p 2 Sin@2 p xD A2 - €€€€ $ %%€€€€%%% b EI T2 s1 +IT2 s2 f p 2 • !a! Sin@2 p xD A2 -
• 1 2 •
3 3

4 • !!!
2 E-IT2 s1 p 4 a Sin@2 p xD A21 A2 + 14 $ %€€€€
%%%% b E-I T2 s1 p 4 a3•2 Sin@2 p xD A21 A2 -
• 2 •
3

• !!! 4
2 p a Sin@p xD A1 A2 A2 + €€€€€€ $ %%%
€€€€ % b p 4 a3•2 Sin@p xD A1 A2 A2 -
• 68 2 •
8
3 3

• !!! IT2 s1 4
p a Sin@2 p xD A22 A2 + €€€€€€ $ %%% €€€€ % b EI T2 s1 p 4 a3•2 Sin@2 p xD A22 A2,
• 40 2 •
48 2 E

• !!! • !!! • !!!

- 2 E-I T2 s1 HD21A1 L Sin@p xD - 2 HD21A2 L Sin@2 p xD - 2 I 2 E-I T2 s1 HD2A1 L Sin@p xD w1 -
3 3

• !!! • !!! • !!!

2 I 2 E-I T2 s1 m Sin@p xD A1 w1 - 2 I 2 HD2 A2 L Sin@2 p xD w2 - 2 I 2 m Sin@2 p xD A2 w2 -
• • !!! •

• !2! 2 • !!!
EI T2 s1 +IT2 s2 f p 2 Sin@p xD A1 b EI T2 s1+I T2 s2 f p 2 a Sin@p xD A1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ -
6

• !!! -IT2 s1 4
p a Sin@p xD A21 A1 + 5 $ %%%%
€€€€ % b E-I T2 s1 p 4 a3•2 Sin@p xD A21 A1 -
• 3 •
3 2 E
2

• !!! 4
2 p a Sin@2 p xD A1 A2 A1 + €€€€€€ $ %%%
€€€€ % b p 4 a3•2 Sin@2 p xD A1 A2 A1 -
• 68 2 •
8
3 3

• !!! IT2 s1 4
p a Sin@p xD A22 A1 + 14 $ %%%%
€€€€ % b EI T2 s1 p 4 a3•2 Sin@p xD A22 A1 -
• 2 •
4 2 E
3

2 • !!! € % b EI T2 s2 f p 2 • !!!
2 EIT2 s2 f p 2 Sin@2 p xD A2 - €€€€ $ %€€€%%%
• 1 2 •
a Sin@2 p xD A2 -
3 3

4 • !!!
2 E-2I T2 s1 p 4 a Sin@2 p xD A21 A2 + 14 $ %€€€€
%%%% b E-2 I T2 s1 p 4 a3•2 Sin@2 p xD A21 A2 -
• 2 •
3

• !!! -IT2 s1 4
p a Sin@p xD A1 A2 A2 + €€€€€€€ $ %€€€%%%
€ % b E-I T2 s1 p 4 a3•2 Sin@p xD A1 A2 A2 -
• 68 2 •
8 2 E
3 3

• !!! 4
2 p a Sin@2 p xD A22 A2 + €€€€€€€ $ %€€€%%%
€ % b p 4 a3•2 Sin@2 p xD A22 A2>
• 40 2 •
48
3 3
Chapter8.nb 263

Demanding that ST2[[i]] be orthogonal to the Fi @xD, we obtain the solvability conditions

SCond2 =
Table@int@ST2@@kDD F k@xD, 8x, 0, 1<D == 0, 8k, 2<D •. intRule1 ••. intRule2 •. modeshapes •.
values1 •. int -> Integrate •. D@SCond1Rule, T1 D;
SCond2 •. displayRule

9-2 I HD2A1 L w1 - 2 I m A1 w1 - E2 I T2 s1 +I T2 s2 f p 2 A1 + 12 p 4 a A21 A1 +

• •

24 E2 I T2 s1 p 4 a A22 A1 + €€€€€€€€€ p4 a A1 A2 A2 == 0, -2 I HD2 A2L w2 - 2 I m A2 w2 +

• 112 •
3
112 • 8 • • 64 •
€€€€€€€€€ p 4 a A1 A2 A1 - €€€€ EI T2 s2 f p 2 A2 + 24 E-2 IT2 s1 p 4 a A21 A2 - €€€€€€€ p 4 a A22 A2 == 0=
3 3 3

Ÿ 8.1.3 The Function MMSDirect11

Collecting the commands in the preceding section, we build a function MMSDirect11, specifically for the nonlinear trans-
verse vibrations of a hinged-hinged beam around its first buckled mode (EOM and BC1), to determine the solvability condi-
tions for different combinations of frequencies, modeshapes, and ResonanceConds in which the first mode might be involved
in either a one-to-one or a three-to-one internal resonance with one of the other modes.

MMSDirect11@frequencies_, modeshapes_, ResonanceConds_D :=

ModuleA8<,
multiScales = 8u@x_, tD -> u@x, T0, T1 , T2 D,
Derivative@m_, n_D@u_D@x_, tD -> dt@nD@D@u@x, T0, T1 , T2 D, 8x, m<DD, t -> T0<;
solRule = u -> IEvaluateASumAej uj@#1, #2, #3, #4D, 8j, 3<EE &M;
scaling = 9c -> 2 e2 m , f -> e2 f=;
eqs = HJoin@8EOM<, BC1D •. multiScales •. solRule •. scaling •. intRule1 ••. intRule2 ••
TrigToExp •• ExpandAllL •. en_•;n>3 -> 0;
eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eqs •• Transpose •• Rest;
sol1 = 9u1 -> FunctionA8x, T0 , T1 , T2<,
SumAAi@T1 , T2 D Fi@xD Exp@I wi T0 D + Ai @T1 , T2D Fi @xD Exp@-I wi T0D, 8i, 2<E •• EvaluateE=;
•

order2Eq = HeqEps@@1, 1, 1DD •. u1 -> u2 L == HeqEps@@1, 1, 1DD •. u1 -> u2 L - eqEps@@2, 1, 1DD •.

sol1 •. intRule1 ••. intRule2 •• ExpandAll;
omgList = 8w1, w2<;
OmgRule =
Solve@ResonanceConds, Complement@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž omgList;
expRule1@i_D := Exp@arg_D :> ExpAExpand@arg •. OmgRule@@iDDD •. e2 T0 -> T2E;
ST1 = CoefficientAorder2Eq@@2DD •. expRule1@#D, EI w# T0 E & •ž 81, 2<;

SCond1 = TableAà ST1@@kDD F k @xD â x == 0, 8k, 2<E •. modeshapes;

1

H1,0L H1,0L
SCond1Rule = SolveASCond1, 9A1 @T1, T2 D, A2 @T1 , T2 D=E@@1DD;
0

ccSCond1Rule = SCond1Rule •. conjugateRule;

order2Eqm = order2Eq •. SCond1Rule •. ccSCond1Rule •. modeshapes •. int -> Integrate •.
Integrate -> int;
264 Chapter8.nb

rhsTerms = Cases@order2Eqm@@2DD, #D & •ž 9a_ f_@b_ xD :> f@b xD •; FreeQ@a, T0D,

_ Ea_ T0+b_. f_@c_ xD -> Ea T0 +b f@c xD= •• Flatten •• Union;
symbolList = Table@ai, 8i, Length@rhsTermsD<D;
sol2Form = u2 -> Function@8x, T0 , T1, T2<, symbolList.rhsTerms •• EvaluateD;
symbolRule =
Solve@Coefficient@Subtract žž order2Eqm •. sol2Form •. intRule1 ••. intRule2 •.
int -> Integrate, rhsTermsD == 0 •. Exp@_ T0 + _.D -> 0 •• Thread, symbolListD@@1DD;
values1 = Solve@Select@ResonanceConds, FreeQ@#, WD &D •. e -> 0 •. frequencies, bD ••

symbolRuleN = symbolRule •. W -> 2 w2 •. frequencies •. bucklingLoad •. values1;

Flatten;

sol2 = sol2Form •. symbolRuleN;

order3Eq = HeqEps@@1, 1, 1DD •. u1 -> u3 L ==
HHeqEps@@1, 1, 1DD •. u1 -> u3L - eqEps@@3, 1, 1DD •. sol1 •. sol2 •. intRule1 ••. intRule2 •.
modeshapes •. int -> Integrate •• ExpandL;
ST2 = CoefficientAorder3Eq@@2DD •. expRule1@#D, EI w# T0 E & •ž 81, 2<;
SCond2 =
Table@int@ST2@@kDD F k@xD, 8x, 0, 1<D == 0, 8k, 2<D •. intRule1 ••. intRule2 •. modeshapes •.
values1 •. int -> Integrate •. D@SCond1Rule, T1 D;
Return@SCond2 •. Ai_ -> HAi @#2D &L •. f_@T1, T2D -> f@T2 DD
E

Ÿ 8.1.4 Three-to-One Internal Resonances Between the First Two Modes

In this section, we consider the case w2 » 3 w1 and W » 2 w2 . To describe the nearness of the resonances, we introduce the
detuning parameters si defined by

ResonanceConds1 = 9w2 == 3 w1 + e2 s1, W == 2 w2 + e2 s2 =;

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies, modeshapes, ResonanceConds1D •• Timing

921.501 Second,

9-2 I m w1 A1 @T2D + 12 p 4 a A1 @T2D2 A1@T2 D + €€€€€€€€€ p 4 a A1@T2 D A2@T2 D A2@T2 D - 2 I w1 A¢1@T2 D == 0,

• 624 •
35
-2 I m w2 A2 @T2D + €€€€€€€€€ p a A1 @T2D A2 @T2D A1 @T2D -
624 4 •
35
f p A2 @T2D - €€€€€€€€€ p 4 a A2 @T2D2 A2@T2 D - 2 I w2 A¢2@T2 D == 0==
2•
72 I T2 s2 576 •
€€€€€€€ E
35 35

which indicates that this three-to-one internal resonance is not activated.

Next, we consider the case in which the frequency of the first mode is three times that of the second mode; that is, the
buckling load is above the first crossover value. To accomplish this, we redefine the frequencies and denote the frequency of
the first mode by w2 and that of the second mode by w1 as follows:
Chapter8.nb 265

frequencies2 = 8w2 -> Hw •. wSol1L, w1 -> Hw •. wSol2 •. n -> 2L<

• !!! •!!
9w2 ® b p 2 a , w1 ® 2 3 p 2=

Moreover, we redefine the modeshapes according to

modeshapes2 = Join@modeshape,
8F1 -> Function@x, shape2 •• EvaluateD, F2 -> Function@x, shape1 •• EvaluateD<D;

To describe the nearness of the resonances, we introduce the two detuning parameters si defined by

ResonanceConds2 = 9w2 == 3 w1 + e2 s1, W == 2 w2 + e2 s2 =;

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies2, modeshapes2, ResonanceConds2D •• Timing

921.721 Second,

9-2 I m w1 A1 @T2D + €€€€€€ p 4 a A1 @T2D2 A1@T2 D - €€€€€€€€€ p 4 a A1 @T2 D A2 @T2 D A2 @T2 D - 2 I w1 A¢1@T2 D == 0,
64 • 496 •
5 5
-2 I m w2 A2 @T2D - €€€€€€€€€ p a A1 @T2D A2 @T2D A1 @T2D -
496 4 •

f p A2 @T2D + 12 p4 a A2@T2 D2 A2 @T2 D - 2 I w2 A¢2 @T2 D == 0==

5
I T2 s2 2• •
E

Again, this three-to-one internal resonance is not activated.

Ÿ 8.1.5 One-to-One and Three-to-One Internal Resonances Between the First and Third Modes

Variations of the first and third natural frequencies with the buckling level are plotted as follows:
300

250

200

w 150

100 C

50

b1:3 b1:1 b3:1

b•!!!
0
0 5 10 15 20 25 30
a

• !!!!
• !!!!
It follows from this figure that the frequencies of the first and third modes are equal when b a = b1:1 , the frequency of the

• !!!!
third mode is three times that of the first mode when b a = b 1:3 , and the frequency of the first mode is three times that of
the third mode when b a = b 3:1 . Next, we use the function MMSDirect11 to treat these three cases.
266 Chapter8.nb

Ÿ One-to-One Internal Resonance

We denote the frequency of the third mode by w2 and that of the first mode by w1 as follows:

frequencies3 = 8w1 -> Hw •. wSol1L, w2 -> Hw •. wSol2 •. n -> 3L<

9w1 ® b p 2 • !!!
a , w2 ® 6 • !2! p 2=

Moreover, we define fk and the first and third eigenmodes of the buckled beam as

modeshapes3 = Join@modeshape,
8F1 -> Function@x, shape1 •• EvaluateD, F2 -> Function@x, shape3 •• EvaluateD<D;

We consider a combination of a one-to-one internal resonance between the first and third modes and a principal parametric
resonance of the third mode; that is,

ResonanceConds3 = 9w2 == w1 + e2 s1 , W == 2 w2 + e2 s2 =;

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies3, modeshapes3, ResonanceConds3D •• Timing

826.809 Second,
8-2 I m w1 A1 @T2D - E2 I T2 s1 +I T2 s2 f p 2 A1 @T2D + 12 p 4 a A1 @T2D2 A1@T2 D + 144 E2 I T2 s1 p 4 a A2 @T2D2 A1@T2 D +
• • •

144 p 4 a A1 @T2D A2 @T2D A2 @T2D - 2 I w1 A¢1 @T2D == 0,

•

-2 I m w2 A2 @T2D + 144 p 4 a A1 @T2D A2 @T2D A1 @T2D - 6 EIT2 s2 f p 2 A2 @T2 D +

• •

144 E-2 I T2 s1 p 4 a A1@T2 D2 A2 @T2D - 108 p 4 a A2 @T2D2 A2@T2 D - 2 I w2 A¢2@T2 D == 0<<

• •

which indicates that the one-to-one internal resonance between the first and third modes is activated.

Ÿ Three-to-One Internal Resonances

To treat the case in which the natural frequency of the third mode is approximately three times that of the first mode using the
function MMSDirect11, we denote these frequencies by w2 and w1 , respectively, according to

frequencies4 = 8w1 -> Hw •. wSol1L, w2 -> Hw •. wSol2 •. n -> 3L<

• !!! •!!
9w1 ® b p 2 a , w2 ® 6 2 p 2=

Moreover, we define fk and the first and third eigenmodes of the buckled beam as

modeshapes4 = Join@modeshape,
8F1 -> Function@x, shape1 •• EvaluateD, F2 -> Function@x, shape3 •• EvaluateD<D;

We consider a combination of a three-to-one internal resonance between the first and third modes and a principal parametric
resonance of the third mode; that is,

ResonanceConds4 = 9w2 == 3 w1 + e2 s1, W == 2 w2 + e2 s2 =;

Chapter8.nb 267

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies4, modeshapes4, ResonanceConds4D •• Timing

921.651 Second,

9-2 I m w1 A1 @T2D + 12 p 4 a A1 @T2D2 A1@T2 D + €€€€€€€€€€€€ p 4 a A1@T2 D A2@T2 D A2@T2 D - 2 I w1 A¢1@T2 D == 0,

• 1584 •
35

-2 I m w2 A2 @T2D + €€€€€€€€€€€€ p a A1 @T2D A2 @T2D A1 @T2D -

1584 4 •
35
f p 2 A2 @T2D - €€€€€€€€€€€€ p 4 a A2 @T2D2 A2@T2 D - 2 I w2 A¢2@T2 D == 0==
162 I T2 s2 • 2916 •
€€€€€€€€€ E
35 35

which indicates that this three-to-one internal resonance between the first and third modes is not activated.

When the frequency of first mode is approximately three times that of the third mode, we define the frequencies as

frequencies5 = 8w2 -> Hw •. wSol1L, w1 -> Hw •. wSol2 •. n -> 3L<

9w2 ® b p 2 • !!!
a , w1 ® 6 • !2! p 2=

Moreover, we define fk and the third and first eigenmodes of the buckled beam as

modeshapes5 = Join@modeshape,
8F1 -> Function@x, shape3 •• EvaluateD, F2 -> Function@x, shape1 •• EvaluateD<D;

We consider a combination of a three-to-one internal resonance between the first and third modes and a principal parametric
resonance of the third mode; that is,

ResonanceConds5 = 9w2 == 3 w1 + e2 s1, W == 2 w2 + e2 s2 =;

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies5, modeshapes5, ResonanceConds5D •• Timing

921.712 Second,

9-2 I m w1 A1 @T2D + €€€€€€€€€ p 4 a A1 @T2D2 A1@T2 D - €€€€€€€€€€€€ p4 a A1@T2 D A2@T2 D A2@T2 D - 2 I w1 A¢1@T2 D == 0,
324 • 2736 •
5 5
-2 I m w2 A2 @T2D - €€€€€€€€€€€€ p a A1 @T2D A2 @T2D A1 @T2D -
2736 4 •

EI T2 s2 f p2 A2 @T2D + 12 p4 a A2@T2 D2 A2 @T2 D - 2 I w2 A¢2 @T2 D == 0==

5
• •

which indicates that this three-to-one internal resonance between the first and third modes is not activated.

Ÿ 8.1.6 One-to-One and Three-to-One Internal Resonances Between the First and Fourth Modes

Variations of the first and fourth natural frequencies with the buckling level are plotted as follows:
268 Chapter8.nb

500

400

300
w
200
C

100

b1:3 b1:1 b3:1

b•!!!
0
0 10 20 30 40 50
a

Again, there exist buckling levels for which the frequencies of the first and fourth modes are in the ratio of 1:1, 3:1, and 1:3.
We use the function MMSDirect11 to treat these cases.

Ÿ One-to-One Internal Resonance

When the frequencies of the first and fourth modes are approximately equal, we denote them by w1 and w2, respectively,
according to

frequencies6 = 8w1 -> Hw •. wSol1L, w2 -> Hw •. wSol2 •. n -> 4L<

• !!! • ! !!!!
9w1 ® b p 2 a , w2 ® 4 15 p 2=

Moreover, we define fk and the first and fourth eigenmodes of the buckled beam as

modeshapes6 = Join@modeshape,
8F1 -> Function@x, shape1 •• EvaluateD, F2 -> Function@x, shape4 •• EvaluateD<D;

We consider a combination of a one-to-one internal resonance between the first and fourth modes and a principal parametric
resonance of the fourth mode; that is,

ResonanceConds6 = 9w2 == w1 + e2 s1 , W == 2 w2 + e2 s2 =;

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies6, modeshapes6, ResonanceConds6D •• Timing

929.522 Second,
9-2 I m w1 A1 @T2D - E2 I T2 s1 +I T2 s2 f p 2 A1 @T2D + 12 p 4 a A1 @T2D2 A1@T2 D + 480 E2 I T2 s1 p 4 a A2 @T2D2 A1@T2 D +
• • •

€€€€€€€€€€€€ p 4 a A1@T2 D A2@T2 D A2@T2 D - 2 I w1 A¢1@T2 D == 0,

1216 •
3
-2 I m w2 A2 @T2D + €€€€€€€€€€€€ p 4 a A1 @T2D A2 @T2D A1 @T2D - €€€€€€ EI T2 s2 f p 2 A2@T2 D +
1216 • 32 •
3 3

p a A1@T2 D A2 @T2D - €€€€€€€€€€€€ p a A2 @T2D2 A2@T2 D - 2 I w2 A¢2@T2 D == 0==

-2 I T2 s1 4 2• 1024 4 •
480 E
3

which indicates that this one-to-one internal resonance between the first and fourth modes is activated.
Chapter8.nb 269

Ÿ Three-to-One Internal Resonances

When the frequency of the fourth mode is approximately three times that of the first mode, we denote these frequencies by
w2 and w1, respectively, as

frequencies7 = 8w1 -> Hw •. wSol1L, w2 -> Hw •. wSol2 •. n -> 4L<

• !!! • ! !!!!
9w1 ® b p 2 a , w2 ® 4 15 p 2=

Moreover, we define fk and the first and fourth eigenmodes of the buckled beam as

modeshapes7 = Join@modeshape,
8F1 -> Function@x, shape1 •• EvaluateD, F2 -> Function@x, shape4 •• EvaluateD<D;

We consider a combination of a three-to-one internal resonance between the first and fourth modes and a principal paramet-
ric resonance of the fourth mode; that is,

ResonanceConds7 = 9w2 == 3 w1 + e2 s1, W == 2 w2 + e2 s2 =;

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies7, modeshapes7, ResonanceConds7D •• Timing

922.653 Second,

9-2 I m w1 A1 @T2D + 12 p 4 a A1 @T2D2 A1@T2 D + €€€€€€€€€€€€ p 4 a A1@T2 D A2@T2 D A2@T2 D - 2 I w1 A¢1@T2 D == 0,

• 3264 •
35
-2 I m w2 A2 @T2D + €€€€€€€€€€€€ p a A1 @T2D A2 @T2D A1 @T2D -
3264 4 •
35
f p A2 @T2D - €€€€€€€€€€€€ p 4 a A2 @T2D2 A2@T2 D - 2 I w2 A¢2@T2 D == 0==
2•
288 I T2 s2 9216 •
€€€€€€€€€ E
35 35

which indicates that this three-to-one internal resonance between the first and fourth modes is not activated.

When the frequency of the first mode is approximately three times that of the fourth mode, we denote these frequencies by
w2 and w1, respectively, as

frequencies8 = 8w2 -> Hw •. wSol1L, w1 -> Hw •. wSol2 •. n -> 4L<

• !!! • ! !!!!
9w2 ® b p 2 a , w1 ® 4 15 p 2=

Moreover, we define fk and the fourth and first eigenmodes of the buckled beam as

modeshapes8 = Join@modeshape,
8F1 -> Function@x, shape4 •• EvaluateD, F2 -> Function@x, shape1 •• EvaluateD<D;

We consider a combination of a three-to-one internal resonance between the first and fourth modes and a principal paramet-
ric resonance of the fourth mode; that is,

ResonanceConds8 = 9w2 == 3 w1 + e2 s1, W == 2 w2 + e2 s2 =;

270 Chapter8.nb

Using MMSDirect11, we obtain the solvability conditions

MMSDirect11@frequencies8, modeshapes8, ResonanceConds8D •• Timing

922.072 Second,

9-2 I m w1 A1 @T2D + €€€€€€€€€€€€ p 4 a A1 @T2D2 A1@T2 D - €€€€€€€€€€€€ p 4 a A1@T2 D A2@T2 D A2@T2 D - 2 I w1 A¢1@T2 D == 0,
1024 • 8896 •
5 5

-2 I m w2 A2 @T2D - €€€€€€€€€€€€ p a A1 @T2D A2 @T2D A1 @T2D -

8896 4 •

f p A2 @T2D + 12 p4 a A2@T2 D2 A2 @T2 D - 2 I w2 A¢2 @T2 D == 0==

5
I T2 s2 2• •
E

which indicates again that this three-to-one internal resonance between the first and fourth modes is not activated.

à 8.2 Circular Cylindrical Shells

Following McIvor (1966) and Raouf and Nayfeh (1990), we write the equations of motion governing the dynamic response of
an infinitely long cylindrical shell to a primary-resonance excitation of one of its two orthogonal flexural modes as

EOM = 9w - yq + wt,t + a2 Hw + 2 wq,q + wq,q,q,q L ==

- €€€€q€€€ - w wq2 - y2t + w y2t - 2 w yq + €€€€ wq2 yq + y2q - w y2q + f e3 Cos@n qD Cos@W tD H1 - w + yqL -
w2 1
2 2
3
w wq,q - w 2 wq,q + €€€€ wq2 wq,q + yq wq,q + w yq wq,q - y2q wq,q + wq yq,q + w wq yq,q -

i
2

2 wq yq yq,q + a2 j
j
j-3 w - €€€€€€€€€€€€€ - 11 w wq,q + 4 yq wq,q - 6 wq,q + 4 wq yq,q - 8 wq wq,q,q +
2 11 wq2 2

k 2
y
z,
8 yq,q wq,q,q + w yq,q,q + 5 wq,q yq,q,q - 4 w wq,q,q,q + 4 yq wq,q,q,q + wq yq,q,q,q z z
{
w 3
wq + yt,t - yq,q == 2 w wq + €€€€q€€ + 2 wt yt - 2 w wt yt - 2 wq yq + 2 w wq yq + wq wq,q +
2
w wq wq,q - 2 wq yq wq,q + 2 w yt,t - w 2 yt,t - 2 w yq,q + w 2 yq,q - wq2 yq,q +
a2 Hwq wq,q - w wq,q,q - wq,q wq,q,q + wq wq,q,q,qL= •. 8w -> w@q, tD, Hs : w È yL m__ :> ¶ m s@q, tD<;

We seek a second-order uniform asymptotic expansion of the solution of EOM in the form

solRule = 9w -> IEvaluateASumAej wj @#1, #2, #3, #4D, 8j, 3<EE &M,
y -> IEvaluateASumAej yj @#1, #2, #3, #4D, 8j, 3<EE &M=

8w ® He w1 @#1, #2, #3, #4D + e2 w2@#1, #2, #3, #4D + e3 w3@#1, #2, #3, #4D &L,
y ® He y1 @#1, #2, #3, #4D + e2 y2@#1, #2, #3, #4D + e3 y3@#1, #2, #3, #4D &L<

where #1 stands for q and #2, #3, and #4 stand for T0 , T1 , and T2, respectively. Transforming the time derivatives in EOM in
terms of the scales T0 , T1, and T2 , substituting solRule into EOM, expanding the result for small e, and discarding terms of
order higher than e3, we obtain
Chapter8.nb 271

eq82a = HEOM •. 8w@x_, tD -> w@x, T0 , T1, T2 D, y@x_, tD -> y@x, T0 , T1, T2D,
Derivative@m_, n_D@w_D@x_, tD -> dt@nD@D@w@x, T0 , T1 , T2D, 8x, m<DD,
t -> T0 < •. solRule •• ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e, we have

eqEps = Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eq82a •• Transpose;

eqEps •. displayRule

982 a2 Hw'' 1 L + a Hw1

2 ''''
L - y'1 + D20 w1 + w1 + a2 w1 == 0, w'1 - y'' 2
1 + D0 y1 == 0<,

9 €€€€ Hw'1 L + €€€€€€€ a2 Hw'1L + 6 a2 Hw'' 1 L + 8 a Hw1L Hw1 L + 2 a Hw2 L + a Hw2 L-

1 2 11 2 2 2 ' ''' 2 '' 2 ''''

Hw''
1 L Hy1 L - 4 a Hw1 L Hy1 L - 4 a Hw1 L Hy'1 L - Hy'1 L - Hw'1L Hy''
1 L - 4 a Hw1L Hy1 L -
2 2
' 2 '' ' 2 '''' 2 2 ' ''

8 a2 Hw'''
1 L Hy1 L - 5 a Hw1 L Hy1 L - a Hw1 L Hy1
'' 2 '' ''' 2 ' ''''
L - y'2 + HD0 y1L2 + D20w2 + 2 HD0 D1w1 L +
Hw''
1 L w1 + 11 a Hw1 L w1 + 4 a Hw1 L w1 + 2 Hy'1 L w1 - a2 Hy'''
1 L w1 + 3 a w1 + w2 + a w2 == 0,
2 '' 2 '''' 2 2 2

-Hw1L Hw1 L - a Hw1L Hw1 L + a Hw1 L Hw1 L - a Hw1 L Hw1 L + w2 + 2 Hw1L Hy1 L - y2 -
' '' 2 ' '' 2 '' ''' 2 ' '''' ' ' ' ''

2 HD0 w1L HD0 y1L + D20y2 + 2 HD0 D1y1 L - 2 Hw'1 L w1 + a2 Hw'''

1 L w1 + 2 Hy1 L w1 - 2 HD0 y1L w1 == 0=,
'' 2

9-f Cos@n qD Cos@T0 WD - €€€€ Hw'1L Hw'' 1 L + Hw1L Hw2 L + 11 a Hw1 L Hw2L + 8 a Hw1 L Hw2L +
3 2 ' ' 2 ' ' 2 ''' '
2

12 a2 Hw''
1 L Hw2 L + 8 a Hw1 L Hw2 L + 2 a Hw3 L + a Hw3 L - €€€€ Hw'1 L Hy'1 L - Hw''
2 L Hy1L -
'' 2 ' ''' 2 '' 2 '''' 1 2 '

4 a2 Hw''
2 L Hy1L - 4 a Hw2 L Hy'1L + Hw''
1 L Hy1 L - Hw2L Hy1 L - 4 a Hw2 L Hy1 L - 8 a Hw2 L Hy1 L +
2
' 2 '''' ' ' 2'' 2 ' '' 2 ''' ''

2 Hw'1 L Hy'1 L Hy''

1 L - 5 a Hw2 L Hy1 L - a Hw2 L Hy1 L - Hw''
1 L Hy2L - 4 a Hw1 L Hy2L -
2 '' ''' 2 ' '''' ' 2 '' '

4 a2 Hw'''' L Hy'2L - 2 Hy'1L Hy'2 L - Hw'1 L Hy''

2 L - 4 a Hw1 L Hy2 L - 8 a Hw1 L Hy2 L -
2 ' '' 2 ''' ''

5 a2 Hw''
1 L Hy2 L - a Hw1 L Hy2 L - y'3 + 2 HD0y1 L HD0y2 L + D20 w3 + 2 HD0y1 L HD1y1 L +
1
''' 2 ' ''''

2 HD0 D1w2 L + D21 w1 + 2 HD0 D2 w1L + Hw'1L w1 + Hw''

2 L w1 + 11 a Hw2 L w1 + 4 a Hw2 L w1 -
2 2 '' 2 ''''

Hw''
1 L Hy1 L w1 + Hy1 L w1 - Hw1 L Hy1 L w1 + 2 Hy2 L w1 - a Hy2 L w1 - HD0 y1 L w1 + Hw1 L w1 +
' ' 2 ' '' ' 2 ''' 2 '' 2

Hw''
1 L w2 + 11 a Hw1 L w2 + 4 a Hw1 L w2 + 2 Hy'1 L w2 - a2 Hy'''
1 L w2 + 6 a w1 w2 + w3 + a w3 == 0,
2 '' 2 '''' 2 2

- €€€€ Hw'1 L - Hw''

1 L Hw2 L - a Hw1 L Hw2L - a Hw1 L Hw'2L - Hw'1L Hw'' 2 L - a Hw1L Hw2 L +
1 3 ' 2 '' ' 2 '''' 2 ' ''

a2 Hw'''
1 L Hw2 L + a Hw1 L Hw2 L - a Hw1 L Hw2 L + w'3 + 2 Hw'1 L Hw''
1 L Hy1 L +
2
'' 2 '' ''' 2 ' '''' '

2 Hw'2 L Hy'1 L + Hw'1 L Hy''

1 L + 2 Hw1 L Hy2L - y3 - 2 HD0w2 L HD0y1 L - 2 HD0 w1 L HD0 y2 L + D0 y3 -
2 ' ' '' 2

2 HD0 y1L HD1 w1L - 2 HD0w1 L HD1y1 L + 2 HD0 D1y2 L + D21 y1 + 2 HD0 D2 y1L - Hw'1L Hw''
1 L w1 -
2 Hw2 L w1 + a Hw2 L w1 - 2 Hw1 L Hy1L w1 + 2 Hy2 L w1 + 2 HD0w1 L HD0y1 L w1 - 2 HD20y2 L w1 -
' 2 ''' ' ' ''

4 HD0 D1y1 L w1 - Hy''

1 L w1 + HD0 y1L w1 - 2 Hw1 L w2 + a Hw1 L w2 + 2 Hy1 L w2 - 2 HD0y1 L w2 == 0==
2 2 2 ' 2 ''' '' 2

Ÿ 8.2.1 First-Order Solution

To determine the solution of the first-order problem, eqEps[[1]], we use the method of separation of variables and let

svRule = 9w1 -> IA EI m #1

EI wm #2 &M, y1 -> IB EI m #1
EI wm #2 &M=;

Substituting the svRule into eqEps[[1]], we obtain the characteristic equations

272 Chapter8.nb

eqEps1lhs = ExpandAE-I m q-I wm T0 #@@1DD •. svRuleE & •ž eqEps@@1DD

8A - I B m + A a2 - 2 A m2 a2 + A m4 a2 - A w2m , I A m + B m2 - B w2m <

Setting the determinant of the coefficient matrix in eqEps1lhs equal to zero yields the following equation governing the
natural frequencies wm of the shell:

HOuter@D, eqEps1lhs, 8A, B<D •• Det •• Collect@#, w m D &L == 0

m2 a2 - 2 m4 a2 + m6 a2 + H-1 - m2 - a2 + 2 m2 a2 - m4 a2 L w2m + w4m == 0

It follows from eqEps1lhs that the ratio Gm = I €€€€BA€€ of the amplitudes of y1 and w1 is given by

GRule = G m -> FactorAI €€€€ •. Solve@eqEps1lhs@@1DD == 0, BD@@1DDE

B
A

1 + a2 - 2 m2 a2 + m4 a2 - w2m
G m ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€
m

Hence, the solution of eqEps[[1]] can be expressed in terms of the linear free-vibration modes

modeShapes = ComplexExpand@8Re@#D, Im@#D<D & •ž 8Exp@I n qD, -I G n Exp@I n qD<

88Cos@n qD, Sin@n qD<, 8Sin@n qD Gn, -Cos@n qD Gn<<

as

sol1Form =
modeShapes.TableAAi,n@T1 , T2D Exp@I w n T0D + Ai,n @T1 , T2D Exp@-I w n T0 D, 8i, 2<E •• Expand
•

8EI T0 wn Cos@n qD A1,n @T1, T2 D + EIT0 wn Sin@n qD A2,n@T1 , T2D +

E-I T0 wn Cos@n qD A1,n @T1, T2 D + E-I T0 wn Sin@n qD A2,n @T1, T2D, EI T0 wn Sin@n qD Gn A1,n@T1 , T2 D -
• •

EI T0 wn Cos@n qD Gn A2,n@T1 , T2D + E-I T0 wn Sin@n qD Gn A1,n @T1 , T2D - E-I T0 wn Cos@n qD Gn A2,n @T1, T2 D<
• •

It follows from the sol1Form that all of the flexural modes are degenerate because two linearly independent eigenfunctions,
namely Cos@n qD and Sin@n qD, correspond to the same frequency wn for each n.

We consider the case of a primary-resonance excitation of the Cos@n qD mode corresponding to the frequency wn ; it is called
the driven mode. Because of the degeneracy, this mode is involved in a one-to-one internal resonance with the Sin@n qD mode,
which is called the companion mode. We assume that neither of these two modes is involved in an internal resonance with
any other mode. Therefore, in the presence of viscous damping, out of the infinite number of modes present in w1 and y1 ,
only the driven mode Cos@n qD and its companion mode Sin@n qD will contribute to the steady-state response (Nayfeh and
Mook, 1979). Consequently, we take the solution of the first-order problem to consist only of these two modes and express it
as

sol1 = 8w1 -> Function@8q, T0 , T1 , T2<, sol1Form@@1DD •• EvaluateD,

y1 -> Function@8q, T0, T1 , T2 <, sol1Form@@2DD •• EvaluateD<;

To describe the nearness of the primary resonance, we introduce the detuning parameter s defined by
Chapter8.nb 273

ResonanceCond = 9W == w n + e2 s=;

where the detuning is ordered at e2 because secular terms appear first at order e3.

Ÿ 8.2.2 Second-Order Solution

Substituting the sol1 into the second-order problem, eqEps[[2]], we obtain

order2Eq =
Table@HeqEps@@1, i, 1DD •. 8w1 -> w2 , y1 -> y2<L == HeqEps@@1, i, 1DD •. 8w1 -> w2, y1 -> y2<L -
HSubtract žž eqEps@@2, iDD •. sol1 •• ExpandL, 8i, 2<D;

Collecting the terms that may lead to secular terms, the terms proportional to EI wn T0 , we have

ST1 = CoefficientA#@@2DD, EI wn T0 E & •ž order2Eq

9-2 I Cos@n qD wn A1,n @T1 , T2D - 2 I Sin@n qD wn A2,n @T1 , T2 D,

H1,0L H1,0L

-2 I Sin@n qD Gn wn A1,n @T1, T2 D + 2 I Cos@n qD Gn wn A2,n @T1 , T2 D=

H1,0L H1,0L

Because the homogeneous parts of order2Eq have a nontrivial solution, the corresponding nonhomogeneous equations
have a solution only if solvability conditions are satisfied. These conditions demand that ST1 be orthogonal to every
solution of the adjoint homogeneous equations. In this case, these equations are self-adjoint and hence solutions of the
adjoint equations are given by the transpose of the modeShapes. Demanding that ST1 be orthogonal to the transpose of the
modeShapes, we obtain the solvability conditions

SCond1 =
SolveAHint@#, 8q, 0, 2 p<D & •ž [email protected] •. intRule1 ••. intRule2 •.
8int@Cos@s_D Sin@s_D, _D -> 0, int@f_@n qD ^2, arg_D -> p<L == 0 ••
Thread, 9AH1,0L
1,n @T1 , T2 D, A2,n @T1, T2D=E@@1DD
H1,0L

9A1,n @T1 , T2 D ® 0, A2,n @T1 , T2D ® 0=

H1,0L H1,0L

Having determined the solvability conditions of the second-order equations, order2Eq, we use a combination of the superpo-
sition principle and the method of undetermined coefficients to determine a particular solution of them. To this end, we
identify the forms of all of the possible terms on the right-hand sides of order2Eq as follows:

collectForm = OuterATimes, 8Cos@n qD, Sin@n qD<,

TableA9Ai,n@T1 , T2 D EI wn T0 , Ai,n @T1 , T2 D E-Iwn T0 =, 8i, 2<E •• FlattenE •• Flatten
•

8EI T0 wn Cos@n qD A1,n @T1, T2 D, E-I T0 wn Cos@n qD A1,n @T1, T2 D,

•

EI T0 wn Cos@n qD A2,n @T1, T2 D, E-I T0 wn Cos@n qD A2,n @T1, T2 D, EI T0 wn Sin@n qD A1,n @T1, T2 D,
•

E-I T0 wn Sin@n qD A1,n @T1, T2 D, EI T0 wn Sin@n qD A2,n @T1, T2 D, E-I T0 wn Sin@n qD A2,n @T1 , T2 D<
• •
274 Chapter8.nb

possibleTerms = Flatten@If@Head@#D === Plus, List žž #, #D & •ž

HOuter@Times, collectForm, collectFormD •• Flatten •• Union •• TrigReduce ••
ExpandLD •. c_? NumberQ form_ -> form •• Union;
possibleTerms •. displayRule

:E2 I T0 wn A21,n , E2 I T0 wn Cos@2 n qD A21,n , E2 IT0 wn Sin@2 n qD A21,n , E2 I T0 wn A1,n A2,n,

E2 I T0 wn Cos@2 n qD A1,n A2,n , E2 I T0 wn Sin@2 n qD A1,n A2,n , E2 I T0 wn A22,n , E2 I T0 wn Cos@2 n qD A22,n,
• • • •
E2 I T0 wn Sin@2 n qD A22,n , A1,n A1,n , Cos@2 n qD A1,n A1,n , Sin@2 n qD A1,n A1,n , A2,n A1,n,
• • •2 •2
Cos@2 n qD A2,n A1,n , Sin@2 n qD A2,n A1,n, E-2 I T0 wn A1,n , E-2 I T0 wn Cos@2 n qD A1,n,
•2 • • • •
E-2 I T0 wn Sin@2 n qD A1,n , A1,n A2,n, Cos@2 n qD A1,n A2,n , Sin@2 n qD A1,n A2,n , A2,n A2,n,
• • • • • •
Cos@2 n qD A2,n A2,n , Sin@2 n qD A2,n A2,n, E-2 I T0 wn A1,n A2,n, E-2 I T0 wn Cos@2 n qD A1,n A2,n,
E-2 I T0 wn Sin@2 n qD A1,n A2,n , E-2 I T0 wn A2,n , E-2 I T0 wn Cos@2 n qD A2,n, E-2 I T0 wn Sin@2 n qD A2,n >
• • •2 •2 •2

possibleTerms •• Length

30

We note that the complex conjugate of each term on the right-hand sides of order2Eq is also a term on their right-hand sides.
Therefore, we use this fact to reduce the algebra and calculate only the particular solutions corresponding to almost half of
the terms because the particular solutions corresponding to the remaining terms are the complex conjugates of those already
calculated. Then, we express the right-hand side of order2Eq[[i]] as the sum of two terms, tempList[i] and its complex
conjugate, for i = 1 and 2, according to

DoAeq2rhsList = List žž order2Eq@@i, 2DD; tempList@iD = eq2rhsList;

IfAterm = # •. conjugateRule;
MemberQ@tempList@iD, termD, tempList@iD = Complement@tempList@iD, 8#<D;

IfA# === term, AppendToAtempList@iD, €€€€ #EEE & •ž Reverse@eq2rhsListD, 8i, 2<E
1
2

Next, we determine the coefficients of the possibleTerms in each of the tempList[i] according to

Do@rhs = Plus žž tempList@iD •. SCond1 •• TrigReduce •• Expand;

coeffs@iD = Coefficient@rhs, possibleTermsD •. HCos È SinL@2 n qD -> 0, 8i, 2<D

and denote them by ci, j as

symbolList1 = MapIndexed@If@#1 =!= 0, cSequencežž #2, 0D &, Array@coeffs, 2D, 82<D

88c1,1 , c1,2, 0, 0, 0, c1,6 , c1,7 , c1,8, 0, c1,10 , c1,11, 0, 0, 0, c1,15, 0,

0, 0, 0, 0, 0, c1,22, c1,23 , 0, 0, 0, 0, 0, 0, 0<, 80, 0, c2,3, 0, c2,5 , 0, 0, 0,
c2,9 , 0, 0, c2,12, 0, c2,14 , 0, 0, 0, 0, 0, 0, 0, 0, 0, c2,24 , 0, 0, 0, 0, 0, 0<<

symbolList2@s_D = Plus žž symbolList1 •. c -> s;

where ci, j = coeffs[i][[j]]. Moreover, we associate with the jth term in the symbolList1[[i]] the undetermined coefficient d i, j
in w2 and the undetermined coefficient ei, j in y2 by replacing the ci, j in symbolList1[[i]] with d i, j and ei, j according to
Chapter8.nb 275

sol2temp = 8w2 -> Function@8q, T0 , T1 , T2<, [email protected] •• EvaluateD,

y2 -> Function@8q, T0, T1 , T2 <, [email protected] •• EvaluateD<;

Substituting the sol2temp into the left-hand sides of the second-order equations, order2Eq, setting the coefficient of the jth
term in possibleTerms in the ith equation equal to ci, j , and solving the resulting algebraic equations, we obtain

deRule = Solve@
Flatten@Table@Coefficient@order2Eq@@i, 1DD •. sol2temp, possibleTermsD •. f_@2 n qD -> 0,
8i, 2<DD == Flatten@symbolList1D •• Thread •• Union,
8symbolList2@dD, symbolList2@eD< •• Flatten •• Union •• RestD@@1DD

Solve::svars : Equations may not give solutions for all "solve" variables.

:e1,1 ® 0, e1,7 ® 0, d1,10 ® €€€€€€€€€€€€€ , d1,22 ® €€€€€€€€€€€€€€ , e1,11 ® 0, e1,15 ® 0,

c1,10 c1,22
1+a 2 1 + a2
-2 n c1,11 - c2,12 -2 n c1,15 + c2,14
4 n H-1 + 4 n L a 4 n2 H-1 + 4 n2L a2
c2,12 c2,14
e1,23 ® 0, e2,12 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€ + €€€€€€€€€€€€ , e2,14 ® €€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ ,
2 2 2 2 4 n 2 4 n 2 2

-2 n c1,23 - c2,24
4 n2 H-1 + 4 n2 L a2
c2,24 c1,1 c1,7
e2,24 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€ + €€€€€€€€€€€€ , d1,1 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€ , d1,7 ® €€€€€€€€€€€€€€€€€€€€€€€€€€ , e1,2 ® 0, e1,6 ® 0,
2 4n 2 1 + a - 4 wn
2 2 1 + a2 - 4 w2n
2 n H-H4 n2 - 4 w2n L c1,2 - 2 n c2,3 L
H4 n2 - 4 w2nL H-4 n2 + H4 n2 - 4 w2n L H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2n LL
c2,3
e1,8 ® 0, e2,3 ® €€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ ,
4 n2 - 4 w2n
H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nL HH4 n2 - 4 w2n L c1,6 - 2 n c2,5 L
2 n H4 n2 - H4 n2 - 4 w2nL H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nLL
c1,6
e2,5 ® €€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ ,

H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nL HH4 n2 - 4 w2n L c1,8 + 2 n c2,9 L

2n

2 n H4 n2 - H4 n2 - 4 w2n L H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nLL
c1,8
e2,9 ® - €€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ ,
2n
-2 n c1,11 - c2,12 -2 n c1,15 + c2,14 -2 n c1,23 - c2,24
2 n H-1 + 4 n L a 2 n H-1 + 4 n L a 2 n H-1 + 4 n2L2 a2
d1,11 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ , d1,15 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ , d1,23 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,
2 2 2 2 2 2

-H4 n2 - 4 w2n L c1,2 - 2 n c2,3

-4 n2 + H4 n2 - 4 w2nL H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nL
d2,12 ® 0, d2,14 ® 0, d2,24 ® 0, d1,2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ ,

H4 n2 - 4 w2n L c1,6 - 2 n c2,5

4 n2 - H4 n2 - 4 w2nL H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nL
d1,6 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€ ,

H4 n2 - 4 w2n L c1,8 + 2 n c2,9

4 n - H4 n - 4 w2nL H1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nL
d1,8 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€ , d2,3 ® 0, d2,5 ® 0, d2,9 ® 0>
2 2

To simplify the expressions in the second-order solution, we introduce the notation

276 Chapter8.nb

paperNotationRule = ThreadAHsymbolList2@cD •• Union •• RestL ->

9c2 , c1, 2 c1 , c2 , -c1, -c4 , - €€€€ c3, -c3 , -c4 , €€€€ c3 , d1, -2 d1 , -d1, - €€€€ d2 , d2 , €€€€ d2=E ~Join~
1 1 1 1
2 2 2 2

9I1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2nM -> €€€€€€€€€€€€€€€€ €€€€€€€€€€ =

2
Dn + 4 n 4 n2
€€€€€€€€ , PowerA-1 + 4 n2 , -2E Power@a, -2D ->
4 n2 - 4 w2n Qn

:c1,1 ® c2, c1,2 ® c1, c1,6 ® 2 c1, c1,7 ® c2, c1,8 ® -c1, c1,10 ® -c4 , c1,11 ® - €€€€€€ , c1,15 ® -c3 ,
c3
2
c3 d2
c1,22 ® -c4 , c1,23 ® €€€€€€ , c2,3 ® d1, c2,5 ® -2 d1 , c2,9 ® -d1 , c2,12 ® - €€€€€€ , c2,14 ® d2 ,
2 2

c2,24 ® €€€€€€ , 1 + a2 - 8 n2 a2 + 16 n4 a2 - 4 w2n ® €€€€€€€€€€€€€€€€€€€€€€€ , €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ® €€€€€€€€€€ >

2
4 n + Dn 4 n2
H-1 + 4 n2 L2 a2
d2 1
2 4 n2 - 4 w2n Qn

Then, we express the second-order solution as

w2Sol = [email protected] •. deRule ••. paperNotationRule;

w2Sol •. displayRule

E2 I T0 wn c2 A21,n E2 I T0 wn Cos@2 n qD H-2 n d1 - c1 H4 n2 - 4 w2nLL A21,n

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ +
1 + a2 - 4 w2n Dn
E2 I T0 wn Sin@2 n qD H4 n d1 + 2 c1 H4 n2 - 4 w2n LL A1,n A2,n E2 I T0 wn c2 A22,n
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
Dn 1 + a2 - 4 w2n

E2 I T0 wn Cos@2 n qD H-2 n d1 - c1 H4 n2 - 4 w2nLL A22,n 2 n Cos@2 n qD In c3 + €€€€€ M A1,n A1,n

• d2 •
c4 A1,n A1,n 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ -
Dn 1 + a2 Qn

2 n Sin@2 n qD H2 n c3 + d2 L A2,n A1,n 2 n Cos@2 n qD I-n c3 - €€€€ € M A2,n A2,n

• • d2 •
c4 A2,n A2,n 2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€
Qn 1 + a2 Qn

y2Sol = [email protected] •. deRule ••. paperNotationRule •. ei__ -> 0;

y2Sol •. displayRule

i 2 n H-2 n d1 - c1 H4 n2 - 4 w2n LL y
E2 I T0 wn Sin@2 n qD j
j
j €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z
z A1,n +
H4 2 - 4 w2L
d1

k {
2
€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 n 2 - 4 w2 D n

i c1 H4 n2 + Dn L H4 n d1 + 2 c1 H4 n2 - 4 w2nLL y
n n n

E2 I T0 wn Cos@2 n qD j
j
j €€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€ z
z
z A1,n A2,n +
k n 2 n Dn H4 n2 - 4 w2n L {
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

i c1 H4 n2 + Dn L H-2 n d1 - c1 H4 n2 - 4 w2n LL y
E2 I T0 wn Sin@2 n qD j
j
j €€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€ z
z
z A2,n +
k H4 2 - 4 w2 L
{
2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 n 2 n D n

i € y
n n

j
j 2 z z
Sin@2 n qD j
j 8 n2 z
z A1,n A1,n +
d2
d2 n c3 + €€€€ •

k {
- €€€€
€ €€€€
€ - €€€€€€€€€€€€€€€€
€€€€€
Q n

i
j d2
d2 y
i d2 2 n c3 + d2 y 2 z
Cos@2 n qD j €€€€€€€ z
j €€€€€€€€€€ + €€€€€€€€€€€€€€€€ z A2,n A1,n + Sin@2 n qD j
j
j €€€€ z
z
z
• -n c3 - €€€€ € •
k4n { k8n {
€€€€
€ €€€€
€ - €€€€€€€€€€€€€€€€
€€€€ A2,n A2,n
2 Qn 2 Qn

We express this solution in pure function form as

sol2 = 8w2 -> Function@8q, T0 , T1 , T2<, w2Sol + Hw2Sol •. conjugateRuleL •• EvaluateD,

y2 -> Function@8q, T0, T1 , T2 <, y2Sol + Hy2Sol •. conjugateRuleL •• EvaluateD<;
Chapter8.nb 277

Ÿ 8.2.3 Solvability Conditions

Substituting the sol1 and sol2 into eqEps[[3]] yields

order3Eq =
Table@HeqEps@@1, i, 1DD •. 8w1 -> w3 , y1 -> y3<L == HeqEps@@1, i, 1DD •. 8w1 -> w3, y1 -> y3<L -
HSubtract žž eqEps@@3, iDD •. sol1 •. sol2 •• TrigToExp •• ExpandL, 8i, 2<D; •• Timing

890.5 Second, Null<

To convert the terms that lead to small-divisor terms in the third-order equations, order3Eq, into terms that lead to secular
terms, we introduce the rule

expRule1 = Exp@arg_D :> ExpAExpand@arg •. HResonanceCond •. Equal -> RuleLD •. e2 T0 -> T2E;

Then, the sources of secular terms in order3Eq, the terms proportional to Exp@± I n q + I wn T0D, are

ST2 =
Expand@Coefficient@#@@2DD •. Exp@c1_. + Complex@0, s_ •; Abs@sD != 1D n qD -> 0 •. expRule1,
Exp@I w n T0 DD •. Exp@c1_. + Complex@0, m_D n qD ->
HCos@n qD + I m Sin@n qDL Exp@c1DD & •ž order3Eq; •• Timing

864.343 Second, Null<

These terms consist of linear and cubic terms in the complex-valued amplitudes Ai. The cubic terms are proportional to

cubicTerms =
Flatten@If@Head@#D === Plus, List žž #, #D & •ž HNest@Outer@Times, collectForm, #D &,
collectForm, 2D •• Flatten •• Union •• TrigReduce •• ExpandLD •.
8f_@s_ n qD -> 0, Exp@c1_. + Complex@0, s_ •; s != 1D w n T0 D -> 0< •.
Exp@_D -> 1 •. c_? NumberQ form_ -> form •• Union •• Rest;
cubicTerms •. displayRule

8Cos@n qD A21,n A1,n , Sin@n qD A21,n A1,n, Cos@n qD A1,n A2,n A1,n, Sin@n qD A1,n A2,n
• • • •
A1,n,
• • • •
Cos@n qD A22,n A1,n , Sin@n qD A22,n A1,n, Cos@n qD A21,n A2,n , Sin@n qD A21,n A2,n ,
A2,n<
• • • •
Cos@n qD A1,n A2,n A2,n , Sin@n qD A1,n A2,n A2,n , Cos@n qD A22,n A2,n , Sin@n qD A22,n

And, their coefficients in ST2 are

coef2 = Outer@Coefficient, ST2, cubicTermsD;

which we denote by

symbolList3 = MapIndexed@If@#1 =!= 0, gSequencežž #2, 0D &, coef2, 82<D;

278 Chapter8.nb

notationRule = HsymbolList3 •• Flatten •• Union •• RestL ->

8E2 , E1, E3 , E3 , E1, E2, -G2 , G1, -G3 , G3, -G1 , G2< •• Thread

8g1,1 ® E2, g1,4 ® E1, g1,5 ® E3, g1,8 ® E3, g1,9 ® E1 , g1,12 ® E2 ,
g2,2 ® -G2 , g2,3 ® G1 , g2,6 ® -G3 , g2,7 ® G3 , g2,10 ® -G1 , g2,11 ® G2 <

In terms of this notation, ST2 can be rewritten as

ST2New = H# •. Thread@cubicTerms -> 0D & •ž ST2L + symbolList3.cubicTerms •. notationRule •.

D@SCond1, T1 D;
ST2New •. displayRule

9 €€€€ EI T2 s f Cos@n qD - 2 I Cos@n qD HD2 A1,n L wn - 2 I HD2A2,n L Sin@n qD wn +

1
2
• • • •
Cos@n qD E2 A21,n A1,n + Sin@n qD E1 A1,n A2,n A1,n + Cos@n qD E3 A22,n A1,n + Sin@n qD E3 A21,n A2,n +
Cos@n qD E1 A1,n A2,n A2,n + Sin@n qD E2 A22,n A2,n , 2 I Cos@n qD HD2A2,n L Gn wn -
• •

2 I HD2 A1,n L Sin@n qD Gn wn - Sin@n qD G2 A21,n A1,n + Cos@n qD G1 A1,n A2,n A1,n -
• •

A2,n =
• • • •
Sin@n qD G3 A22,n A1,n + Cos@n qD G3 A21,n A2,n - Sin@n qD G1 A1,n A2,n A2,n + Cos@n qD G2 A22,n

The solvability conditions of order3Eq demand that the vector ST2New be orthogonal to the transpose of the modeShapes;
that is,

SCond2 = ExpandA €€€€€ int@#, 8q, 0, 2 p<D & •ž [email protected] •. intRule1 ••.
1
p
intRule2 •. 8int@Cos@s_D Sin@s_D, _D -> 0, int@f_@n qD ^2, _D -> p<E == 0 •• Thread;
SCond2 •. displayRule

9 €€€€ EI T2 s f - 2 I HD2A1,n L wn - 2 I HD2 A1,nL G2n wn + E2 A21,n A1,n -

1 •
2
• • • • •
G2 Gn A21,n A1,n + E3 A22,n A1,n - G3 Gn A22,n A1,n + E1 A1,n A2,n A2,n - G1 Gn A1,n A2,n A2,n == 0,
-2 I HD2A2,n L wn - 2 I HD2A2,nL G2n wn + E1 A1,n A2,n A1,n - G1 Gn A1,n A2,n A1,n +
• •
• • • •
E3 A21,n A2,n - G3 Gn A21,n A2,n + E2 A22,n A2,n - G2 Gn A22,n A2,n == 0=

à 8.3 Near-Square Plates

We consider the nonlinear transverse vibrations of a near-square simply supported plate that is initially flat. We introduce a
Cartesian coordinate system x y z, with the x and y axes lying in the undisturbed plane of the middle surface of the plate, and
let u, v, and w denote the displacements of the middle surface in the x, y, and z directions, respectively. Then, the equations
of motion are (Chu and Herrmann, 1956)

Needs@"Calculus`VectorAnalysis`"D

BH@wD = Biharmonic@w@x, y, tD, Cartesian@x, y, zDD

wH0,4,0L@x, y, tD + 2 wH2,2,0L @x, y, tD + wH4,0,0L @x, y, tD

Chapter8.nb 279

EOM = 9

ux,x + wx wx,x + n Hvx,y + w y wx,y L + €€€€ H1 - nL Hu y,y + vx,y + wx w y,y + w y wx,yL == c-2
1
p ut,t ,
2
v y,y + w y w y,y + n Hux,y + wx wx,y L + €€€€ H1 - nL Hvx,x + ux,y + w y wx,x + wx wx,yL == c-2
1
p vt,t ,
2
1
€€€€€€€ h2 BH@wD - Ni1 wx,x - 2 Ni12 wx,y - Ni2 w y,y - ux wx,x -

€€€€ wx wx,x - v y w y,y - €€€€ w 2y w y,y - n i

jv y wx,x + €€€€ w 2y wx,x + ux w y,y + €€€€ wx2 w y,y y
12
z-
1 2 1
k
1 1
{
F I1 - n M
2 2 2 2

H1 - nL Hu y + vx + wx w y L wx,y == -c-2 p Hwx ut,t + w y vt,t + wt,tL + €€€€€€€€€€€€€€€€

2
€€€€€€€
Es h
= •. Hs : u È v È wL m__ :> ¶ m s@x, y, tD;

and the boundary conditions are

BC = 88u@x, y, tD == 0, w@x, y, tD == 0, ¶x,x w@x, y, tD == 0< •. 88x -> 0<, 8x -> a<<,
8v@x, y, tD == 0, w@x, y, tD == 0, ¶ y,y w@x, y, tD == 0< •. 88y -> 0<, 8y -> b<<< •• Flatten

8u@0, y, tD == 0, w@0, y, tD == 0, wH2,0,0L @0, y, tD == 0, u@a, y, tD == 0,

w@a, y, tD == 0, wH2,0,0L @a, y, tD == 0, v@x, 0, tD == 0, w@x, 0, tD == 0,
wH0,2,0L@x, 0, tD == 0, v@x, b, tD == 0, w@x, b, tD == 0, wH0,2,0L@x, b, tD == 0<

where c2p = Es • HrH1 - n2 LL, Es is Young's modulus, n is Poisson's ratio, r and h are the density and thickness of the plate,
respectively, F is the transverse load, the Nki are the uniformly distributed in-plane applied edge loads, and a and b are the
edge lengths, with b » a. EOM are usually called the dynamic analog of the von Karman equations because they reduce to
von Karman's equations (1910) in the absence of the time derivatives.

• !!!
For thin plates, h • a is very small, which we denote by e . Next, we introduce the aspect ratio c = a • b, define the rule

scaleRule1 = 9h2 -> e a2 , a -> c b=;

and nondimensionalize the variables as

• !! !!!! !!!
charT = 12 • e a • c p ; H* characteristic time *L

nondimRule = 9Hdv : u È v È wL -> Ha dv@#1 • a, #2 • b, #3 • charTD &L,

F -> F Es h • I12 a I1 - n2 MM, x -> a x, y -> b y, t -> charT t=;

Substituting nondimRule and scaleRule1 into EOM and BC and considering the case of no in-plane edge loads, we obtain

EOM1 = Ia HSubtract žž #L •. Nik_ -> 0 •. nondimRule ••. scaleRule1 •• ExpandM == 0 & •ž EOM;

BC1 = BC •. nondimRule •. c_ h_@a__, tD :> h@a, tD

8u@0, y, tD == 0, w@0, y, tD == 0, wH2,0,0L @0, y, tD == 0, u@1, y, tD == 0,

w@1, y, tD == 0, wH2,0,0L @1, y, tD == 0, v@x, 0, tD == 0, w@x, 0, tD == 0,
wH0,2,0L@x, 0, tD == 0, v@x, 1, tD == 0, w@x, 1, tD == 0, wH0,2,0L@x, 1, tD == 0<
280 Chapter8.nb

To write EOM1 in a more concise way, we introduce the display rule

display1@expr_D := expr •. 8Derivative@a_, b_, c_D@w_D@x, y, tD :>

Subscript@w, Sequence žž Join@Table@x, 8a<D, Table@y, 8b<D, Table@t, 8c<DDD<

and obtain

EOM1@@1DD •• display1
EOM1@@2DD •• display1
EOM1@@3DD •• display1

1 1 1 1 1
- €€€€€€ e ut,t + ux,x + €€€€ c2 uy,y - €€€€ c2 n uy,y + €€€€ c vx,y + €€€€ c n vx,y +
12 2 2 2 2
1 1 2 1 1
wx wx,x + €€€€ c2 wy wx,y + €€€€ c n wy wx,y + €€€€ c wx wy,y - €€€€ c2 n wx wy,y == 0
2
2 2 2 2
1 1 1 vx,x 1
€€€€ c ux,y + €€€€ c n ux,y - €€€€€€€ e vt,t + €€€€€€€€€€ - €€€€ n vx,x + c2 vy,y +
2 2 12 2 2
1 1 1 1
€€€€ c wy wx,x - €€€€ c n wy wx,x + €€€€ c wx wx,y + €€€€ c n wx wx,y + c3 wy wy,y == 0
2 2 2 2

F 1 1 1 1 1
- €€€€€€ + €€€€€€ e wx ut,t + €€€€€€€ c e wy vt,t + €€€€€€ e wt,t - ux wx,x - c n vy wx,x - €€€€ w2x wx,x - €€€€ c2 n w2y wx,x -
12 12 12 12 2 2
c2 uy wx,y + c2 n uy wx,y - c vx wx,y + c n vx wx,y - c2 wx wy wx,y + c2 n wx wy wx,y - c2 n ux wy,y -
1 1 1 1 1
c3 vy wy,y - €€€€ c2 n w2x wy,y - €€€€ c4 w2y wy,y + €€€€€€€ e wx,x,x,x + €€€€ c2 e wx,x,y,y + €€€€€€€ c4 e wy,y,y,y == 0
2 2 12 6 12

We use the method of multiple scales to directly attack EOM1 and BC1. To transform the time derivatives in EOM1 in terms
of the scales T0 and T1 , we let

multiScales = 8Hdv : u È v È wL@x_, y_, tD -> dv@x, y, T0, T1 D,

Derivative@m_, n_, o_D@u_D@x_, y_, tD -> dt@oD@D@u@x, y, T0, T1D, 8x, m<, 8y, n<DD<;

Then, we seek a second-order approximate solution in the form

solRule = 9Hdv : u È vL -> IEvaluateASumAej+1 dvj@#1, #2, #3, #4D, 8j, 2<EE &M,
w -> IEvaluateASumAej wj @#1, #2, #3, #4D, 8j, 3<EE &M=

8dv : u È v ® He2 dv1@#1, #2, #3, #4D + e3 dv2 @#1, #2, #3, #4D &L,
w ® He w1 @#1, #2, #3, #4D + e2 w2@#1, #2, #3, #4D + e3 w3@#1, #2, #3, #4D &L<

where #1, #2, #3, and #4 stand for x, y, T0, and T1 , respectively.

We introduce the detuning s1 to describe the nearness of the edge lengths a and b, consider the case of primary resonance,
and hence let

scaleRule2 = 9c -> 1 + e s1 , F -> e3 F@x, yD Cos@W T0 D=;

Substituting multiScales , solRule, and scaleRule2 into EOM1 and BC1, expanding the result for small e, and discarding
terms of order higher than e3, we obtain
Chapter8.nb 281

eq83a =
HEOM1 •. multiScales •. solRule •. scaleRule2 •• TrigToExp •• ExpandAllL •. e n_•;n>3 -> 0;

bc83a = BC1 •. multiScales •. solRule;

Equating coefficients of like powers of e in eq83a and bc83a yields

eqEps =
Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq83a •• Transpose •• Rest •• Rest;

bcEpsuv =
Thread@CoefficientList@Subtract žž #, eD == 0D & •ž Select@bc83a, FreeQ@#, wD &D ••
Transpose •• Rest •• Rest

88u1 @0, y, T0 , T1D == 0, u1 @1, y, T0 , T1D == 0, v1 @x, 0, T0 , T1 D == 0, v1@x, 1, T0 , T1 D == 0<,

8u2 @0, y, T0 , T1D == 0, u2 @1, y, T0 , T1D == 0, v2 @x, 0, T0 , T1 D == 0, v2@x, 1, T0 , T1 D == 0<<

bcEpsw = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž

Select@bc83a, ! FreeQ@#, wD &D •• Transpose •• Rest

99w1 @0, y, T0 , T1D == 0, w1 @0, y, T0 , T1 D == 0,

H2,0,0,0L

w1 @1, y, T0 , T1D == 0, @1, y, T0 , T1 D == 0, w1@x, 0, T0, T1 D == 0,

H2,0,0,0L

wH0,2,0,0L@x, 0, T0 , T1 D == 0, w1@x, 1, T0 , T1 D == 0, wH0,2,0,0L @x, 1, T0 , T1D == 0=,

w1

9w2 @0, y, T0 , T1D == 0, wH2,0,0,0L@0, y, T0 , T1 D == 0, w2@1, y, T0, T1 D == 0,

1 1

@1, y, T0 , T1 D == 0, w2@x, 0, T0 , T1 D == 0, w2 @x, 0, T0 , T1D == 0,

H2,0,0,0L H0,2,0,0L
2

w2 @x, 1, T0 , T1D == 0, @x, 1, T0 , T1 D == 0=,

H0,2,0,0L
w2

9w3 @0, y, T0 , T1D == 0, @0, y, T0 , T1 D == 0, w3@1, y, T0, T1 D == 0,

w2
H2,0,0,0L

@1, y, T0 , T1 D == 0, w3@x, 0, T0 , T1 D == 0, w3 @x, 0, T0 , T1D == 0,

w3
H2,0,0,0L H0,2,0,0L

w3 @x, 1, T0 , T1D == 0, wH0,2,0,0L@x, 1, T0 , T1 D == 0==

w3
3

Next, we introduce the display rule

displayRule1 = 9Derivative@a_, b_, c__DAw_i_E@x_, y_, T0 , __D :>

#2@@1DD-1 &, 8c<E; arg1 =!= 1, arg1, ""E,

SequenceFormAIfAarg1 = Times žž MapIndexedAD #1
Subscript@w, Sequence žž Join@8i<, Table@x, 8a<D, Table@y, 8b<DDDE,
Derivative@a__D@Ai__D@__D :> SequenceFormATimes žž MapIndexedAD#1
#2@@1DD &, 8a<E, Ai E,

Derivative@a__D@Ai__D@__D :> SequenceFormATimes žž MapIndexedAD#1

#2@@1DD &, 8a<E, Ai E,
• •

w_i_ @__, T0, __D -> wi , Ai__@__D -> Ai , Ai__@__D -> Ai =;

• •

and write eqEps in a more concise way as

282 Chapter8.nb

HeqEpsm = 8eqEps@@1, 3DD, eqEps@@1, 81, 2<DD, eqEps@@2, 3DD<L •. displayRule1

9 €€€€€€€ Hw1,x,x,x,xL + €€€€ Hw1,x,x,y,y L + €€€€€€€ Hw1,y,y,y,yL + €€€€€€ HD20w1 L == 0,

1 1 1 1
12 6 12 12

9u1,x,x + €€€€ Hu1,y,yL - €€€€ n Hu1,y,yL + €€€€ Hv1,x,y L + €€€€ n Hv1,x,y L + Hw1,x L Hw1,x,x L +
1 1 1 1
2 2 2 2
€€€€ Hw1,y L Hw1,x,y L + €€€€ n Hw1,y L Hw1,x,y L + €€€€ Hw1,x L Hw1,y,y L - €€€€ n Hw1,x L Hw1,y,yL == 0,
1 1 1 1
2 2 2 2
€€€€ Hu1,x,y L + €€€€ n Hu1,x,yL + €€€€ Hv1,x,xL - €€€€ n Hv1,x,x L + v1,y,y + €€€€ Hw1,y L Hw1,x,x L -
1 1 1 1 1
2 2 2 2 2
€€€€ n Hw1,y L Hw1,x,x L + €€€€ Hw1,x L Hw1,x,y L + €€€€ n Hw1,x L Hw1,x,y L + Hw1,y L Hw1,y,y L == 0=,
1 1 1
2 2 2
F@x, yD - Hu1,xL Hw1,x,xL - n Hv1,yL Hw1,x,xL - €€€€ Hw1,xL2 Hw1,x,xL -
1 -I T0 W 1 I T0 W 1
- €€€€€€ E F@x, yD - €€€€€€€ E
24 24 2

€€€€ n Hw1,y L Hw1,x,xL - Hu1,yL Hw1,x,yL + n Hu1,yL Hw1,x,yL - Hv1,xL Hw1,x,yL + n Hv1,xL Hw1,x,yL -
1 2

Hw1,x L Hw1,y L Hw1,x,y L + n Hw1,x L Hw1,y L Hw1,x,yL - n Hu1,x L Hw1,y,yL - Hv1,y L Hw1,y,yL -
2

€€€€ n Hw1,x L2 Hw1,y,yL - €€€€ Hw1,yL2 Hw1,y,y L + €€€€€€ Hw2,x,x,x,x L + €€€€ Hw2,x,x,y,yL +
1 1 1 1
2 2 12 6
€€€€€€€ Hw2,y,y,y,yL + €€€€€€ HD0w2 L + €€€€ HD0 D1 w1L + €€€€ Hw1,x,x,y,y L s1 + €€€€ Hw1,y,y,y,y L s1 == 0=
1 1 2 1 1 1
12 12 6 3 3

Ÿ 8.3.1 First-Order Solution

Linear vibrations are governed by eqEpsm[[1]] and bcEpsw[[1]]; that is,

Horder1Eq = eqEpsm@@1DDL •. displayRule1

€€€€€€€ Hw1,x,x,x,xL + €€€€ Hw1,x,x,y,y L + €€€€€€€ Hw1,y,y,y,yL + €€€€€€ HD20 w1 L == 0

1 1 1 1
12 6 12 12

order1BC = bcEpsw@@1DD

9w1 @0, y, T0 , T1D == 0, w1 @0, y, T0 , T1 D == 0,

H2,0,0,0L

w1 @1, y, T0 , T1D == 0, @1, y, T0 , T1 D == 0, w1@x, 0, T0, T1 D == 0,

H2,0,0,0L

@x, 0, T0 , T1 D == 0, w1@x, 1, T0 , T1 D == 0, w1 @x, 1, T0 , T1D == 0=

H0,2,0,0L H0,2,0,0L
w1
w1

It follows from order1Eq and order1BC that the linear modeshapes and associated frequencies are given by

modeshapes = 8f m_,n_ :> HSin@m p #1D Sin@n p #2D &L<;

frequencies = 9w m_,n_ :> Im 2 + n2 M p 2=;

To investigate the case of one-to-one internal resonance between the mnth and nmth modes (m ¹ n), we let

w1sol = 9w1 -> FunctionA8x, y, T0, T1 <, A1 @T1 D EI wm,n T0 f m,n @x, yD +
A1 @T1D E-I wm,n T0 f m,n @x, yD + A2 @T1D EI wm,n T0 f n,m @x, yD + A2@T1 D E-Iwm,n T0 f n,m @x, yDE=;
• •
Chapter8.nb 283

Ÿ 8.3.2 Second-Order Solution

Substituting the w1sol into eqEpsm[[2]], we obtain

order2Equ = HeqEpsm@@2, 1, 1DD •. w1 -> H0 &LL ==

HHeqEpsm@@2, 1, 1DD •. w1 -> H0 &LL - eqEpsm@@2, 1, 1DD •. w1sol •. modeshapes ••
TrigReduce •• ExpandL; •• Timing

84.39 Second, Null<

order2Eqv = HeqEpsm@@2, 2, 1DD •. w1 -> H0 &LL ==

HHeqEpsm@@2, 2, 1DD •. w1 -> H0 &LL - eqEpsm@@2, 2, 1DD •. w1sol •. modeshapes ••
TrigReduce •• ExpandL; •• Timing

84.28 Second, Null<

The associated boundary conditions are

order2BC = bcEpsuv@@1DD

8u1 @0, y, T0 , T1D == 0, u1 @1, y, T0 , T1D == 0, v1 @x, 0, T0 , T1 D == 0, v1@x, 1, T0 , T1 D == 0<

To obtain particular solutions for u 1 and v1 , we use the method of undetermined coefficients. To accomplish this, we first
look for all possible terms that appear on the right-hand sides of order2Equ and order2Eqv as follows:

possibleTerms1 =
CasesA8order2Equ@@2DD, order2Eqv@@2DD<, _ Sin@a_D E b_ :> Sin@aD E b , ¥E •• Union;

possibleTerms2 = possibleTerms1 •. Exp@_D -> 1 •• Union;

284 Chapter8.nb

possibleTerms = Join@possibleTerms1, possibleTerms2D

8E-2 I T0 wm,n Sin@2 m p xD, E2 I T0 wm,n Sin@2 m p xD, E-2 I T0 wm,n Sin@2 n p xD, E2 I T0 wm,n Sin@2 n p xD,
E-2 I T0 wm,n Sin@2 m p yD, E2 I T0 wm,n Sin@2 m p yD, E-2 I T0 wm,n Sin@2 n p yD, E2 I T0 wm,n Sin@2 n p yD,
E-2 I T0 wm,n Sin@2 n p x - 2 m p yD, E2 I T0 wm,n Sin@2 n p x - 2 m p yD, E-2I T0 wm,n Sin@2 n p x + 2 m p yD,
E2 I T0 wm,n Sin@2 n p x + 2 m p yD, E-2 I T0 wm,n Sin@2 m p x - 2 n p yD, E2 IT0 wm,n Sin@2 m p x - 2 n p yD,
E-2 I T0 wm,n Sin@m p x - n p x - m p y - n p yD, E2 I T0 wm,n Sin@m p x - n p x - m p y - n p yD,
E-2 I T0 wm,n Sin@m p x + n p x - m p y - n p yD, E2 I T0 wm,n Sin@m p x + n p x - m p y - n p yD,
E-2 I T0 wm,n Sin@m p x - n p x + m p y - n p yD, E2 I T0 wm,n Sin@m p x - n p x + m p y - n p yD,
E-2 I T0 wm,n Sin@m p x + n p x + m p y - n p yD, E2 I T0 wm,n Sin@m p x + n p x + m p y - n p yD,
E-2 I T0 wm,n Sin@m p x - n p x - m p y + n p yD, E2 I T0 wm,n Sin@m p x - n p x - m p y + n p yD,
E-2 I T0 wm,n Sin@m p x + n p x - m p y + n p yD, E2 I T0 wm,n Sin@m p x + n p x - m p y + n p yD,
E-2 I T0 wm,n Sin@m p x - n p x + m p y + n p yD, E2 I T0 wm,n Sin@m p x - n p x + m p y + n p yD,
E-2 I T0 wm,n Sin@m p x + n p x + m p y + n p yD, E2 I T0 wm,n Sin@m p x + n p x + m p y + n p yD,
E-2 I T0 wm,n Sin@2 m p x + 2 n p yD, E2 I T0 wm,n Sin@2 m p x + 2 n p yD, Sin@2 m p xD, Sin@2 n p xD,
Sin@2 m p yD, Sin@2 n p yD, Sin@2 n p x - 2 m p yD, Sin@2 n p x + 2 m p yD, Sin@2 m p x - 2 n p yD,
Sin@m p x - n p x - m p y - n p yD, Sin@m p x + n p x - m p y - n p yD, Sin@m p x - n p x + m p y - n p yD,
Sin@m p x + n p x + m p y - n p yD, Sin@m p x - n p x - m p y + n p yD, Sin@m p x + n p x - m p y + n p yD,
Sin@m p x - n p x + m p y + n p yD, Sin@m p x + n p x + m p y + n p yD, Sin@2 m p x + 2 n p yD<

which satisfy order2BC. We then define the following symbol lists to represent the undetermined coefficients.

usymbolList = Table@ci, 8i, Length@possibleTermsD<D

8c1 , c2, c3, c4 , c5 , c6, c7 , c8 , c9, c10, c11 , c12, c13 , c14 , c15, c16 ,

c33 , c34, c35 , c36 , c37, c38 , c39, c40, c41 , c42, c43 , c44 , c45, c46 , c47, c48<
c17 , c18, c19 , c20 , c21, c22 , c23, c24, c25 , c26, c27 , c28 , c29, c30 , c31, c32,

vsymbolList = usymbolList •. c -> d

8d1 , d2, d3, d4 , d5 , d6, d7 , d8 , d9, d10, d11 , d12, d13 , d14 , d15, d16 ,

d33 , d34, d35 , d36 , d37, d38 , d39, d40, d41 , d42, d43 , d44 , d45, d46 , d47, d48<
d17 , d18, d19 , d20 , d21, d22 , d23, d24, d25 , d26, d27 , d28 , d29, d30 , d31, d32,

Then, we express the solution as

u1sol = u1 -> Function@8x, y, T0, T1 <, usymbolList.possibleTerms •• EvaluateD;

v1sol = v1 -> Function@8x, y, T0, T1 <, vsymbolList.possibleTerms •• EvaluateD;

Substituting u1sol and v1sol into order2Equ and order2Eqv, collecting coefficients of possibleTerms, and solving for the
usymbolList and vsymbolList, we obtain

algEq1 = Coefficient@Subtract žž order2Equ •. u1sol •. v1sol, possibleTermsD == 0 •.

Exp@_ T0D -> 0 •• Thread; •• Timing

867.45 Second, Null<

Chapter8.nb 285

algEq2 = Coefficient@Subtract žž order2Eqv •. u1sol •. v1sol, possibleTermsD == 0 •.

Exp@_ T0D -> 0 •• Thread; •• Timing

865.91 Second, Null<

symbolRule1 = Solve@Join@algEq1, algEq2D, Join@usymbolList, vsymbolListDD@@1DD •• Simplify

p Hm2 - n2 nL A1 @T1 D p Hm2 - n2 nL A1 @T1D2 p Hn2 - m2 nL A2@T1 D

:c1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• 2 • 2
€€€€€€€€€€€€€€€ , c2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ , c3 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ ,

p Hn - m nL A2 @T1 D p Hm - n nL A1 @T1D A1 @T1D p Hn2 - m2 nL A2 @T1D A2 @T1D

16 m 16 m 16 n
2 2 2 2 2 • •
c4 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ , c33 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , c34 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm - n nL A2 @T1 D p Hm - n nL A2 @T1D p Hn - m nL A1@T1 D

16 n 8m 8n
2 2
• 2 2 2 2 2 2
• 2
d5 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ , d6 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ , d7 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ ,

p Hn2 - m2 nL A1 @T1 D2 p Hm2 - n2 nL A2 @T1D A2 @T1D p Hn2 - m2 nL A1 @T1D A1 @T1D

16 m 16 m 16 n
• •
d8 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ , d35 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , d36 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
16 n 8m 8n
c5 ® 0, c6 ® 0, c7 ® 0, c8 ® 0, c35 ® 0, c36 ® 0, d1 ® 0, d2 ® 0, d3 ® 0, d4 ® 0,

d33 ® 0, d34 ® 0, c9 ® €€€€€€€ n p A2@T1 D , c10 ® €€€€€€€ n p A2 @T1D2 , c11 ® €€€€€€€ n p A2 @T1D ,
1 • 2 1 1 • 2

32 32 32
c12 ® €€€€€€ n p A2@T1 D , c13 ® €€€€€€ m p A1@T1 D , c14 ® €€€€€€€ m p A1 @T1D , c31 ® €€€€€€€ m p A1 @T1D ,
1 2 1 • 2 1 2 1 • 2

32 32 32 32

c32 ® €€€€€€ m p A1@T1 D2, c37 ® €€€€€€ n p A2@T1 D A2@T1 D, c38 ® €€€€€€ n p A2 @T1 D A2 @T1 D,
1 1 • 1 •
32 16 16

c39 ® €€€€€€ m p A1@T1 D A1@T1 D, c48 ® €€€€€€€ m p A1 @T1D A1 @T1D, d9 ® - €€€€€€ m p A2@T1 D ,
1 • 1 • 1 • 2

16 16 32
d10 ® - €€€€€€€ m p A2 @T1D , d11 ® €€€€€€ m p A2@T1 D , d12 ® €€€€€€ m p A2@T1 D2, d13 ® - €€€€€€ n p A1@T1 D ,
1 2 1 • 2 1 1 • 2

32 32 32 32
d14 ® - €€€€€€€ n p A1 @T1D2 , d31 ® €€€€€€ n p A1@T1 D , d32 ® €€€€€€ n p A1@T1 D2, d37 ® - €€€€€€ m p A2@T1 D A2@T1 D,
1 1 • 2 1 1 •
32 32 32 16
d38 ® €€€€€€ m p A2@T1 D A2@T1 D, d39 ® - €€€€€€€ n p A1 @T1 D A1 @T1 D, d48 ® €€€€€€ n p A1@T1 D A1@T1 D,
1 • 1 • 1 •
16 16 16

c15 ® €€€€€€ Hm - nL p H-1 + nL A1 @T1D A2 @T1D, c16 ® €€€€€€ Hm - nL p H-1 + nL A1 @T1D A2 @T1D,
1 • • 1
32 32

c21 ® €€€€€€ Hm + nL p H-1 + nL A1 @T1D A2 @T1D, c22 ® €€€€€€ Hm + nL p H-1 + nL A1 @T1D A2 @T1D,
1 • • 1
32 32
c25 ® €€€€€€ Hm + nL p H-1 + nL A1 @T1D A2 @T1D, c26 ® €€€€€€ Hm + nL p H-1 + nL A1 @T1D A2 @T1D,
1 • • 1
32 32
c27 ® €€€€€€ Hm - nL p H-1 + nL A1 @T1D A2 @T1D, c28 ® €€€€€€ Hm - nL p H-1 + nL A1 @T1D A2 @T1D,
1 • • 1
32 32
c40 ® €€€€€€ Hm - nL p H-1 + nL HA2 @T1D A1 @T1D + A1@T1 D A2@T1 DL,
1 • •
32

c43 ® €€€€€€ Hm + nL p H-1 + nL HA2 @T1D A1 @T1D + A1@T1 D A2@T1 DL,

1 • •
32

c45 ® €€€€€€ Hm + nL p H-1 + nL HA2 @T1D A1 @T1D + A1@T1 D A2@T1 DL,

1 • •
32
c46 ® €€€€€€ Hm - nL p H-1 + nL HA2 @T1D A1 @T1D + A1@T1 D A2@T1 DL,
1 • •
32
d15 ® - €€€€€€€ Hm + nL p H-1 + nL A1@T1 D A2@T1 D, d16 ® - €€€€€€€ Hm + nL p H-1 + nL A1@T1 D A2@T1 D,
1 • • 1
32 32
286 Chapter8.nb

d21 ® €€€€€€ Hm - nL p H-1 + nL A1 @T1D A2 @T1D, d22 ® €€€€€€ Hm - nL p H-1 + nL A1 @T1D A2 @T1D,
1 • • 1
32 32

d25 ® - €€€€€€€ Hm - nL p H-1 + nL A1@T1 D A2@T1 D, d26 ® - €€€€€€€ Hm - nL p H-1 + nL A1@T1 D A2@T1 D,
1 • • 1
32 32

d27 ® €€€€€€ Hm + nL p H-1 + nL A1 @T1D A2 @T1D, d28 ® €€€€€€ Hm + nL p H-1 + nL A1 @T1D A2 @T1D,
1 • • 1
32 32
d40 ® - €€€€€€€ Hm + nL p H-1 + nL HA2@T1 D A1@T1 D + A1 @T1 D A2 @T1 DL,
1 • •
32
d43 ® €€€€€€ Hm - nL p H-1 + nL HA2 @T1D A1 @T1D + A1@T1 D A2@T1 DL,
1 • •
32
d45 ® - €€€€€€€ Hm - nL p H-1 + nL HA2@T1 D A1@T1 D + A1 @T1 D A2 @T1 DL,
1 • •
32

d46 ® €€€€€€ Hm + nL p H-1 + nL HA2 @T1D A1 @T1D + A1@T1 D A2@T1 DL,

1 • •

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1 @T1 D A2 @T1 D

32
• •

32 Hm + nL
c17 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1 @T1 D A2 @T1 D

32 Hm + nL
c18 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1 @T1 D A2 @T1 D

• •

32 Hm - nL
c19 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1 @T1 D A2 @T1 D

32 Hm - nL
c20 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1 @T1 D A2 @T1 D

• •

32 Hm - nL
c23 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1 @T1 D A2 @T1 D

32 Hm - nL
c24 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1 @T1 D A2 @T1 D

• •

32 Hm + nL
c29 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1 @T1 D A2 @T1 D

32 Hm + nL
c30 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL HA2 @T1D A1 @T1D + A1 @T1D A2 @T1DL

• •

32 Hm + nL
c41 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL HA2 @T1D A1 @T1D + A1 @T1D A2 @T1DL

• •

32 Hm - nL
c42 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL HA2 @T1D A1 @T1D + A1 @T1D A2 @T1DL

• •

32 Hm - nL
c44 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL HA2 @T1D A1 @T1D + A1 @T1D A2 @T1DL

• •

32 Hm + nL
c47 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1@T1 D A2@T1 D

• •

32 Hm + nL
d17 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1@T1 D A2@T1 D

32 Hm + nL
d18 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1 @T1 D A2 @T1 D

• •

32 Hm - nL
d19 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
Chapter8.nb 287

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1 @T1 D A2 @T1 D

32 Hm - nL
d20 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1@T1 D A2@T1 D

• •

32 Hm - nL
d23 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL A1@T1 D A2@T1 D

32 Hm - nL
d24 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1 @T1 D A2 @T1 D

• •

32 Hm + nL
d29 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL A1 @T1 D A2 @T1 D

32 Hm + nL
d30 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL HA2@T1 D A1@T1 D + A1 @T1D A2 @T1DL

• •

32 Hm + nL
d41 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL HA2 @T1D A1 @T1D + A1 @T1D A2 @T1DL

• •

32 Hm - nL
d42 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL HA2@T1 D A1@T1 D + A1 @T1D A2 @T1DL

• •

32 Hm - nL
d44 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL HA2 @T1D A1 @T1D + A1 @T1D A2 @T1DL

€€€€€€€€€ >
• •

32 Hm + nL
d47 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

Ÿ 8.3.3 Solvability Conditions

Substituting the w1sol, u1sol, and v1sol into eqEpsm[[3]], we obtain

order3Eq = Horder1Eq@@1DD •. w1 -> w2 L ==

HHorder1Eq@@1DD •. w1 -> w2 L - eqEpsm@@3, 1DD •. w1sol •. u1sol •. v1sol •. modeshapes ••
ExpandL; •• Timing

84.18 Second, Null<

The associated boundary conditions are

order3BC = bcEpsw@@2DD

9w2 @0, y, T0 , T1D == 0, wH2,0,0,0L@0, y, T0 , T1 D == 0,

w2 @1, y, T0 , T1D == 0, w2 @1, y, T0 , T1 D == 0, w2@x, 0, T0, T1 D == 0,
H2,0,0,0L
2

@x, 0, T0 , T1 D == 0, w2@x, 1, T0 , T1 D == 0, w2 @x, 1, T0 , T1D == 0=

H0,2,0,0L H0,2,0,0L
w2

To describe the nearness of the primary resonance, we introduce the detuning parameter s2 defined by

ResonanceCond = 8W -> w m,n + e s2<;

Because the homogeneous parts of order3Eq and order3BC have a nontrivial solution, the corresponding nonhomogeneous
problem has a solution only if solvability conditions are satisfied. These conditions demand that the right-hand side of
order3Eq be orthogonal to every solution of the adjoint homogeneous problem. In this case, the problem is self-adjoint.
288 Chapter8.nb

To determine the solvability conditions, we collect the terms that may lead to secular terms, the terms proportional to
EI wm,n T0 . To this end, we use the rule

expRule = Exp@a_D :> Exp@Expand@a •. ResonanceCondD •. e T0 -> T1D;

and obtain

ST = CoefficientAorder3Eq@@2DD •. expRule, EI wm,n T0 E •• TrigReduce; •• Timing

8132.53 Second, Null<

Among all the spatial-dependent non-forcing terms in ST, only the following terms are not orthogonal to the modeshapes:

terms1 = Cases@ST, Cos@m p x + a_. n p y •; Abs@aD == 1D, ¥D •• Union

8Cos@m p x - n p yD, Cos@m p x + n p yD<

terms2 = Cases@ST, Cos@n p x + a_. m p y •; Abs@aD == 1D, ¥D •• Union

8Cos@n p x - m p yD, Cos@n p x + m p yD<

The coefficients of these terms in ST are

terms1Coef = Coefficient@ST, terms1D;

terms2Coef = Coefficient@ST, terms2D;

The projection of terms1 onto Sin@m p xD Sin@n p yD can be obtained as

trigRule = 8Cos@2 _. pD -> 1, Sin@_. pD -> 0<;

proj1Coef = à à Sin@m p xD Sin@n p yD # â x â y & •ž terms1 •. trigRule

1 1

0 0

9 €€€€ , - €€€€ =
1 1
4 4

and the projection of terms2 onto Sin@n p xD Sin@m p yD can be obtained as

proj2Coef = à à Sin@n p xD Sin@m p yD # â x â y & •ž terms2 •. trigRule

1 1

0 0

9 €€€€ , - €€€€ =
1 1
4 4

The forcing term in ST is

f0 = ST •. Cos@_D -> 0

1
€€€€€€€ EI T1 s2 F@x, yD
24

Then, the solvability conditions demand that ST be orthogonal to every linear eigenfunction, modeshapes; that is,
Chapter8.nb 289

intRule3 = 8int@a_ fun_, a1_, b1_D :> a int@fun, a1, b1D •; FreeQ@a, First@a1D È First@b1DD<;

SCond1 =
Expand@24 Hint@Sin@m p xD Sin@n p yD f0, 8x, 0, 1<, 8y, 0, 1<D + proj1Coef.terms1Coef •.
symbolRule1 ••. intRule3 •. int -> IntegrateLD == 0;

SCond2 =
Expand@24 Hint@Sin@n p xD Sin@m p yD f0, 8x, 0, 1<, 8y, 0, 1<D + proj2Coef.terms2Coef •.
symbolRule1 ••. intRule3 •. int -> IntegrateLD == 0;

To express these solvability conditions in a more readable form, we define

basicTerms = 9A1@T1 D EI wm,n T0 , A1 @T1 D E-I wm,n T0 , A2@T1 D EI wm,n T0 , A2 @T1 D E-I wm,n T0 =;
• •

cubicTerms =
IE-I wm,n T0 Nest@Outer@Times, basicTerms, #D &, basicTerms, 2D •• FlattenM •. Exp@_ T0D -> 0 ••
Union •• Rest

8A1 @T1D2 A1@T1 D, A1 @T1 D A2 @T1 D A1 @T1 D, A2 @T1D2 A1@T1 D,

• • •

A1 @T1D2 A2@T1 D, A1 @T1 D A2 @T1 D A2 @T1 D, A2 @T1D2 A2@T1 D<

• • •

Then, we collect the coefficients of cubicTerms from SCond1 and SCond2 and obtain

coef1 = Coefficient@SCond1@@1DD, cubicTermsD •• Simplify

9 €€€€ p 4 H-4 m2 n2 n + m4 H-3 + n2L + n4 H-3 + n2 LL, 0,

9
8
€€€€ p 4 H4 m2 n2 H-2 + n2 L + m4 H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2 LL, 0,
3
8

€€€€ p 4 H4 m2 n2 H-2 + n2 L + m4 H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2 LL, 0=

3
4

coef2 = Coefficient@SCond2@@1DD, cubicTermsD •• Simplify

90, €€€€ p 4 H4 m2 n2 H-2 + n2 L + m4 H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2LL,

3
4
0, €€€€ p 4 H4 m2 n2 H-2 + n2 L + m4 H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2LL,
3
8

0, €€€€ p 4 H-4 m2 n2 n + m4 H-3 + n2L + n4 H-3 + n2LL=

9
8

Hence, SCond1 and SCond2 can be rewritten and combined as

290 Chapter8.nb

SCond = 8Collect@SCond1@@1DD •. Thread@cubicTerms -> 0D, s1 A1@T1 DD + coef1.cubicTerms == 0,

Collect@SCond2@@1DD •. Thread@cubicTerms -> 0D, s1 A2@T1 DD + coef2.cubicTerms == 0<;
SCond •. displayRule1

:EI T1 s2 à à F@x, yD Sin@m p xD Sin@n p yD â y â x +

1 1

0 0

H-2 m2 n2 p 4 - 2 n4 p 4 L A1 s1 + €€€€ p 4 H-4 m2 n2 n + m4 H-3 + n2L + n4 H-3 + n2LL A21 A1 +

9 •
8

€€€€ p H4 m n H-2 + n L + m H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2 LL A22 A1 +

3 4 2 2 2 4 •
8
€€€€ p 4 H4 m2 n2 H-2 + n2L + m4 H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2 LL A1 A2 A2 - I HD1A1 L w m,n == 0,
3 •
4

EI T1 s2 à à F@x, yD Sin@n p xD Sin@m p yD â y â x + H-2 m4 p 4 - 2 m2 n2 p 4 L A2 s1 +

1 1

0 0

€€€€ p 4 H4 m2 n2 H-2 + n2L + m4 H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2 LL A1 A2 A1 +

3 •
4
€€€€ p 4 H4 m2 n2 H-2 + n2L + m4 H-1 - 2 n + n2 L + n4 H-1 - 2 n + n2 LL A21 A2 +
3 •
8
€€€€ p 4 H-4 m2 n2 n + m4 H-3 + n2L + n4 H-3 + n2 LL A22 A2 - I HD1 A2L w m,n == 0>
9 •
8

Ÿ 8.3.4 Mixed Approach

Alternatively, we can use a mixed approach, a combination of discretization and direct approaches, to attack the problem,
EOM1 and BC1. We first define

scaleRule3 = 9Hdv : u È vL -> Ie2 dv@#1, #2, #3D &M, w -> He w@#1, #2, #3D &L=;

Substituting scaleRule2 and scaleRule3 into EOM1, expanding the result for small e, and discarding terms of order higher
than e3, we obtain

eq83b = HEOM1 •. scaleRule2 •. scaleRule3 •• TrigToExp •• ExpandAllL •. e n_•;n>3 -> 0;

To write eq83b in a more concise way, we introduce the dispaly rule

display2@expr_D := expr •. 8Derivative@a_, b_, c_D@w_D@x, y, tD :>

Subscript@w, Sequence žž Join@Table@x, 8a<D, Table@y, 8b<D, Table@t, 8c<DDD<

Hence,

IuEOM = CoefficientAeq83b@@1, 1DD, e2E == 0M •• display2

uy,y 1 vx,y 1 1 1 1 1
ux,x + €€€€€€€€€€ - €€€€ n uy,y + €€€€€€€€€€ + €€€€ n vx,y + wx wx,x + €€€€ wy wx,y + €€€€ n wy wx,y + €€€€ wx wy,y - €€€€ n wx wy,y == 0
2 2 2 2 2 2 2 2

IvEOM = CoefficientAeq83b@@2, 1DD, e2E == 0M •• display2

ux,y 1 vx,x 1 1 1 1 1
€€€€€€€€€€ + €€€€ n ux,y + €€€€€€€€€€ - €€€€ n vx,x + vy,y + €€€€ wy wx,x - €€€€ n wy wx,x + €€€€ wx wx,y + €€€€ n wx wx,y + wy wy,y == 0
2 2 2 2 2 2 2 2
Chapter8.nb 291

IwEOM = eq83b@@3DD •• ThreadAe-2 #, EqualE & •• ExpandAllM •• display2

1 1 wt,t
- €€€€€€ E-I T0 W e F@x, yD - €€€€€€€EI T0 W e F@x, yD + €€€€€€€€€€ - e ux wx,x -
24 24 12
1 2 1 2
e n vy wx,x - €€€€ e wx wx,x - €€€€ e n wy wx,x - e uy wx,y + e n uy wx,y - e vx wx,y + e n vx wx,y -
2 2
1 1
e wx wy wx,y + e n wx wy wx,y - e n ux wy,y - e vy wy,y - €€€€ e n w2x wy,y - €€€€ e w2y wy,y +
2 2
1 1 1 1 1
€€€€€€€ wx,x,x,x + €€€€ wx,x,y,y + €€€€ e s1 wx,x,y,y + €€€€€€ wy,y,y,y + €€€€ e s1 wy,y,y,y == 0
12 6 3 12 3

Linear vibrations are governed by wEOM with e = 0; that is,

Horder1Eqw = wEOM •. e -> 0L •• display2

wt,t 1 1 1
€€€€€€€€€€ + €€€€€€€ wx,x,x,x + €€€€ wx,x,y,y + €€€€€€ wy,y,y,y == 0
12 12 6 12

subject to the boundary conditions BC1. It follows from order1Eqw and BC1 that the linear modeshapes and associated
frequencies are given by

f m,n @x, yD •. modeshapes

Sin@m p xD Sin@n p yD

w m,n •. frequencies

Hm2 + n2 L p 2

To investigate the case of one-to-one internal resonance between the mnth and nmth modes (m ¹ n), we let

w1sol = 8w -> Function@8x, y, t<, h@tD Sin@m p xD Sin@n p yD + z@tD Sin@n p xD Sin@m p yDD<;

Substituting the w1sol into uEOM and vEOM, we obtain

order2Equ =
HuEOM@@1DD •. w -> H0 &LL == HHuEOM@@1DD •. w -> H0 &LL - uEOM@@1DD •. w1sol •• TrigReduceL;
order2Equ •• display2

uy,y 1 vx,y 1
ux,x + €€€€€€€€€€ - €€€€ n uy,y + €€€€€€€€€€ + €€€€ n vx,y ==
2 2 2 2
€€€€€€€ H4 n p Sin@2 n p xD z@tD - 4 m2 n p 3 n Sin@2 n p xD z@tD2 -
1 3 3 2
16
2 m2 n p 3 Sin@2 n p x - 2 m p yD z@tD2 - 2 n3 p 3 Sin@2 n p x - 2 m p yD z@tD2 -
2 m2 n p 3 Sin@2 n p x + 2 m p yD z@tD2 - 2 n3 p 3 Sin@2 n p x + 2 m p yD z@tD2 +
m3 p 3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD - m2 n p3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD +
m n2 p 3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD - n3 p3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD -
m3 p 3 n Sin@m p x - n p x - m p y - n p yD z@tD h@tD + m2 n p 3 n Sin@m p x - n p x - m p y - n p yD
z@tD h@tD - m n2 p 3 n Sin@m p x - n p x - m p y - n p yD z@tD h@tD +
n3 p 3 n Sin@m p x - n p x - m p y - n p yD z@tD h@tD - m3 p3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD -
3 m2 n p 3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD -
292 Chapter8.nb

3 m n2 p 3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD -

n3 p 3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD + m3 p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD -
m2 n p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD -
m n2 p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD +
n3 p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD - m3 p3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
3 m2 n p 3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD -
3 m n2 p 3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
n3 p 3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD + m3 p 3 n Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
m2 n p 3 n Sin@m p x - n p x + m p y - n p yD z@tD h@tD - m n2 p 3 n Sin@m p x - n p x + m p y - n p yD
z@tD h@tD - n3 p 3 n Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
m3 p 3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD + m2 n p3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD +
m n2 p 3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD + n3 p3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD -
m3 p 3 n Sin@m p x + n p x + m p y - n p yD z@tD h@tD - m2 n p 3 n Sin@m p x + n p x + m p y - n p yD
z@tD h@tD - m n2 p 3 n Sin@m p x + n p x + m p y - n p yD z@tD h@tD -
n3 p 3 n Sin@m p x + n p x + m p y - n p yD z@tD h@tD - m3 p3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD +
3 m2 n p 3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD -
3 m n2 p 3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD +
n3 p 3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD + m3 p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD +
m2 n p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD -
m n2 p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD -
n3 p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD + m3 p3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD +
m2 n p 3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD +
m n2 p 3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD + n3 p3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD -
m3 p 3 n Sin@m p x + n p x - m p y + n p yD z@tD h@tD - m2 n p 3 n Sin@m p x + n p x - m p y + n p yD
z@tD h@tD - m n2 p 3 n Sin@m p x + n p x - m p y + n p yD z@tD h@tD -
n3 p 3 n Sin@m p x + n p x - m p y + n p yD z@tD h@tD + m3 p3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD -
m2 n p 3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD +
m n2 p 3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD - n3 p3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD -
m3 p 3 n Sin@m p x - n p x + m p y + n p yD z@tD h@tD + m2 n p 3 n Sin@m p x - n p x + m p y + n p yD
z@tD h@tD - m n2 p 3 n Sin@m p x - n p x + m p y + n p yD z@tD h@tD +
n3 p 3 n Sin@m p x - n p x + m p y + n p yD z@tD h@tD - m3 p3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
3 m2 n p 3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
3 m n2 p 3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
n3 p 3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD + m3 p 3 n Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
m2 n p 3 n Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
m n2 p 3 n Sin@m p x + n p x + m p y + n p yD z@tD h@tD + n3 p 3 n Sin@m p x + n p x + m p y + n p yD
z@tD h@tD + 4 m3 p 3 Sin@2 m p xD h@tD2 - 4 m n2 p 3 n Sin@2 m p xD h@tD2 -
2 m3 p 3 Sin@2 m p x - 2 n p yD h@tD2 - 2 m n2 p 3 Sin@2 m p x - 2 n p yD h@tD2 -
2 m3 p 3 Sin@2 m p x + 2 n p yD h@tD2 - 2 m n2 p 3 Sin@2 m p x + 2 n p yD h@tD2 L

order2Eqv =
HvEOM@@1DD •. w -> H0 &LL == HHvEOM@@1DD •. w -> H0 &LL - vEOM@@1DD •. w1sol •• TrigReduceL;
order2Eqv •• display2
Chapter8.nb 293

ux,y 1 vx,x 1
€€€€€€€€€€ + €€€€ n ux,y + €€€€€€€€€€ - €€€€ n vx,x + vy,y ==
2 2 2 2
€€€€€€€ H4 m p Sin@2 m p yD z@tD2 - 4 m n2 p 3 n Sin@2 m p yD z@tD2 +
1 3 3
16
2 m3 p 3 Sin@2 n p x - 2 m p yD z@tD2 + 2 m n2 p 3 Sin@2 n p x - 2 m p yD z@tD2 -
2 m3 p 3 Sin@2 n p x + 2 m p yD z@tD2 - 2 m n2 p 3 Sin@2 n p x + 2 m p yD z@tD2 -
m3 p 3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD - m2 n p3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD -
m n2 p 3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD - n3 p3 Sin@m p x - n p x - m p y - n p yD z@tD h@tD +
m3 p 3 n Sin@m p x - n p x - m p y - n p yD z@tD h@tD + m2 n p 3 n Sin@m p x - n p x - m p y - n p yD
z@tD h@tD + m n2 p 3 n Sin@m p x - n p x - m p y - n p yD z@tD h@tD +
n3 p 3 n Sin@m p x - n p x - m p y - n p yD z@tD h@tD + m3 p3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD +
3 m2 n p 3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD +
3 m n2 p 3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD +
n3 p 3 Sin@m p x + n p x - m p y - n p yD z@tD h@tD - m3 p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD +
m2 n p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD +
m n2 p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD -
n3 p 3 n Sin@m p x + n p x - m p y - n p yD z@tD h@tD - m3 p3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
3 m2 n p 3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD -
3 m n2 p 3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
n3 p 3 Sin@m p x - n p x + m p y - n p yD z@tD h@tD + m3 p 3 n Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
m2 n p 3 n Sin@m p x - n p x + m p y - n p yD z@tD h@tD - m n2 p 3 n Sin@m p x - n p x + m p y - n p yD
z@tD h@tD - n3 p 3 n Sin@m p x - n p x + m p y - n p yD z@tD h@tD +
m3 p 3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD - m2 n p3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD +
m n2 p 3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD - n3 p3 Sin@m p x + n p x + m p y - n p yD z@tD h@tD -
m3 p 3 n Sin@m p x + n p x + m p y - n p yD z@tD h@tD + m2 n p 3 n Sin@m p x + n p x + m p y - n p yD
z@tD h@tD - m n2 p 3 n Sin@m p x + n p x + m p y - n p yD z@tD h@tD +
n3 p 3 n Sin@m p x + n p x + m p y - n p yD z@tD h@tD + m3 p3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD -
3 m2 n p 3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD +
3 m n2 p 3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD -
n3 p 3 Sin@m p x - n p x - m p y + n p yD z@tD h@tD - m3 p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD -
m2 n p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD +
m n2 p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD +
n3 p 3 n Sin@m p x - n p x - m p y + n p yD z@tD h@tD - m3 p3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD +
m2 n p 3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD -
m n2 p 3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD + n3 p3 Sin@m p x + n p x - m p y + n p yD z@tD h@tD +
m3 p 3 n Sin@m p x + n p x - m p y + n p yD z@tD h@tD - m2 n p 3 n Sin@m p x + n p x - m p y + n p yD
z@tD h@tD + m n2 p 3 n Sin@m p x + n p x - m p y + n p yD z@tD h@tD -
n3 p 3 n Sin@m p x + n p x - m p y + n p yD z@tD h@tD + m3 p3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD +
m2 n p 3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD +
m n2 p 3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD + n3 p3 Sin@m p x - n p x + m p y + n p yD z@tD h@tD -
m3 p 3 n Sin@m p x - n p x + m p y + n p yD z@tD h@tD - m2 n p 3 n Sin@m p x - n p x + m p y + n p yD
z@tD h@tD - m n2 p 3 n Sin@m p x - n p x + m p y + n p yD z@tD h@tD -
n3 p 3 n Sin@m p x - n p x + m p y + n p yD z@tD h@tD - m3 p3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
3 m2 n p 3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
3 m n2 p 3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
n3 p 3 Sin@m p x + n p x + m p y + n p yD z@tD h@tD + m3 p 3 n Sin@m p x + n p x + m p y + n p yD z@tD h@tD -
294 Chapter8.nb

m2 n p 3 n Sin@m p x + n p x + m p y + n p yD z@tD h@tD -

m n2 p 3 n Sin@m p x + n p x + m p y + n p yD z@tD h@tD + n3 p 3 n Sin@m p x + n p x + m p y + n p yD
z@tD h@tD + 4 n3 p 3 Sin@2 n p yD h@tD2 - 4 m2 n p 3 n Sin@2 n p yD h@tD2 +
2 m2 n p 3 Sin@2 m p x - 2 n p yD h@tD2 + 2 n3 p 3 Sin@2 m p x - 2 n p yD h@tD2 -
2 m2 n p 3 Sin@2 m p x + 2 n p yD h@tD2 - 2 n3 p 3 Sin@2 m p x + 2 n p yD h@tD2 L

To obtain the particular solutions for u and v, we use the method of undetermined coefficients. To accomplish this, we first
look for all possible terms that appear on the right-hand sides of order2Equ and order2Eqv as follows:

possibleTerms =
Cases@8order2Equ@@2DD, order2Eqv@@2DD<, #, ¥D & •ž 9_ Sin@a_D h1_@tD2 -> Sin@aD h1@tD2 ,
_ Sin@a_D h1_@tD h2_@tD -> Sin@aD h1@tD h2@tD= •• Flatten •• Union

8Sin@2 n p xD z@tD2, Sin@2 m p yD z@tD2, Sin@2 n p x - 2 m p yD z@tD2 , Sin@2 n p x + 2 m p yD z@tD2,

Sin@m p x - n p x - m p y - n p yD z@tD h@tD, Sin@m p x + n p x - m p y - n p yD z@tD h@tD,
Sin@m p x - n p x + m p y - n p yD z@tD h@tD, Sin@m p x + n p x + m p y - n p yD z@tD h@tD,
Sin@m p x - n p x - m p y + n p yD z@tD h@tD, Sin@m p x + n p x - m p y + n p yD z@tD h@tD,

Sin@2 m p xD h@tD2, Sin@2 n p yD h@tD2, Sin@2 m p x - 2 n p yD h@tD2 , Sin@2 m p x + 2 n p yD h@tD2<

Sin@m p x - n p x + m p y + n p yD z@tD h@tD, Sin@m p x + n p x + m p y + n p yD z@tD h@tD,

which satisfy BC1. We then define the following symbol lists to represent the undetermined coefficients.

usymbolList = Table@ci, 8i, Length@possibleTermsD<D

8c1 , c2, c3, c4 , c5 , c6, c7 , c8 , c9, c10, c11 , c12, c13 , c14 , c15, c16 <

vsymbolList = usymbolList •. c -> d

8d1 , d2, d3, d4 , d5 , d6, d7 , d8 , d9, d10, d11 , d12, d13 , d14 , d15, d16 <

Then, we express the solution as

usol = u -> Function@8x, y, t<, usymbolList.possibleTerms •• EvaluateD;

vsol = v -> Function@8x, y, t<, vsymbolList.possibleTerms •• EvaluateD;

Substituting usol and vsol into order2Equ and order2Eqv, collecting coefficients of possibleTerms, and solving for usymbol-
List and vsymbolList, we obtain

algEq1 = Coefficient@Subtract žž order2Equ •. usol •. vsol, possibleTermsD == 0 •• Thread;

algEq2 = Coefficient@Subtract žž order2Eqv •. usol •. vsol, possibleTermsD == 0 •• Thread;

Chapter8.nb 295

HsymbolRule2 =
Solve@Join@algEq1, algEq2D, Join@usymbolList, vsymbolListDD@@1DD •• SimplifyL •• Timing

p Hn2 - m2 nL p Hm2 - n2 nL p Hm2 - n2 nL p Hn2 - m2 nL

:2.48 Second, :c1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , c13 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , d2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , d14 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
16 n 16 m 16 m 16 n
np np mp mp mp
c2 ® 0, c14 ® 0, d1 ® 0, d13 ® 0, c3 ® €€€€€€€€ , c4 ® €€€€€€€€ , c15 ® €€€€€€€€ , c16 ® €€€€€€€€ , d3 ® - €€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL

32 32 32 32 32

32 Hm - nL
mp np np
d4 ® €€€€€€€€ , d15 ® - €€€€€€€€ , d16 ® €€€€€€€€ , c7 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL

32 32 32

32 Hm + nL
c12 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL

32 Hm - nL
d7 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL

32 Hm + nL
d12 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL

c5 ® €€€€€€ Hm - nL p H-1 + nL, c6 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
32 Hm + nL
1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL

32

c8 ® €€€€€€ Hm + nL p H-1 + nL, c9 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

32 Hm - nL
1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
32

c10 ® €€€€€€ Hm + nL p H-1 + nL, c11 ® €€€€€€ Hm - nL p H-1 + nL,

1 1

p Hm2 H-1 + nL + n2 H-1 + nL - 2 m n H1 + nLL

32 32

d5 ® - €€€€€€€ Hm + nL p H-1 + nL, d6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

32 Hm + nL
1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

p Hm2 H-1 + nL + n2 H-1 + nL + 2 m n H1 + nLL

32

d8 ® €€€€€€ Hm - nL p H-1 + nL, d9 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

32 Hm - nL
1
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
32

d10 ® - €€€€€€€ Hm - nL p H-1 + nL, d11 ® €€€€€€ Hm + nL p H-1 + nL>>

1 1
32 32

Ÿ Solvability Conditions

We introduce two time scales T0 and T1 and seek a first-order uniform asymptotic expansion of the solution of wEOM in the
form

solRule = 9w -> FunctionA8x, y, T0 , T1<, SumAej-1 wj@x, y, T0 , T1 D, 8j, 2<EE=;

Transforming the time derivatives in wEOM in terms of the scales T0 and T1, substituting solRule into wEOM, expanding the
result for small e, and discarding terms of order higher than e, we obtain

eq83c =
HwEOM •. 8Hs : u È v È wL@x, y, tD -> s@x, y, T0, T1D, Derivative@a_, b_, c_D@h_D@x, y, tD ->
dt@cD@D@h@x, y, T0 , T1D, 8x, a<, 8y, b<DD< •. solRule •• ExpandAllL •. e n_•;n>1 -> 0;

Equating coefficients of like powers of e, we have

eqEps = CoefficientList@eq83c@@1DD, eD == 0 •• Thread;

296 Chapter8.nb

Accounting for the two time scales T0 and T1 , we create a rule for h@tD and z@tD

ampRule = 9h@tD -> A1@T1 D EI wm,n T0 + A1@T1 D E-I wm,n T0 , z@tD -> A2 @T1D EI wm,n T0 + A2 @T1D E-I wm,n T0 =;
• •

and rewrite w1sol, usol, and vsol as

sol1 = 8w1 -> Function@8x, y, T0, T1 <, w@x, y, tD •. w1sol •. ampRule •• EvaluateD,
u -> Function@8x, y, T0 , T1<, u@x, y, tD •. usol •. ampRule •• EvaluateD,
v -> Function@8x, y, T0 , T1<, v@x, y, tD •. vsol •. ampRule •• EvaluateD<;

Substituting the sol1 into the second-order problem, eqEps[[2]], we obtain

order2Eqw = HeqEps@@1, 1DD •. w1 -> w2 L == HeqEps@@1, 1DD •. w1 -> w2 L - eqEps@@2, 1DD •. sol1 ••
ExpandAll; •• Timing

811.09 Second, Null<

Collecting the terms that may lead to secular terms, the terms proportional to EI wm,n T0 , we have

ST = CoefficientAorder2Eqw@@2DD •. expRule, EI wm,n T0 E •• TrigReduce; •• Timing

8156.76 Second, Null<

Among all the spatial-dependent non-forcing terms in ST, only the following terms are not orthogonal to the modeshapes:

terms1 = Cases@ST, Cos@m p x + a_. n p y •; Abs@aD == 1D, ¥D •• Union

8Cos@m p x - n p yD, Cos@m p x + n p yD<

terms2 = Cases@ST, Cos@n p x + a_. m p y •; Abs@aD == 1D, ¥D •• Union

8Cos@n p x - m p yD, Cos@n p x + m p yD<

The coefficients of these terms in ST are

terms1Coef = Coefficient@ST, terms1D;

terms2Coef = Coefficient@ST, terms2D;

The projection of terms1 onto Sin@m p xD Sin@n p yD can be obtained as

p1Coef = à à Sin@m p xD Sin@n p yD # â x ây & •ž terms1 •. trigRule

1 1

0 0

9 €€€€ , - €€€€ =
1 1
4 4

and the projection of terms2 onto Sin@n p xD Sin@m p yD can be obtained as

Chapter8.nb 297

p2Coef = à à Sin@n p xD Sin@m p yD # â x ây & •ž terms2 •. trigRule

1 1

0 0

9 €€€€ , - €€€€ =
1 1
4 4

The forcing term in ST is

f0 = ST •. Cos@_D -> 0

1
€€€€€€€ EI T1 s2 F@x, yD
24

Then, the solvability conditions demand that ST be orthogonal to every linear eigenfunction, modeshapes; that is,

SC1 = Expand@24 Hint@Sin@m p xD Sin@n p yD f0, 8x, 0, 1<, 8y, 0, 1<D + p1Coef.terms1Coef •.
symbolRule2 ••. intRule3 •. int -> IntegrateLD == 0;

SC2 = Expand@24 Hint@Sin@n p xD Sin@m p yD f0, 8x, 0, 1<, 8y, 0, 1<D + p2Coef.terms2Coef •.
symbolRule2 ••. intRule3 •. int -> IntegrateLD == 0;

They agree with the results obtained in the previous section:

8SC1, SC2< === 8SCond1, SCond2<

True
Chapter 9
Higher Approximations of Continuous
Systems Having Two-to-One Internal
Resonances
In this chapter, we use the method of multiple scales to determine second-order uniform asymptotic expansions of the
solutions of continuous systems with quadratic and cubic nonlinearities possessing two-to-one internal resonances. Higher-
order treatments of such systems lead to inconsistent results if the time derivatives in their governing equations are
expressed in second-order rather than first-order form. Therefore, we express the time derivatives in the governing equations
in first-order form before treating them with the method of multiple scales. We describe the methodology using two examples:
two-mode interactions in hinged-hinged buckled beams and four-mode interactions in suspended cables.

à Preliminaries
Off@General::spell1, Integrate::generD

To determine second-order uniform asymptotic expansions of the solutions of buckled beams and suspended cables by
using the method of multiple scales, taking into account two-to-one internal resonances, we introduce the three time scales
T0 = t, T1 = e t, and T2 = e2 t, where e is a small nondimensional parameter and the order of the amplitude of oscillations.
Moreover, we symbolize these scales according to

Needs@"Utilities`Notation`"D

Symbolize@T0 D; Symbolize@T1D; Symbolize@T2 D;

timeScales = 8T0 , T1, T2 <;

In terms of the time scales T0, T1 , and T2, the time derivatives can be expressed as

dt@0D@expr_D := expr; dt@1D@expr_D := SumAei D@expr, timeScales@@i + 1DDD, 8i, 0, 2<E;

To speed up manipulations of the integrals involved in the governing equations with Mathematica, we introduce the rules

intRule1 = 8int@fun_, arg2__D :> int@Expand@funD, arg2D<;

intRule2 = 8int@a_ + b_, arg2__D :> int@a, arg2D + int@b, arg2D,
int@e n_. fun_, arg2__D :> e n int@fun, arg2D,
int@a_ fun_, a1_, b1___D :> a int@fun, a1, b1D •; FreeQ@a, First@a1DD,
int@int@a1_, a2_D fun_., a3__D :> int@a1, a2D int@fun, a3D •; FreeQ@Rest@a2D, First@a2DD<;
298 Chapter9.nb

In the ensuing analysis, we express the solutions in terms of complex-variable functions and hence we introduce the complex-
conjugate rule

conjugateRule = 9As_ :> As , As_ :> As, Complex@m_, n_D :> Complex@m, -nD=;
• •

To display the outputs in easily read expressions, we introduce the rules

displayRule = 9Derivative@a_, b__DAw_i__ E@_, T0 , __D :>

#2@@1DD-1 &, 8b<E; arg1 =!= 1, arg1, ""E,

SequenceFormAIfAarg1 = Times žž MapIndexedAD #1
wi SequenceForm žž Table@"'",8a<DE,
Derivative@a__D@Ai__D@__D :> SequenceFormATimes žž MapIndexedAD#1
#2@@1DD &, 8a<E, Ai E,

Derivative@a__D@Ai__D@__D :> SequenceFormATimes žž MapIndexedAD#1

#2@@1DD &, 8a<E, Ai E,
• •

w_i__ @_, T0, __D -> wi , Ai__@__D -> Ai , Ai__@__D -> Ai ,

• •

Exp@a_. + b_. Complex@0, m_D T0 + c_. Complex@0, n_D T0D -> Exp@a + Hm * b + n * cL I T0 D=;

à 9.1 Two-Mode Interactions in Buckled Beams

In this section, we consider the nonlinear response of a hinged-hinged buckled beam possessing a two-to-one internal
resonance to a principal parametric resonance of the higher mode. The analysis assumes a unimodal static buckled deflec-
tion. The nondimensional equation of motion is (Section 8.1.1)

i 1 y ²
EOM = ut,t + ux,x,x,x + P k ux,x - 2 b2 a j
jà ux f¢k @xD âxz
z f k @xD ==
k 0 {
i y i y
-c ut + b a j
jà u2x âxz z f²k @xD + 2 b a j
jà ux f¢k@xD â xz
z ux,x +
1 1

k 0 { k 0 {
i y
ajjà u2x âxz z ux,x + Cos@t WD f ux,x + b Cos@t WD f f²k @xD •. u m__ :> ¶ m u@x, tD;
1

k 0 {

and the associated boundary conditions are

BC = 9u@0, tD == 0, uH2,0L @0, tD == 0, u@1, tD == 0, uH2,0L@1, tD == 0=;

where Pk is the kth critical Euler buckling load, fk is the kth buckling mode shape, and b is the nondimensional buckling level.

Because the two-to-one internal resonance is activated by the quadratic nonlinearities, as indicated in Chapter 5, second-
and higher-order treatments of such systems lead to inconsistent results if the time derivatives in their governing equations
are expressed in second-order rather than first-order form. Therefore, we express the time derivatives in first-order form by
letting v@x, tD == ¶t u@x, tD, adding it to EOM, and obtaining

EOM1 = 8¶t u@x, tD - v@x, tD == 0, EOM •. ¶t,t u@x, tD -> ¶t v@x, tD<;

We use the method of multiple scales to directly attack EOM1 and BC. To transform the time derivatives in EOM1 in terms of
the scales T0, T1 , and T2, we define
Chapter9.nb 299

multiScales = 8u@x_, tD -> u@x, T0 , T1, T2 D, v@x, tD -> v@x, T0 , T1, T2D,
Derivative@m_, n_D@u_D@x_, tD -> dt@nD@D@u@x, T0, T1 , T2 D, 8x, m<DD, t -> T0<;

Then, we seek a second-order approximate solution of EOM1 and BC in the form

solRule = h : u È v -> IEvaluateASumAej hj @#1, #2, #3, #4D, 8j, 3<EE &M;

where #1 stands for x and #2, #3, and #4 stand for T0 , T1, and T2 , respectively.

We consider the case of a two-to-one internal resonance between the lowest two modes and a principal parametric resonance
of the second mode; that is,

omgList = 8w1 , w2<;

ResonanceConds = 8w2 == 2 w1 + e s1 , W == 2 w2 + e s2 <;

where the si are detuning parameters that describe the nearness of the resonances. Because the influence of the two-to-one
internal resonance appears at O(e), we scale the damping and forcing terms as

scaling = 8c -> 2 e m, f -> e f<;

Substituting multiScales, solRule, and scaling into EOM1, expanding the result for small e, and discarding terms of order
higher than e3, we obtain

eq91a =
HEOM1 •. Integrate -> int •. multiScales •. solRule •. scaling •. intRule1 ••. intRule2 ••
TrigToExp •• ExpandAllL •. en_•;n>3 -> 0;

Equating coefficients of like powers of e, we obtain

eqEps = Thread@CoefficientList@Subtract žž #, eD == 0D & •ž eq91a •• Transpose •• Rest;

eqEps •. displayRule

99D0 u1 - v1 == 0, u'''' + D0 v1 + Hu''

1 L Pk - €€€€ b E f f²k @xD -
1 -IT0 W
1
2

€€€€ b EI T0 W f f²k @xD - 2 b2 a int@Hu'1 L f¢k @xD, 8x, 0, 1<D f²k @xD == 0=,
1
2
9D0 u2 + D1u1 - v2 == 0, - €€€€ E-I T0 W f Hu''1 L - €€€€ E f Hu''1 L-
1 1 I T0 W

2 b a int@Hu'1L f¢k @xD, 8x, 0, 1<D Hu'' 1 L + u2 + 2 m HD0u1 L + D0 v2 + D1 v1 + Hu''

2 L Pk -
2 2
''''

b a intBHu'1L , 8x, 0, 1<F f²k @xD - 2 b2 a int@Hu'2 L f¢k @xD, 8x, 0, 1<D f²k @xD == 0=,
2

9D0 u3 + D1u2 + D2 u1 - v3 == 0, -a intBHu'1 L , 8x, 0, 1<F Hu''

1 L-
2

2 b a int@Hu'2L f¢k @xD, 8x, 0, 1<D Hu''

1 L - €€€€ E f Hu''
2 L - €€€€ E f Hu''
2 L-
1 -I T0 W 1 I T0 W

2 b a int@Hu'1L f¢k @xD, 8x, 0, 1<D Hu''

2 L + u3 + 2 m HD0u2 L + D0 v3 + 2 m HD1u1 L + D1 v2 + D2 v1 +
2 2
''''

Hu''
3 L Pk - 2 b a int@Hu1 L Hu2L, 8x, 0, 1<D fk @xD - 2 b a int@Hu3L fk @xD, 8x, 0, 1<D fk @xD == 0==
' ' ² 2 ' ¢ ²
300 Chapter9.nb

Ÿ 9.1.1 First-Order Solution

The homogeneous part of the first-order equations, eqEps[[1]], can be written as

linearSys = #@@1DD & •ž eqEps@@1DD •. f -> 0;

linearSys •. displayRule

8D0 u1 - v1, u''''

1 + D0 v1 + Hu''
1 L Pk - 2 b a int@Hu1L fk @xD, 8x, 0, 1<D fk @xD<
2 ' ¢ ²

Because in the presence of damping all of the modes that are not directly or indirectly excited decay with time (Nayfeh and
Mook, 1979), the solution of eqEps[[1]] is taken to consist of the lowest two modes; that is,

solForm = Y@xD Exp@I W T0D + Y@xD Exp@-I W T0 D +

SumAAi @T1, T2 D Fi@xD Exp@I wi T0 D + Ai@T1 , T2D Fi @xD Exp@-I wi T0D, 8i, 2<E;
•

sol1 = 8u1 -> Function@8x, T0 , T1 , T2<, solForm •• EvaluateD,

v1 -> Function@8x, T0, T1 , T2 <, D@solForm, T0D •• EvaluateD<

8u1 ® Function@8x, T0, T1 , T2 <, E-IT0 W Y@xD + EI T0 W Y@xD + EI T0 w1 A1@T1 , T2 D F1 @xD +

EI T0 w2 A2@T1 , T2D F2 @xD + E-I T0 w1 F1@xD A1 @T1, T2 D + E-I T0 w2 F2@xD A2 @T1, T2 DD,
• •

v1 ® Function@8x, T0, T1 , T2 <, -I E-I T0 W W Y@xD + I EI T0 W W Y@xD + I EIT0 w1 w1 A1@T1 , T2D F1 @xD +
I EI T0 w2 w2 A2 @T1, T2 D F2@xD - I E-IT0 w1 w1 F1@xD A1 @T1 , T2 D - I E-I T0 w2 w2 F2 @xD A2@T1 , T2DD<
• •

where the Fi are the eigenmodes, which satisfy the orthonormality condition Ù0 Fn @xDFm @xD â x = dn,m, and Y is uniquely
1

determined by the following boundary-value problem:

eq91b = CoefficientA
eqEps@@1, 2, 1DD •. Isol1 •. 9Ai_ -> H0 &L, Ai_ -> H0 &L=M •. intRule1 ••. intRule2,
•

EI W T0 E == 0

-W2 Y@xD + Pk Y² @xD - €€€€ b f f²k @xD - 2 b2 a int@Y¢@xD f¢k @xD, 8x, 0, 1<D f²k @xD + YH4L @xD == 0
1
2

bc91b = BC •. u -> Y

8Y@0, tD == 0, YH2,0L @0, tD == 0, Y@1, tD == 0, YH2,0L@1, tD == 0<

In what follows, we consider nonlinear vibrations of a first-mode buckled beam; hence, the critical buckling load is

bucklingLoad = P k -> p 2 ;

We note that, as the buckling level increases from zero, the symmetric mode is the first mode. However, as the buckling level
exceeds the first cross-over value, the symmetric mode becomes the second mode. Hence, the second natural frequency w2
depends on the buckling level. For buckling levels greater than the cross-over value, the first static buckled deflection and
the first two normalized mode shapes and associated natural frequencies are
Chapter9.nb 301

• !! •!!
modeshapes = 9f k -> HSin@p #D &L, F1 -> I 2 Sin@2 p #D &M, F2 -> I 2 Sin@p #D &M=;
• !!! • !!!
frequencies = 9w2 -> b p 2 a , w1 -> 2 3 p 2 =;

The buckling level at which w2 = 2 w1 is

values1 = Solve@w2 == 2 w1 •. frequencies, bD@@1DD

• !!!
:b ® €€€€€€€€
€€€€€€ >
• !!!
4 3
a

Substituting the bucklingLoad and modeshapes into eq91b yields

eq91c = eq91b •. bucklingLoad •. modeshapes

1
€€€€ b f p 2 Sin@p xD +

2 b2 p 2 a int@p Cos@p xD Y¢ @xD, 8x, 0, 1<D Sin@p xD - W2 Y@xD + p2 Y² @xD + YH4L@xD == 0

2

Inspection of eq91c and bc91b suggests that Y@xD has the form

Ysol = Y -> HG Sin@p #D &L;

Substituting Ysol into eq91c, collecting the coefficient of Sin@p xD, solving the resulting equation for G, and simplifying it
using frequencies and values1, we obtain

GRule = Solve@Coefficient@eq91c@@1DD •. Ysol •. int -> Integrate, Sin@p xDD == 0, GD@@1DD •.

W -> 2 w2 •. frequencies •. values1

:G ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
• !!! 2 • !!!
f
24 3 p a

Because the buckled-beam problem is self-adjoint, we have

adjoint = 88-I w1 F1 @xD, F1@xD<, 8-I w2 F2 @xD, F2 @xD<<;

whose complex conjugate is

adjointC = adjoint •. conjugateRule

88I w1 F1@xD, F1 @xD<, 8I w2 F2 @xD, F2@xD<<

Ÿ 9.1.2 Second-Order Solution

Substituting sol1, Ysol, and some of the parameter values into eqEps[[2]] yields
302 Chapter9.nb

order2Eq =
HlinearSys •. u_1 -> u2 L == HHlinearSys •. u_1 -> u2 L - H#@@1DD & •ž eqEps@@2DDL •. sol1 •.
Ysol •. GRule •. intRule1 ••. intRule2 •. modeshapes •.
int -> Integrate •• ExpandL •. bucklingLoad •. values1 •• Thread;
order2Eq •. displayRule

:D0 u2 - v2 == - • !!!
2 EIT0 w2 HD1 A2L Sin@p xD - • !2! E-I T0 w2 HD1 A2L Sin@p xD -
•

• !!!
2 EI T0 w1 HD1 A1L Sin@2 p xD - • !2! E-I T0 w1 HD1 A1L Sin@2 p xD,
•

p 2 Hu''
2 L + u2
''''
+ D0v2 - 96 int@Hu'2L f¢k @xD, 8x, 0, 1<D f²k @xD ==

• !!! €€€€€€€€
• !!!€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• !!! • ! €€€€€€€€
! €€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• !!! • !!!€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• !!! • !!!€€€€€€€€€€€ -
f2 Sin@p xD E-2 I T0 W f2 Sin@p xD E2 I T0 W f2 Sin@p xD I E-I T0 W f m W Sin@p xD
- €€€€€€€€€€€€€€€€ €€€€€€€€
16 3 a 32 3 a 32 3 a 12 3 p 2 a

€€€€€€€€€€€€€€€€€ - €€€€ • !2! EI T0 H-W+w1L f p 2 Sin@2 p xD A1 - €€€€ • !2! EI T0 HW+w1L f p 2 Sin@2 p xD A1 -

• !!! 2 • !!!
I EI T0 W f m W Sin@p xD 8 8
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• !!! 2I T0 w1 4 • ! ! • !!! • !!!

12 3 p a 3 3
a Sin@p xD A21 - 2 EI T0 H-W+w2 L f p 2 Sin@p xD A2 - 2 EI T0 HW+w2 L f p 2 Sin@p xD A2 -
32 • !!! 3 EIT0 Hw1+w2 L p 4 • !!! a Sin@2 p xD A1 A2 - 12 • !3! E2 I T0 w2 p 4 • !!!
16 3 E p

• !!! IT0 w1 • !!! -I T0 w1

a Sin@p xD A22 -
HD1 A1L Sin@2 p xD w1 + I 2 E HD1A1 L Sin@2 p xD w1 -
•

2 I • !!! 2 EI T0 w1 m Sin@2 p xD A1 w1 - I • !!! 2 EIT0 w2 HD1 A2L Sin@p xD w2 + I • !!! 2 E-I T0 w2 HD1A2 L Sin@p xD w2 -
I 2 E
•

• !!! 8 •!!
2 I 2 EI T0 w2 m Sin@p xD A2 w2 - €€€€ 2 EI T0 H-W-w1L f p 2 Sin@2 p xD A1 -
•

8 • ! ! I T0 HW-w1L • !!! • !!!

3
• •
€€€€ 2 E f p Sin@2 p xD A1 - 32 3 p4 a Sin@p xD A1 A1 -
2

32 • !!! 3 EIT0 H-w1+w2 L p 4 • !!! a Sin@2 p xD A2 A1 + 2 I • !2! E-I T0 w1 m Sin@2 p xD w1 A1 -

3
• •

16 • !!! 3 E-2 I T0 w1 p 4 • !a! Sin@p xD A1 - • !!! 2 EI T0 H-W-w2 L f p 2 Sin@p xD A2 -

•2 •
• !!! I T0 HW-w2L • !!! • !!!
f p 2 Sin@p xD A2 - 32 3 EI T0 Hw1-w2 L p 4 a Sin@2 p xD A1 A2 -
• •

24 • !!! 3 p 4 • !!! a Sin@p xD A A + 2 I • !!!

2 E
• •

• !!! • !!! •!! • !!!

2 E-I T0 w2 m Sin@p xD w A -
32 3 EIT0 H-w1-w2 L p 4 a Sin@2 p xD A1 A2 - 12 3 E-2 I T0 w2 p 4 a Sin@p xD A2>
2 2 2 2
• • •2

In order to collect the terms that may lead to secular terms from the right-hand sides of order2Eq, we define the rules

OmgRule = Solve@ResonanceConds, Complement@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž omgList

98w2 ® e s1 + 2 w1, W ® 2 e s1 + e s2 + 4 w1 <, 9w1 ® €€€€ H-e s1 + w2 L, W ® e s2 + 2 w2==

1
2

expRule1@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

Collecting the terms that may lead to secular terms, the terms proportional to EI wi T0 , we have

ST11 = CoefficientA#@@2DD & •ž order2Eq •. expRule1@1D, EI w1 T0 E;

ST11 •. displayRule
• !!!
9- 2 HD1 A1 L Sin@2 p xD,
• !!! • !!! •! ! • !!!
-I 2 HD1 A1L Sin@2 p xD w1 - 2 I 2 m Sin@2 p xD A1 w1 - 32 3 EI T1 s1 p 4 a Sin@2 p xD A2 A1=
•
Chapter9.nb 303

ST12 = CoefficientA#@@2DD & •ž order2Eq •. expRule1@2D, EI w2 T0 E;

ST12 •. displayRule
• !!! • !!! •!!
9- 2 HD1 A2 L Sin@p xD, -16 3 E-I T1 s1 p 4 a Sin@p xD A21 -
• !!! •!! • !!!
I 2 HD1A2 L Sin@p xD w2 - 2 I 2 m Sin@p xD A2 w2 - 2 EI T1 s2 f p 2 Sin@p xD A2 =
•

Demanding that ST11 be orthogonal to the adjointC[[1]], we obtain the solvability condition

SCond11 =
SolveAint@adjointC@@1DD.ST11, 8x, 0, 1<D == 0 •. intRule1 ••. intRule2 •. modeshapes •.
H1,0L
int -> Integrate, A1 @T1 , T2DE@@1DD •• ExpandAll

8 I • !!! 6 EIT1 s1 p 4 •!a! A2 @T1, T2 D A1@T1 , T2 D

:A1 @T1 , T2 D ® -m A1 @T1, T2 D + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
H1,0L
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w1

Demanding that ST12 be orthogonal to the adjointC[[2]], we obtain the solvability condition

SCond12 =
SolveAint@adjointC@@2DD.ST12, 8x, 0, 1<D == 0 •. intRule1 ••. intRule2 •. modeshapes •.
int -> Integrate, AH1,0L@T1 , T2DE@@1DD •• ExpandAll

• !!! •! !
2

4 I 6 E-IT1 s1 p 4 a A1 @T1, T2 D2 I EI T1 s2 f p 2 A2 @T1, T2 D

:A2 @T1 , T2 D ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€ - m A2 @T1 , T2D + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ >
H1,0L
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
w2 2 w2

Hence, we have the solvability conditions

SCond1 = Join@SCond11, SCond12D;

whose complex conjugates are

ccSCond1 = SCond1 •. conjugateRule;

To remove D1A1 and D1 A2 from the right-hand sides of order2Eq, we first define the rules

sigRule = Solve@ResonanceConds, 8s1, s2 <D@@1DD

9s1 ® - €€€€€€€€1€€€€€€€€€2€€€ , s2 ® - €€€€€€€€€€€€€€€€€€€€ =

2w -w -W + 2 w2
e e

expRule2 = Exp@a_D :> Exp@a •. sigRule •. T1 -> e T0 •• ExpandD;

Substituting SCond1, ccSCond1, and expRule2 into order2Eq, we obtain

304 Chapter9.nb

order2Eqm = order2Eq •. SCond1 •. ccSCond1 •. expRule2 •• ExpandAll;

order2Eqm •. displayRule
•!! • !!!
9D0 u2 - v2 == 2 EI T0 w1 m Sin@2 p xD A1 + 2 EIT0 w2 m Sin@p xD A2 -
• !!! • !!!
I EI T0 H-W+w2 L f p 2 Sin@p xD A2
•!!
8 I 3 E2 I T0 w1 p 4 a Sin@p xD A21
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +
w2
• !!! • !!!
2 w2

• !!! -IT0 w1 16 I 3 EI T0 H-w1+w2 L p 4 a Sin@2 p xD A2 A1

•
•
2 E m Sin@2 p xD A1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

• !!! • !!!
w1

8 I 3 E-2 I T0 w1 p 4 a Sin@p xD A1 • !!! -I T0 w2

• 2
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ + 2 E m Sin@p xD A2 +

16 I • !!! 3 EI T0 Hw1 -w2L p 4 • !a! Sin@2 p xD A1 A2

w2
I EI T0 HW-w2 L f p 2 Sin@p xD A2
• •

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
w1 2 w2
p 2 Hu''
2 L + u2
''''
+ D0v2 - 96 int@Hu'2L f¢k @xD, 8x, 0, 1<D f²k @xD ==

• !!! • !!! • !!! • ! ! • !!! • !!! • !!! • !!!

f2 Sin@p xD E-2 I T0 W f2 Sin@p xD E2 I T0 W f2 Sin@p xD I E-I T0 W f m W Sin@p xD
- €€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€ -
a 32 3 a a 12 3 p 2 a
8 •!! 8 •!!
16 3 32 3

€€€€€€€€€€€€€€€€€ - €€€€ 2 EI T0 H-W+w1L f p 2 Sin@2 p xD A1 - €€€€ 2 EI T0 HW+w1L f p 2 Sin@2 p xD A1 -

• !!! 2 • !!!
I EI T0 W f m W Sin@p xD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
12 3 p a 3 3

• !!! 2I T0 w1 4 • ! ! EI T0 H-W+w2 L f p 2 Sin@p xD A2 • ! ! I T0 H-W+w2L

• !!!
8 3 E p a Sin@p xD A21 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 2 E f p 2 Sin@p xD A2 -

• !!! I T0 HW+w2L • !!! I T0 Hw1+w2 L 4 • !!!

2

• !!! •!! • !!!

2
2 E f p Sin@p xD A2 - 32 3 E p a Sin@2 p xD A1 A2 -
12 3 E2I T0 w2 p 4 a Sin@p xD A22 - I 2 EI T0 w1 m Sin@2 p xD A1 w1 -
• !!! 8 •!!
I 2 EIT0 w2 m Sin@p xD A2 w2 - €€€€ 2 EI T0 H-W-w1 L f p 2 Sin@2 p xD A1 -
•

8 • ! ! I T0 HW-w1L • !!! • !!!

3
• •
f p Sin@2 p xD A1 - 32 3 p4 a Sin@p xD A1 A1 -
2
€€€€ 2 E
• !!! • !!! •!!
16 3 EIT0 H-w1+w2 L p 4 a Sin@2 p xD A2 A1 + I 2 E-I T0 w1 m Sin@2 p xD w1 A1 -
3
• •

• !!! •!! •2 • !!! EI T0 HW-w2 L f p 2 Sin@p xD A2

•
8 3 E-2 I T0 w1 p 4 a Sin@p xD A1 - 2 EI T0 H-W-w2 L f p 2 Sin@p xD A2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• !!!
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!! I T0 HW-w2L • !!! I T0 Hw1-w2 L 4 • !!!

2
• •
• !!! 4 • !!! • !!!
2
2 E f p Sin@p xD A2 - 16 3 E p a Sin@2 p xD A1 A2 -
• •

32 • !!!
3 EIT0 H-w1-w2 L p 4 • !!!
a Sin@2 p xD A1 A2 - 12 • !3! E-2 I T0 w2 p 4 • !!!
-I T0 w2
24 3 p a Sin@p xD A2 A2 + I 2 E m Sin@p xD w2 A2 -
a Sin@p xD A2=
• • •2

We use the method of undetermined coefficients to determine the solution of order2Eqm and their associated boundary
conditions. To implement this, we first determine all possible forms of the terms on the right-hand sides of order2Eqm as
follows:
Chapter9.nb 305

sol2Form = TableACases@order2Eqm@@k, 2DD, #D & •ž 9_ Ea_ T0 +b_. f_@c_ xD -> EaT0 +b f@c xD,
a_ f_@b_ xD :> f@b xD •; FreeQ@a, T0 D=, 8k, 2<E •• Flatten •• Union

8Sin@p xD, E-I T0 W Sin@p xD, EI T0 W Sin@p xD, E-2 I T0 W Sin@p xD, E2 I T0 W Sin@p xD,
E-2 I T0 w1 Sin@p xD, E2 I T0 w1 Sin@p xD, E-IT0 w2 Sin@p xD, EI T0 w2 Sin@p xD, E-2 I T0 w2 Sin@p xD,
E2 I T0 w2 Sin@p xD, E-I T0 W-IT0 w2 Sin@p xD, EI T0 W-I T0 w2 Sin@p xD, E-I T0 W+I T0 w2 Sin@p xD,
EI T0 W+I T0 w2 Sin@p xD, E-I T0 w1 Sin@2 p xD, EI T0 w1 Sin@2 p xD, E-I T0 W-I T0 w1 Sin@2 p xD,
EI T0 W-I T0 w1 Sin@2 p xD, E-I T0 W+I T0 w1 Sin@2 p xD, EI T0 W+I T0 w1 Sin@2 p xD, E-IT0 w1-I T0 w2 Sin@2 p xD,
EI T0 w1 -I T0 w2 Sin@2 p xD, E-I T0 w1+I T0 w2 Sin@2 p xD, EI T0 w1+I T0 w2 Sin@2 p xD<

where all the spatial dependent functions satisfy the hinged-hinged boundary conditions. We note that sol2Form consists
of (a) the terms that are proportional to E-I w1 T0 , EI w1 T0 , E-I w2 T0 , EI w2 T0 and (b) the remainder of the terms. The
positions of the former terms, which are orthogonal to the adjoint, in sol2Form are

pos@1D =
HPosition@# sol2Form •. expRule1@1D •. Exp@_. + _ T0D -> 0, a_ •; a =!= 0, 1D •• Flatten ••
RestL & •ž 9E-I w1 T0 , EI w1 T0 =

8817, 24<, 816, 23<<

pos@2D =
HPosition@# sol2Form •. expRule1@2D •. Exp@_. + _ T0D -> 0, a_ •; a =!= 0, 1D •• Flatten ••
RestL & •ž 9E-I w2 T0 , EI w2 T0 =

887, 9, 13<, 86, 8, 14<<

Next, we seek the solution of u 2 and v2 in two parts. First, we consider the part of sol2Form that is not related to the secular
terms

sol2Forma = Delete@sol2Form, 8#< & •ž Flatten@Array@pos, 2DDD

8Sin@p xD, E-I T0 W Sin@p xD, EI T0 W Sin@p xD, E-2 I T0 W Sin@p xD, E2 I T0 W Sin@p xD,
E-2 I T0 w2 Sin@p xD, E2 I T0 w2 Sin@p xD, E-IT0 W-I T0 w2 Sin@p xD, EI T0 W+I T0 w2 Sin@p xD,
E-I T0 W-I T0 w1 Sin@2 p xD, EI T0 W-I T0 w1 Sin@2 p xD, E-I T0 W+I T0 w1 Sin@2 p xD,
EI T0 W+I T0 w1 Sin@2 p xD, E-I T0 w1 -I T0 w2 Sin@2 p xD, EI T0 w1 +IT0 w2 Sin@2 p xD<

and define the undetermined coefficients as

uSymbola = Table@yi, 8i, Length@sol2FormaD<D;

vSymbola = uSymbola •. y -> h;

The general solution in terms of the sol2Forma can be written as

sol2a = 8u2 -> Function@8x, T0 , T1 , T2<, uSymbola.sol2Forma •• EvaluateD,

v2 -> Function@8x, T0, T1 , T2 <, vSymbola.sol2Forma •• EvaluateD<;

Second, we consider the part of sol2Form that is related to the secular terms
306 Chapter9.nb

sol2Formb1 = sol2Form@@pos@1D@@1DDDD

8EI T0 w1 Sin@2 p xD, E-I T0 w1+I T0 w2 Sin@2 p xD<

sol2Formb2 = sol2Form@@pos@2D@@1DDDD

8E2 I T0 w1 Sin@p xD, EI T0 w2 Sin@p xD, EI T0 W-I T0 w2 Sin@p xD<

and define the undetermined coefficients as

uSymbolb1 = Table@ys1,i, 8i, Length@sol2Formb1D<D;

uSymbolb2 = Table@ys2,i, 8i, Length@sol2Formb2D<D;

where the undetermined coefficients for v2 are related to uSymbolb1 and uSymbolb2 due to the orthogonality condition.
Therefore, the general solution in terms of the sol2Formb1 and sol2Formb2 can be written as

sol2b =
8u2 -> Function@8x, T0, T1 , T2 <, uSymbolb1.sol2Formb1 + uSymbolb2.sol2Formb2 •• EvaluateD,
v2 -> Function@8x, T0, T1 , T2 <,
-I w1 uSymbolb1.sol2Formb1 - I w2 uSymbolb2.sol2Formb2 •• EvaluateD<;

The total solution is the sum of sol2a, sol2b, and the complex conjugate of sol2b.

Next, we solve for these undetermined coefficients. Substituting sol2a into order2Eqm and equating the coefficients of
sol2Forma, we have

algEqa =
Flatten@Coefficient@Subtract žž # •. sol2a •. intRule1 ••. intRule2 •. modeshapes •. int ->
Integrate, sol2FormaD & •ž order2EqmD == 0 •. Exp@_ T0 + _.D -> 0 •• Thread;

Substituting some of the parameter values into algEqa and solving for uSymbola and vSymbola, we obtain
Chapter9.nb 307

symbolaRule =
Solve@algEqa, uSymbola~ Join~vSymbolaD@@1DD •. W -> 2 w2 •. frequencies •. values1
• !!! 2 • !!! • !!!
3 f + 1536 3 p 4 a A1 @T1, T2 D A1@T1 , T2 D + 1152 3 p4 a A2@T1 , T2 D A2@T1 , T2D
:h1 ® 0, y1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
• •

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ ,
2304 p 4 a

• !!! 2 • !!! • !!! 2 • !!! • !!! •!!

fm fm I f2 I f2
h2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , h3 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , h4 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€ , h5 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

• !!! • • !!!
9 3 p a 9 3 p a 1440 p 2 a 1440 p 2 a

h6 ® -2 I p 2 a A2 @T1, T2D , h7 ® 2 I p 2 a A2 @T1, T2 D2 , y2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€ ,

• !!!
2 I fm

• !!! •
216 p 4 a
a A2 @T1, T2D
2

216 p 4 • !!! 23040 • !!! 3 p4 • !!! 23040 • !3! p 4 • !a! 4 • !3!

Ifm f2 f2
y3 ® €€€€€€€€€€€€€€€€€€€€€€€€€ , y4 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ , y5 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€ , y6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

•!!
a a

a A2@T1 , T2 D2
€€€€€€€€€€€€€€€€€€€€€ , h8 ® - €€€€€€€ I $ %%€€€€%%% f A2@T1 , T2 D, h9 ® €€€€€€€ I $ %€€€€ %%%% f A2 @T1 , T2 D,
• !!!
1 3 • 1 3
y7 ® €€€€€€€€€€€€€€€€
16 2 16 2

5 I f A1@T1 , T2 D I f A1 @T1, T2 D I f A1 @T1, T2D 5 I f A1@T1 , T2 D

4 3
• •

• !!! •!! • !!! • !!!

h10 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , h11 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , h12 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , h13 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

f A2@T1 , T2 D
9 6 6 6 9 6

h14 ® -6 I p 2 • !!! a A1 @T1, T2 D A2@T1 , T2 D, h15 ® 6 I p 2 • !!! a A1 @T1, T2 D A2@T1 , T2 D, y8 ® €€€€€€€€€€€€€€€€

•

• !!!
• •
€€€€€€€€€€€€€ ,
192 2 p 2
f A2 @T1, T2 D f A1@T1 , T2 D f A1 @T1, T2D f A1@T1 , T2 D
• •

• €€€€€€€€
!!! €2€€€€ , y10 ® €€€€€€€€€€€€€€€€ • !!! • !!! • !!!
y9 ® €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€ , y11 ® €€€€€€€€€€€€€€€€ €€€€€€€€€€€€€ , y12 ® €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ ,

• !!! • • !!!
192 2 p 54 2 p 2 18 2 p 2 18 2 p 2
f A1 @T1, T2 D a A1 @T1, T2D A2 @T1, T2 D a A1 @T1, T2 D A2@T1 , T2 D
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
•

• !!! 2 •! ! •!!
y13 ® €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ , y14 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , y15 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
54 2 p 3 3

Substituting sol2b into either of the two equations in order2Eqm, equating the coefficients of sol2Formb1 and sol2Formb2,
and solving for uSymbolb1 and uSymbolb2, we have

symbolbRule =
Solve@Coefficient@Subtract žž order2Eqm@@1DD •. sol2b, sol2Formb1~Join~ sol2Formb2D ==
0 •. Exp@_ T0 + _.D -> 0 •• Thread,
uSymbolb1~ Join~uSymbolb2D@@1DD •. W -> 2 w2 •. frequencies •. values1

I m A1@T1 , T2 D 2 • !!!a A2 @T1, T2D A1 @T1, T2 D

:ys1,1
•

• !!! • !!!
® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , ys1,2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

• !!!
2 6 p2 3
I m A2@T1 , T2 D f A2@T1 , T2 D a A1@T1 , T2D2
® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , ys2,3 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , ys2,1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
•

• !!! 2 • !!! 2 • !!!

ys2,2
4 6 p 96 2 p 4 3

Substituting these symbol values into sol2a and sol2b yields

sol2aForm = 8u2 @x, T0, T1 , T2 D, v2@x, T0 , T1 , T2D< •. sol2a •. symbolaRule;

sol2bForm = 8u2 @x, T0, T1 , T2 D, v2@x, T0 , T1 , T2D< •. sol2b •. symbolbRule;
308 Chapter9.nb

sol2 = 8u2 ® Function@8x, T0 , T1 , T2<,

sol2aForm@@1DD + sol2bForm@@1DD + Hsol2bForm@@1DD •. conjugateRuleL •• Expand ••
EvaluateD, v2 ® Function@8x, T0 , T1 , T2<, sol2aForm@@2DD + sol2bForm@@2DD +
Hsol2bForm@@2DD •. conjugateRuleL •• Expand •• EvaluateD<

:u2 ® FunctionB8x, T0, T1 , T2 <,

• ! ! • ! ! • !!! • !!! • ! ! • ! ! 216 p 4 • !!!

f2 Sin@p xD E-2I T0 W f2 Sin@p xD E2 I T0 W f2 Sin@p xD I E-I T0 W f m Sin@p xD
- €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ +
768 3 p 4 a 23040 3 p 4 a 23040 3 p 4 a a
E-I T0 W+I T0 w1 f Sin@2 p xD A1 @T1, T2 D EIT0 W+I T0 w1 f Sin@2 p xD A1@T1 , T2 D
• !!! • !!! • !!!
I EI T0 W f m Sin@p xD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ -

• !!!
216 p 4 a 18 2 p 2 54 2 p 2
I EI T0 w1 m Sin@2 p xD A1@T1 , T2D E2 I T0 w1 a Sin@p xD A1@T1 , T2 D2
•!! •!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ -
2 6 p2
E-I T0 W+I T0 w2 f Sin@p xD A2 @T1 , T2 D EI T0 W+I T0 w2 f Sin@p xD A2@T1 , T2 D
4 3

• !!! •!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!!
96 2 p 2 192 2 p 2
I EI T0 w2 m Sin@p xD A2@T1 , T2 D EI T0 w1+I T0 w2 a Sin@2 p xD A1@T1 , T2D A2 @T1, T2D
• !!! •!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

• !!!
4 6 p2 3
E2 IT0 w2 a Sin@p xD A2 @T1, T2D2 E-I T0 W-IT0 w1 f Sin@2 p xD A1@T1 , T2D
•

• !!! • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ +
54 2 p 2
EI T0 W-I T0 w1 f Sin@2 p xD A1 @T1 , T2 D I E-I T0 w1 m Sin@2 p xD A1 @T1, T2D
4 3
• •

• !!! • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!! • !!!
18 2 p 2 2 6 p2
2 a Sin@p xD A1@T1 , T2 D A1@T1 , T2D 2 E-I T0 w1 +I T0 w2 a Sin@2 p xD A2 @T1, T2D A1 @T1, T2 D
• •

• !!! • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ -

• !!!
3 3

E-2 IT0 w1 a Sin@p xD A1 @T1, T2 D E-I T0 W-I T0 w2 f Sin@p xD A2@T1 , T2D

• •

•!! • !!!
2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
192 2 p 2
EI T0 W-I T0 w2 f Sin@p xD A2 @T1, T2 D I E-I T0 w2 m Sin@p xD A2 @T1 , T2D
4 3
• •

96 • !!! 4 • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!!
2 p2 6 p2
2 EI T0 w1 -I T0 w2 a Sin@2 p xD A1 @T1, T2 D A2@T1 , T2 D 1 • !!! • ! !
€€€€€€€€€€€€€€€€€€€€ - €€€€ 3 a Sin@p xD A2@T1 , T2D A2 @T1, T2 D +
•

• !!!
•
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• !!! • !!!
3 2

E-I T0 w1 -I T0 w2 a Sin@2 p xD A1 @T1, T2 D A2@T1 , T2 D E-2 I T0 w2 a Sin@p xD A2@T1 , T2 D

€€€€€€€€€€€€€ F,
• • •

•!! • !!!
2
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 4 3

v2 ® FunctionB8x, T0, T1 , T2 <, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• ! ! •! !
I E-2 I T0 W f2 Sin@p xD I E2I T0 W f2 Sin@p xD
€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ -
1440 p 2 a 1440 p 2 a
I E-I T0 W+I T0 w1 f Sin@2 p xD A1 @T1, T2 D
• !!! 2 • !!! • !!! 2 • !!! • !!!
E-I T0 W f m Sin@p xD EI T0 W f m Sin@p xD
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ +
9 3 p a 9 3 p a
5 I EIT0 W+I T0 w1 f Sin@2 p xD A1@T1 , T2 D EI T0 w1 m Sin@2 p xD w1 A1@T1 , T2 D
6

9 • !!! 2 • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ +

I E2 IT0 w1 • !!!
6 6 p2

a Sin@p xD w2 A1@T1 , T2 D2
€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€ I $ %%€€€€%%% EI T0 W+I T0 w2 f Sin@p xD A2 @T1, T2 D -
• !!!
1 3
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 3 16 2
Chapter9.nb 309

I E-IT0 W+I T0 w2 f Sin@p xD w2 A2 @T1, T2D EI T0 w2 m Sin@p xD w2 A2@T1 , T2D

96 • !2! p 2 4 • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ +

I T0 w1 +I T0 w2 2 • !!!
a Sin@2 p xD A1@T1 , T2 D A2@T1 , T2D + 2 I E2 I T0 w2 p 2 • !!!
6 p2
p a Sin@p xD A2@T1 , T2 D2 -
5 I E-I T0 W-I T0 w1 f Sin@2 p xD A1@T1 , T2 D I EI T0 W-I T0 w1 f Sin@2 p xD A1 @T1 , T2D
6IE
• •

•!! •!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!!
9 6 6
E-I T0 w1 m Sin@2 p xD w1 A1 @T1, T2 D 2 I E-I T0 w1+I T0 w2 a Sin@2 p xD w1 A2 @T1, T2D A1 @T1, T2 D
• •

• !!! • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!!
2 6 p2 3

I E-2I T0 w1 a Sin@p xD w2 A1@T1 , T2D

€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€ I $ %%€€€€%%% E-I T0 W-I T0 w2 f Sin@p xD A2 @T1, T2 D +
•

• !!!
2
1 3 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
16 2

I EI T0 W-I T0 w2 f Sin@p xD w2 A2 @T1 , T2D E-I T0 w2 m Sin@p xD w2 A2@T1 , T2D

4 3
• •

• !!! • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!!
96 2 p 2 4 6 p2
2 I EIT0 w1-I T0 w2 a Sin@2 p xD w1 A1 @T1 , T2D A2 @T1, T2 D • !!!
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 6 I E-I T0 w1 -I T0 w2 p 2 a
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

•!!
Sin@2 p xD A1@T1 , T2 D A2 @T1 , T2D - 2 I E-2 I T0 w2 p 2 a Sin@p xD A2@T1 , T2D F>
3
• • • 2

Ÿ 9.1.3 Solvability Conditions

Substituting sol1, Ysol, sol2, and some of the parameter values into eqEps[[3]] yields

order3Eq = HlinearSys •. u_1 -> u3 L ==

HHlinearSys •. u_1 -> u3 L - H#@@1DD & •ž eqEps@@3DDL •. sol1 •. Ysol •. GRule •. sol2 •.
intRule1 ••. intRule2 •• ExpandL •. values1 •• Thread; •• Timing

821.141 Second, Null<

Collecting the terms that may lead to secular terms, the terms proportional to EI wi T0 , we have

ST21 =
CoefficientA#@@2DD & •ž order3Eq •. expRule1@1D, EI w1 T0 E •. modeshapes •. int -> Integrate;

ST22 =
CoefficientA#@@2DD & •ž order3Eq •. expRule1@2D, EI w2 T0 E •. modeshapes •. int -> Integrate;

Demanding that ST21 be orthogonal to the adjointC[[1]], we obtain the solvability condition
310 Chapter9.nb

SCond21 = SolveAHint@adjointC@@1DD.ST21, 8x, 0, 1<D •. intRule1 ••. intRule2L == 0,

H0,1L
A1 @T1 , T2DE@@1DD •. modeshapes •.
int -> Integrate •. SCond1 •. frequencies •. values1 •• ExpandAll;
SCond21 •. displayRule

€ - €€€€€€€€€€€€€€€€1€€€€ + I • !!!
:D2 A1 ® €€€€€€€€€€€€€€€€€€€€€€€€
• !!! 2 • !!! 2
13 I f2 A1 I m2 A •
3 p 2 a A21 A1 -

• !!! • •
648 3 p 4 3 p

$ %%€€€€%%% EI T1 s1 • !a! m A2 A1 - €€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€ >

•

• !!! €€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

•€€€€€€€€€€€€€€€€
!!
3 • 8 I p 2 a A1 A2 A2 31 I EI T1 s1+I T1 s2 f a A1 A2
2 3 72 2

Demanding that ST22 be orthogonal to the adjointC[[2]], we obtain the solvability condition

SCond22 = SolveAHint@adjointC@@2DD.ST22, 8x, 0, 1<D •. intRule1 ••. intRule2L == 0,

AH0,1L@T1 , T2DE@@1DD •. modeshapes •. int -> Integrate •.
SCond1 •. W -> 2 w2 •. frequencies •. values1 •• ExpandAll;
2

SCond22 •. displayRule

3 -I T1 s1 • !!!
:D2 A2 ® €€€€ $ %€€€€ %%%% E
•! ! •!!
1 19 I f2 A2 I m2 A
a m A21 - €€€€€€€€€€€€€€€€€€€€€€€€ €€€€ - €€€€€€€€€€€€€€€€2€€€€ -

• ! ! •2
4 2 9216 3 p 2 8 3 p2

€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€ I • !!!

3 p 2 a A22 A2>
• •

•!! • !!!
4 I p 2 a A1 A2 A1 31 I EI T1 s1 +I T1 s2 f a A1 5 EI T1 s2 f m A2 1 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 288 2 288 p 2 2

Because in the absence of damping the system is conservative, the modulation equations, solvability conditions, must
satisfy symmetry conditions. To check for these symmetries, we let

formList1a = 9A1 @T1 , T2 D A2@T1 , T2D, A1@T1 , T2D A2 @T1, T2 D A2 @T1 , T2 D=;
• • •

a1coefs = 8s11 , s12< -> Coefficient@H2 w1 IL SCond21@@1, 2DD, formList1aD •• Thread

• !!!
:s11 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ , s12 ® €€€€€€€€€€€€€€€€€€€€€€€ >
• !!! • !!!
31 EI T1 s1 +I T1 s2 f a w1 16 p 2 a w1
36 2 3

formList2a = 9A1 @T1 , T2 D A2@T1 , T2D A1 @T1, T2D, A1 @T1, T2 D =;

• • 2

a2coefs = 8s21 , s22< -> Coefficient@H2 w2 IL SCond22@@1, 2DD, formList2aD •• Thread

• !!!
:s21 ® €€€€€€€€€€€€€€€€€2€€€ , s22 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ >
•!! •!!
8 p2 a w 31 EIT1 s1+I T1 s2 f a w2
3 144 2

The symmetry condition demands that s12 = s21 , which is true because

s12 - s21 •. a1coefs •. a2coefs •. frequencies •. values1

0
Chapter9.nb 311

à 9.2 Four-Mode Interactions in Suspended Cables

We determine a second-order uniform asymptotic expansion of the three-dimensional response of a suspended cable to a
transverse harmonic distributed excitation of one of the first two vertical or horizontal modes when their frequencies are in
the ratio of either two-to-one or one-to-one and no other internal resonance is activated. Benedettini, Rega, and Alaggio
(1995) derived nondimensional equations governing three-dimensional finite-amplitude vibrations of a suspended homoge-
neous elastic cable excited by harmonic distributed vertical and horizontal loads. The time derivatives are expressed in
second-order form. As discussed in Chapter 5, second- and higher-order treatments of such equations lead to inconsistent
results unless the time derivatives are expressed in first-order rather than second-order form. Therefore, we rewrite these
equations as

di_Integer,j_Integer := If@i == j, 1, 0D

eq1 = Table@uj,t - vj == 0, 8j, 2<D;

eq2 = TableAvj,t + 2 mj vj - uj,x,x - a Hb y² @xD d1,j + uj,x,x L à ij b y¢ @xD u1,x + €€€€ Iu21,x + u22,x My
z âx ==
1

0 k
1
{
Pj @xD Cos@W t + tjD, 8j, 2<E;
2

EOM = Transpose@8eq1, eq2<D •. 9u_s_,m__ :> ¶ m us@x, tD, vs_ :> vs @x, tD= •• Flatten

:-v1@x, tD + u1 @x, tD == 0, 2 m1 v1@x, tD + v1 @x, tD - u1 @x, tD -

H0,1L H0,1L H2,0L

i 1i 2 y y
aj
jà jb y¢ @xD uH1,0L @x, tD + €€€€ Ju1 @x, tD + u2 @x, tD Nz â xzz Ib y² @xD + uH2,0L@x, tDM ==
H1,0L H1,0L

k 0 k { {
1 2
1 1

Cos@t W + t1 D P1 @xD, -v2 @x, tD + u2 @x, tD == 0, 2 m2 v2 @x, tD + v2 @x, tD - u2 @x, tD -

H0,1L H0,1L H2,0L
2

i
jà i 2 y y H2,0L
aj jb y¢ @xD uH1,0L@x, tD + €€€€ Ju1 @x, tD + u2 @x, tD Nz â xzz u2 @x, tD ==
H1,0L H1,0L

k 0 k { {
1 1 2
1

Cos@t W + t2 D P2 @xD>
2

yRule = 8y -> H4 # H1 - #L &L<;

Moreover, we write the associated boundary conditions as

BC = 8uj @0, tD == 0, uj@1, tD == 0<;

displacement; the m j are the viscous damping coefficients; the P j @xD are the distributed amplitudes of the excitation with the
Here, the index j = 1 refers to the vertical component of the displacement and j = 2 refers to the horizontal component of the

frequency W; di, j denotes the Kronecker delta; and b y@xD, where y@xD = 4 x H1 - xL, defined in the yRule, is the initial static

• !!!!!!!!
tic time { · • E , where · and E are the cable density and modulous of elasticity, respectively.
configuration. The spatial and temporal variables were nondimensionalized using the span of the cable { and the characteris -

natural frequencies are in the ratio of two-to-one; that is, b2 a » p2 • 16. When b2 a = p2 • 16, the natural frequencies are
We seek an asymptotic expansion of the response of the cable when its span is such that the first vertical and horizontal

params = 9wj_•;j!=3 :> 2 p, w3 -> p, b2 a -> p 2 • 16=;

and the corresponding normalized modeshapes are

312 Chapter9.nb

i
j$ % 2 y
modeshapes = 9F1 -> j
j %%%% H1 - Cos@2 p #DL &z
z
z
j €€€€ z,
k {
• !!!
3
• !!! • !!!
F2 -> I 2 Sin@2 p #D &M, F3 -> I 2 Sin@p #D &M, F4 -> I 2 Sin@2 p #D &M=;

We use the method of multiple scales and seek a second-order uniform asymptotic expansion of the response of the cable to
a primary-resonance excitation of the first vertical mode in the form

multiScales = 9u_j_@x, tD -> uj @x, T0 , T1, T2 D,

Derivative@m_, n_D@u_D@x, tD -> dt@nD@D@u@x, T0, T1 , T2 D, 8x, m<DD, t -> T0=;

solRule = 9ui_ -> IEvaluateASumAej ui,j @#1, #2, #3, #4D, 8j, 3<EE &M,
vi_ -> IEvaluateASumAej vi,j@#1, #2, #3, #4D, 8j, 3<EE &M=;

Taking into account the resonance conditions, we scale the damping and forcing terms as

scaling = 9mj_ -> e2 mj , Pj_@xD -> e3 Pj@xD=;

where the forcing has been scaled at order e3 so that its influence first appears at the same order as the nonlinear shift in the
frequencies in the absence of the internal resonance.

We consider the case in which the orders of magnitude of the forcing levels for both the vertical and horizontal motions are
the same. Otherwise, a primary resonance of the vertical modes (current study) could also initiate a subharmonic resonance
of order one-half of the first horizontal mode and a primary resonance of the second horizontal mode at different levels of
approximation.

Substituting the multiScales, solRule, and scaling into EOM, expanding the result for small e, and discarding terms of order
higher than e3, we obtain

eq92a =
HHEOM •. Integrate -> int •. multiScales •. solRule •. scaling •• ExpandAllL ••. intRule2 ••
ExpandAllL •. e n_•;n>3 -> 0;

Equating coefficients of like powers of e in eq92a, we obtain

eqEps =
Rest@Thread@CoefficientList@Subtract žž #, eD == 0DD & •ž eq92a •• Transpose •• TrigToExp;

Ÿ 9.2.1 First-Order Solution

The homogeneous parts of the first-order equations are

HlinearSys = #@@1DD & •ž eqEps@@1DDL •. displayRule

8D0 u1,1 - v1,1, -Hu''

1,1L + D0v1,1 - b a int@Hu1,1 L y @xD, 8x, 0, 1<D y @xD,
2 ' ¢ ²

2,1L + D0v2,1 <

D0 u2,1 - v2,1, -Hu''
Chapter9.nb 313

The first-order problems are identical to the linear eigenvalue problems. To account for the interactions arising from the
multiple internal resonances and the primary resonance, we include the lowest two modes in each plane; the other modes will
decay due to damping. Hence, the solution of eqEps[[1]] can be expressed as

sol1u = 9u1,1 -> FunctionA8x, T0 , T1 , T2<,

SumAAi @T1 , T2 D Fi@xD Exp@I wi T0 D + Ai @T1 , T2D Fi @xD Exp@-I wi T0D, 8i, 1, 2<E •• EvaluateE,
•

u2,1 -> FunctionA8x, T0 , T1 , T2<,

SumAAi @T1 , T2 D Fi@xD Exp@I wi T0 D + Ai @T1 , T2D Fi @xD Exp@-I wi T0D, 8i, 3, 4<E •• EvaluateE=;
•

sol1v = Table@
v k,1 ® Function@8x, T0 , T1 , T2<, D@u k,1 @x, T0, T1, T2 D, T0 D •. sol1u •• EvaluateD, 8k, 2<D;

sol1 = Join@sol1u, sol1vD

8u1,1 ® Function@8x, T0 , T1 , T2<, EIT0 w1 A1 @T1, T2 D F1@xD +

EI T0 w2 A2@T1 , T2D F2 @xD + E-I T0 w1 F1@xD A1 @T1, T2 D + E-I T0 w2 F2@xD A2 @T1, T2 DD,
• •

u2,1 ® Function@8x, T0 , T1 , T2<, EIT0 w3 A3 @T1, T2 D F3@xD + EI T0 w4 A4 @T1, T2 D F4@xD +

E-I T0 w3 F3 @xD A3@T1 , T2 D + E-I T0 w4 F4@xD A4 @T1, T2 DD,
• •

v1,1 ® Function@8x, T0 , T1 , T2<, I EI T0 w1 w1 A1 @T1 , T2D F1 @xD + I EI T0 w2 w2 A2 @T1, T2 D F2@xD -

I E-I T0 w1 w1 F1 @xD A1 @T1, T2D - I E-I T0 w2 w2 F2 @xD A2@T1 , T2 DD,
• •

v2,1 ® Function@8x, T0 , T1 , T2<, I EI T0 w3 w3 A3 @T1 , T2D F3 @xD + I EI T0 w4 w4 A4 @T1, T2 D F4@xD -

I E-I T0 w3 w3 F3 @xD A3 @T1, T2D - I E-I T0 w4 w4 F4 @xD A4@T1 , T2 DD<
• •

where F1 @xD and F2 @xD are the lowest symmetric and antisymmetric vertical eigenmodes, and F3@xD and F4@xD are the lowest
symmetric and antisymmetric horizontal eigenmodes. The Fn @xD are orthogonal; they are normalized so that they satisfy the
orthonormality condition Ù0 Fn @xD Fm @xD â x = dn,m .
1

Because the problem is self-adjoint, we have

adjoint = Table@8-I wk F k @xD, F k@xD<, 8k, 4<D;

whose complex conjugate is

adjointC = adjoint •. conjugateRule

88I w1 F1@xD, F1 @xD<, 8I w2 F2 @xD, F2@xD<, 8I w3 F3@xD, F3 @xD<, 8I w4 F4 @xD, F4 @xD<<

Ÿ 9.2.2 Second-Order Solution

Substituting sol1 into eqEps[[2]] yields

314 Chapter9.nb

order2Eq = IlinearSys •. u_ k_,1 -> u k,2M ==

IIlinearSys •. u_ k_,1 -> u k,2M - H#@@1DD & •ž eqEps@@2DDL •. sol1 •. intRule1 ••. intRule2 •.

modeshapes •. yRule •. int -> Integrate •• ExpandM •• Thread;

order2Eq •. displayRule

:D0 u1,2 - v1,2 == - $ %%%

€€€€ % EI T0 w1 HD1A1 L + $ %€€€€
%%%% E Cos@2 p xD HD1 A1L - $ %%%
€€€€ % E-IT0 w1 HD1 A1L +
2 2 I T0 w1 2 •
3 3 3

$ %%€€€€%%% E-I T0 w1 Cos@2 p xD HD1A1 L - • !2! EI T0 w2 HD1 A2 L Sin@2 p xD - • !!!

2 E-I T0 w2 HD1A2 L Sin@2 p xD,
2 • •

1,2L + D0v1,2 - b a int@Hu1,2 L y @xD, 8x, 0, 1<D y @xD ==

3
¢ ²
-Hu'' 2 '

64 b EI T0 Hw1 +w2 L p 2 a Sin@2 p xD A1 A2

•!!
16 64
- €€€€€€ b E2 I T0 w1 p 2 a A21 + €€€€€€ b E2 I T0 w1 p 2 a Cos@2 p xD A21 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ -
3 3 3

16 b E2 I T0 w2 p 2 a A22 - 4 b E2 IT0 w3 p 2 a A23 - 16 b E2 I T0 w4 p 2 a A24 - I $ %%€€€€%%% EI T0 w1 HD1 A1L w1 +

2
3

I $ %%€€€€%%% EI T0 w1 Cos@2 p xD HD1A1 L w1 + I $ %%% €€€€ % E-I T0 w1 HD1A1 L w1 - I $ %%% €€€€ % E-IT0 w1 Cos@2 p xD HD1A1 L w1 -
2 2 • 2 •

• !!! • !!!
3 3 3

I 2 EIT0 w2 HD1 A2L Sin@2 p xD w2 + I 2 E-I T0 w2 HD1A2 L Sin@2 p xD w2 - €€€€€€€ b p 2 a A1 A1 +

• 32 •

I T0 H-w1 +w2L 2
3
•

• ! !
128 • 64 b E p a Sin@2 p xD A A 16 •2
€€€ - €€€€€€€ b E-2 I T0 w1 p2 a A1 +
2 1
€€€€€€€€€ b p 2 a Cos@2 p xD A1 A1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€
3 3 3
I T0 Hw1 -w2L 2 •

• !!!
64 •2 64 b E p a Sin@2 p xD A1 A2
€€€€€€€ b E-2 I T0 w1 p 2 a Cos@2 p xD A1 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ -
3 3
64 b EI T0 H-w1-w2 L p 2 a Sin@2 p xD A1 A2
• •

• !!!
• •2
32 b p 2 a A2 A2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€ - 16 b E-2 I T0 w2 p 2 a A2 -
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3
• •2 • •2
• !!! • !!!
-2 I T0 w3 2
2
8 b p a A3 A3 - 4 b E p a A3 - 32 b p2 a A4 A4 - 16 b E-2 I T0 w4 p 2 a A4 ,
D0 u2,2 - v2,2 == - 2 EI T0 w3 HD1A3 L Sin@p xD - 2 E-I T0 w3 HD1 A3 L Sin@p xD -
•
• !!!
2 EI T0 w4 HD A L Sin@2 p xD - • !2! E-I T0 w4 HD A L Sin@2 p xD,
•

I T0 Hw1 +w3L
64 b EI T0 Hw1+w4 L p 2 a Sin@2 p xD A1 A4
1 4 1 4

2,2L + D0v2,2 == - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ • !!! • !!!

2
16 b E p a Sin@p xD A1 A3
-Hu'' €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€ -

I • !!!2 EIT0 w3 HD1 A3L Sin@p xD w3 + I • !!! 2 E-I T0 w3 HD1A3 L Sin@p xD w3 -

3 3
•
• !!! • !!!
I 2 EIT0 w4 HD1 A4L Sin@2 p xD w4 + I 2 E-I T0 w4 HD1A4 L Sin@2 p xD w4 -
•

16 b EIT0 H-w1+w3 L p 2 a Sin@p xD A3 A1 64 b EI T0 H-w1 +w4L p 2 a Sin@2 p xD A4 A1

• •

• !!! • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€ -
3 3
16 b EIT0 Hw1-w3 L p 2 a Sin@p xD A1 A3 16 b EI T0 H-w1 -w3L p 2 a Sin@p xD A1 A3
• • •

•!! •!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€ -
3 3
IT0 Hw1-w4 L 2 I T0 H-w1 -w4L 2
€€€€€€€€€€ >
• • •

•!! • !!!
64 b E p a Sin@2 p xD A1 A4 64 b E p a Sin@2 p xD A1 A4
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€ €€€€€€€
€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€
3 3

We consider the case of primary resonance of the first vertical mode, a two-to-one internal resonance between the first
vertical and horizontal modes, a one-to-one internal resonance between the first and second vertical modes, and a one-to-one
Chapter9.nb 315

internal resonance between the first vertical and second horizontal modes. In order to collect the terms that may lead to
secular terms from the right-hand sides of order2Eq, we define the rules

omgList = Table@wi, 8i, 4<D;

ResonanceConds = 8W == w1 + e s1, w2 == w1 + e s2 , 2 w3 == w1 + e s3 , w4 == w1 + e s4<;

OmgRule = Solve@ResonanceConds, Complement@omgList, 8#<D ~Join~ 8W<D@@1DD & •ž omgList

99w2 ® e s2 + w1, w3 ® €€€€ He s3 + w1 L, w4 ® e s4 + w1 , W ® e s1 + w1 =,

1
2

9w1 ® -e s2 + w2 , w3 ® €€€€ H-e s2 + e s3 + w2L, w4 ® -e s2 + e s4 + w2, W ® e s1 - e s2 + w2 =,

1

8w1 ® -e s3 + 2 w3 , w2 ® e s2 - e s3 + 2 w3, w4 ® -e s3 + e s4 + 2 w3, W ® e s1 - e s3 + 2 w3 <,

2

9w1 ® -e s4 + w4 , w2 ® e s2 - e s4 + w4, w3 ® €€€€ He s3 - e s4 + w4 L, W ® e s1 - e s4 + w4==

1
2

expRule1@i_D := Exp@arg_D :> Exp@Expand@arg •. OmgRule@@iDDD •. e T0 -> T1 D

Collecting the terms that may lead to secular terms, the terms proportional to EI wi T0 , from the equations governing the
vertical modes, we have

ST1v = CoefficientA#@@2DD & •ž order2Eq@@81, 2<DD •. expRule1@#D, EI w# T0 E & •ž 81, 2<;

ST1v •. displayRule

• !!!
::- $ %%%
€€€€ % HD1A1 L + $ %%€€€€%%% Cos@2 p xD HD1 A1 L - 2 EIT1 s2 HD1 A2L Sin@2 p xD, -4 b EI T1 s3 p 2 a A23 -
2 2
3 3

€€€€ % Cos@2 p xD HD1 A1L w1 - I • !!!

I $ %%€€€€%%% HD1A1 L w1 + I $ %%% 2 EI T1 s2 HD1 A2L Sin@2 p xD w2>,
2 2
3 3

:- $ %%%
€€€€ % E-I T1 s2 HD1A1 L + $ %€€€€
%%%% E Cos@2 p xD HD1 A1L - • !!!
2 HD1 A2L Sin@2 p xD,
2 2 -I T1 s2
3 3

-4 b E-I T1 s2 +I T1 s3 p 2 a A23 - I $ %€€€€

%%%% E HD1 A1 L w1 +
2 -I T1 s2
3

• !!!
I $ %%€€€€%%% E-I T1 s2 Cos@2 p xD HD1A1 L w1 - I 2 HD1 A2L Sin@2 p xD w2>>
2
3

Collecting the terms that may lead to secular terms from the equations governing the horizontal modes, we have

ST1h = CoefficientA#@@2DD & •ž order2Eq@@83, 4<DD •. expRule1@#D, EI w# T0 E & •ž 83, 4<;

ST1h •. displayRule

• !!! • !!!
::- 2 HD1 A3L Sin@p xD, -I 2 HD1 A3 L Sin@p xD w3 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >,
•

• !!!
16 b E-IT1 s3 p 2 a Sin@p xD A1 A3

9- • !!!
2 HD A L Sin@2 p xD, -I • !!!
2 HD A L Sin@2 p xD w =>
3
1 4 1 4 4

Demanding that ST1v[[i]] be orthogonal to the adjointC[[i]], we obtain the solvability conditions
316 Chapter9.nb

SCond1v =
SolveAHint@adjointC@@#DD.ST1v@@#DD, 8x, 0, 1<D •. intRule1 ••. intRule2 •. modeshapes •.
H1,0L
params •. int -> IntegrateL == 0, A # @T1, T2 DE@@1DD & •ž 81, 2< •• Flatten;
SCond1v •. displayRule

:D1 A1 ® I $ %%€€€€%%% b EI T1 s3 p a A23 , D1A2 ® 0>

2
3

Demanding that ST1h[[i]] be orthogonal to the adjointC[[i+2]], we obtain the solvability conditions

SCond1h =
SolveAHint@adjointC@@# + 2DD.ST1h@@#DD, 8x, 0, 1<D •. intRule1 ••. intRule2 •. modeshapes •.
H1,0L
params •. int -> IntegrateL == 0, A #+2 @T1, T2 DE@@1DD & •ž 81, 2< •• Flatten;
SCond1h •. displayRule

:D1 A3 ® 4 I $ %%€€€€%%% b E-I T1 s3 p a A1 A3 , D1 A4 ® 0>

2 •
3

SCond1 = Join@SCond1v, SCond1hD;

whose complex conjugates are

ccSCond1 = SCond1 •. conjugateRule;

Next, we use the solvability conditions to eliminate the D1 Ai from the right-hand sides of order2Eq. To this end, we define
the rules

sigRule = Solve@ResonanceConds, Table@si , 8i, 4<DD@@1DD

9s1 ® - €€€€€€€€€€€€€€€€€ , s2 ® - €€€€€€€€€€€€€€€€ , s3 ® - €€€€€€€€€€€€€€€€€€€€ , s4 ® - €€€€€€€€€€€€€€€€ =

-W + w1 w1 - w2 w1 - 2 w3 w1 - w4
e e e e

expRule2 = Exp@a_D :> Exp@a •. sigRule •. T1 -> e T0 •• ExpandD;

Substituting SCond1, ccSCond1, and expRule2 into order2Eq, we obtain

order2Eqm = order2Eq •. SCond1 •. ccSCond1 •. expRule2;

To determine the solution of order2Eqm and the associated boundary conditions, we use the method of undetermined
coefficients. To accomplish this, we first determine the forms of the terms on the right-hand sides of order2Eqm by picking
up the x dependent or x and T0 dependent terms by defining the rule

baits = 9_ Ea_ T0+b_. f_@c_ xD :> Ea T0 +b f@c xD, a_ f_@b_ xD :> f@b xD •; FreeQ@a, T0 D=;

Then, we collect the solution forms for the horizontal modes according to
Chapter9.nb 317

sol2Formh = Table@Cases@order2Eqm@@k, 2DD, #D & •ž baits, 8k, 3, 4<D •• Flatten •• Union

8E-I T0 w1 -I T0 w3 Sin@p xD, EI T0 w1-I T0 w3 Sin@p xD, E-I T0 w1+I T0 w3 Sin@p xD, EIT0 w1+I T0 w3 Sin@p xD,
E-I T0 w1 -I T0 w4 Sin@2 p xD, EI T0 w1-I T0 w4 Sin@2 p xD, E-I T0 w1+I T0 w4 Sin@2 p xD, EI T0 w1+I T0 w4 Sin@2 p xD<

Each of these terms satisfies the boundary conditions. However, not all of the solution forms

sol2Formv = Table@Cases@order2Eqm@@k, 2DD, #D & •ž baits, 8k, 2<D •• Flatten •• Union

8Cos@2 p xD, E-2 I T0 w1 Cos@2 p xD, E2 I T0 w1 Cos@2 p xD, E-2 I T0 w3 Cos@2 p xD, E2 I T0 w3 Cos@2 p xD,
E-I T0 w1 -I T0 w2 Sin@2 p xD, EI T0 w1-I T0 w2 Sin@2 p xD, E-I T0 w1+I T0 w2 Sin@2 p xD, EI T0 w1+I T0 w2 Sin@2 p xD<

for the vertical modes satisfy the boundary conditions. Besides, there exist terms that are not functions of x on the right-hand
sides of order2Eqm[[{1,2}]], namely

#@@2DD & •ž order2Eqm@@81, 2<DD •. Thread@sol2Formv -> 0D •. displayRule

9- €€€€ I b E2 I T0 w3 p a A23 + €€€€ I b E-2 I T0 w3 p a A3 ,

2 2 •2
3 3
16
- €€€€€€ b E2 I T0 w1 p 2 a A21 - 16 b E2 IT0 w2 p 2 a A22 - 4 b E2 I T0 w3 p 2 a A23 - 16 b E2 I T0 w4 p 2 a A24 +
3
2 32 • 16 •2 • •2
€€€€ b E2 I T0 w3 p a A23 w1 - €€€€€€€ b p 2 a A1 A1 - €€€€€€€ b E-2 I T0 w1 p 2 a A1 - 32 b p 2 a A2 A2 - 16 b E-2I T0 w2 p 2 a A2 -
3 3 3
p a A3 + €€€€ b E-2 IT0 w3 p a w1 A3 - 32 b p 2 a A4 A4 - 16 b E-2 I T0 w4 p 2 a A4=
2 • -2 I T0 w3 2 •2 2 •2 • •2
8 b p a A3 A3 - 4 b E
3

Hence, we need to either consider some algebraic forms or combined algebraic and sinusoidal forms for x to manually adjust
the forms in sol2Formv or use a more general approach to let Mathematica pick up the right forms. The latter approach is
used here.

Ÿ Vertical Modes

Instead of looking for the x dependent forms, we seek all of the T0 dependent forms as

sol2Formv1 =
Join@81<, Table@Cases@order2Eqm@@k, 2DD, Exp@_D, ¥D, 8k, 2<D •• Flatten •• UnionD

81, E-2 I T0 w1 , E2 I T0 w1 , E-2 I T0 w2 , E2 I T0 w2 , E-I T0 w1-I T0 w2 ,

EI T0 w1 -I T0 w2 , E-IT0 w1+I T0 w2 , EI T0 w1 +I T0 w2 , E-2 I T0 w3 , E2 I T0 w3 , E-2I T0 w4 , E2 I T0 w4 <

where the first argument in sol2Formv1 is included to account for the T0 independent terms.

We define the undetermined coefficients for u 1, 2 and v1, 2 as

uSymbolv = Table@y1,i@xD, 8i, Length@sol2Formv1D<D;

vSymbolv = uSymbolv •. y -> h;

and write the general solution in the form

318 Chapter9.nb

sol2v = 8u1,2 -> Function@8x, T0 , T1 , T2<, uSymbolv.sol2Formv1 •• EvaluateD,

v1,2 -> Function@8x, T0 , T1 , T2<, vSymbolv.sol2Formv1 •• EvaluateD<;

Substituting sol2v into order2Eqm[[1]] and equating the coefficients of the sol2Formv1 on both sides yields

expr1 = Subtract žž order2Eqm@@1DD •. sol2v;

eqv1 = Join@8expr1 == 0 •. Exp@_D -> 0<, Coefficient@expr1, Rest@sol2Formv1DD == 0 •• ThreadD

9-h1,1@xD == 0, -h1,2@xD - 2 I w1 y1,2 @xD == 0, -h1,3 @xD + 2 I w1 y1,3 @xD == 0,

-h1,4@xD - 2 I w2 y1,4 @xD == 0, -h1,5 @xD + 2 I w2 y1,5 @xD == 0,
-h1,6@xD - I w1 y1,6 @xD - I w2 y1,6 @xD == 0, -h1,7 @xD + I w1 y1,7@xD - I w2 y1,7 @xD == 0,
-h1,8@xD - I w1 y1,8 @xD + I w2 y1,8 @xD == 0, -h1,9 @xD + I w1 y1,9@xD + I w2 y1,9 @xD == 0,

- €€€€ I b p a A3@T1 , T2 D + €€€€ I b p a Cos@2 p xD A3 @T1, T2 D - h1,10@xD - 2 I w3 y1,10@xD == 0,

2 • 2 2 • 2

3 3

€€€€ I b p a A3 @T1, T2 D2 - €€€€ I b p a Cos@2 p xD A3 @T1, T2D2 - h1,11 @xD + 2 I w3 y1,11 @xD == 0,
2 2

-h1,12@xD - 2 I w4 y1,12@xD == 0, -h1,13@xD + 2 I w4 y1,13 @xD == 0=

3 3

Solving for vSymbolv fromeqv1, we obtain

eta1Sol = Solve@eqv1, vSymbolvD@@1DD

9h1,1 @xD ® 0, h1,2 @xD ® -2 I w1 y1,2@xD, h1,3 @xD ® 2 I w1 y1,3 @xD, h1,4 @xD ® -2 I w2 y1,4@xD,
h1,5 @xD ® 2 I w2 y1,5@xD, h1,6 @xD ® -I Hw1 y1,6@xD + w2 y1,6 @xDL, h1,7 @xD ® I Hw1 y1,7 @xD - w2 y1,7 @xDL,
h1,8 @xD ® -I Hw1 y1,8 @xD - w2 y1,8@xDL, h1,9 @xD ® I Hw1 y1,9 @xD + w2 y1,9 @xDL,

h1,10 @xD ® €€€€ I J-b p a A3 @T1 , T2 D + b p a Cos@2 p xD A3 @T1, T2 D - 3 w3 y1,10 @xDN,

2 • 2 • 2

h1,11 @xD ® - €€€€ I H-b p a A3@T1 , T2D2 + b p a Cos@2 p xD A3@T1 , T2 D2 - 3 w3 y1,11 @xDL,
2

h1,12 @xD ® -2 I w4 y1,12 @xD, h1,13@xD ® 2 I w4 y1,13 @xD=

3

Substituting sol2v and eta1Sol into order2Eqm[[2]] and equating the coefficients of the sol2Formv1 on both sides yields

expr2 = Subtract žž order2Eqm@@2DD •. sol2v •. eta1Sol •. intRule1 ••. intRule2 •.

int@y¢ @xD y1,k_ ¢@xD, _D -> b k •. yRule;
Chapter9.nb 319

eqv2 = Join@8expr2 == 0 •. Exp@_D -> 0<,

Coefficient@expr2, Rest@sol2Formv1DD == 0 •• ThreadD •. params

: €€€€€€€€€€€€ + €€€€€€€ b p 2 a A1 @T1 , T2D A1 @T1, T2 D -

p 2 b1 32 •
2 3

€€€€€€€€€ b p 2 a Cos@2 p xD A1 @T1, T2 D A1@T1 , T2 D + 32 b p 2 a A2 @T1 , T2D A2 @T1, T2 D +

128 • •

8 b p 2 a A3 @T1, T2D A3 @T1, T2 D + 32 b p 2 a A4@T1 , T2D A4 @T1, T2D - y²1,1 @xD == 0,

3
• •

+ €€€€€€€ b p 2 a A1 @T1 , T2D - €€€€€€€ b p 2 a Cos@2 p xD A1 @T1 , T2D - 16 p 2 y1,2@xD - y²1,2 @xD == 0,
p 2 b2 16 • 2 64 • 2
€€€€€€€€€€€€
2 3 3

+ €€€€€€€ b p 2 a A1 @T1 , T2D2 - €€€€€€€ b p 2 a Cos@2 p xD A1 @T1 , T2D2 - 16 p 2 y1,3@xD - y²1,3 @xD == 0,
p 2 b3 16 64
€€€€€€€€€€€€
2 3 3

+ 16 b p 2 a A2 @T1, T2 D - 16 p 2 y1,4 @xD - y²1,4 @xD == 0,

p 2 b4 • 2
€€€€€€€€€€€€
2

+ 16 b p 2 a A2 @T1, T2 D2 - 16 p 2 y1,5 @xD - y²1,5 @xD == 0,

p 2 b5
€€€€€€€€€€€€

64 b p 2 a Sin@2 p xD A1 @T1 , T2D A2 @T1, T2 D

2

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 16 p 2 y1,6 @xD - y²1,6 @xD == 0,

• •

• !!!
p 2 b6
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2

64 b p 2 a Sin@2 p xD A1 @T1 , T2D A2 @T1, T2 D

3

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - y²1,7 @xD == 0,

•

• !!!
p 2 b7
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2

64 b p a Sin@2 p xD A2 @T1 , T2D A1 @T1, T2 D

3

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - y²1,8 @xD == 0,

•

• !!!
2 2
p b8
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2

64 b p 2 a Sin@2 p xD A1 @T1 , T2D A2 @T1, T2 D

3

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - 16 p 2 y1,9 @xD - y²1,9 @xD == 0,

• !!!
p 2 b9
€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
2 3

+ €€€€ b p 2 a A3 @T1, T2D + €€€€ b p 2 a Cos@2 p xD A3 @T1, T2D - 4 p 2 y1,10@xD - y²1,10 @xD == 0,
p 2 b10 4 • 2 8 • 2
€€€€€€€€€€€€€€
2 3 3

€€€€€€€€€€€€€€ + €€€€ b p 2 a A3 @T1, T2D2 + €€€€ b p 2 a Cos@2 p xD A3 @T1, T2D2 - 4 p 2 y1,11@xD - y²1,11 @xD == 0,
p 2 b11 4 8
2 3 3

€€€€€€€€€€€€€€ + 16 b p 2 a A4 @T1, T2 D - 16 p 2 y1,12 @xD - y²1,12 @xD == 0,

p 2 b12 • 2

+ 16 b p 2 a A4 @T1, T2 D2 - 16 p 2 y1,13 @xD - y²1,13 @xD == 0>

p 2 b13
€€€€€€€€€€€€€€
2

Next, we determine the solutions of the boundary-value problems: eqv2 and the associated boundary conditions. First, we
obtain the information about the solution forms uSymbolv according to
320 Chapter9.nb

yForm = DSolve@8#@@1DD, #@@2DD == 0 •. x -> 0, #@@2DD == 0 •. x -> 1<, #@@2DD, xD@@1DD •.

C@i_D -> 0 & •ž Transpose@8eqv2, uSymbolv<D •• Flatten •• ExpandAll;
yForm •. displayRule

:y1,1 @xD ® - €€€€ p 2 x b1 + €€€€ p 2 x2 b1 - €€€€€€€ b a A1 A1 - €€€€€€ b p 2 x a A1 A1 +

1 1 32 • 16 •
4 4 3 3
16 • 32 • • •
€€€€€€€ b p x a A1 A1 + €€€€€€ b a Cos@2 p xD A1 A1 - 16 b p 2 x a A2 A2 + 16 b p2 x2 a A2 A2 -
2 2
3 3
• • • •
4 b p 2 x a A3 A3 + 4 b p 2 x2 a A3 A3 - 16 b p 2 x a A4 A4 + 16 b p 2 x2 a A4 A4 ,

y1,2 @xD ® €€€€€€ - €€€€€€€ Cos@4 p xD b2 + €€€€ b a A1 - €€€€€€€ b a Cos@2 p xD A1 + €€€€€€ b a Cos@4 p xD A1 ,
b2 1 1 •2 16 •2 13 •2
32 32 3 9 9

y1,3 @xD ® €€€€ b a A1 - €€€€€€ b a Cos@2 p xD A1 + €€€€€€€ b a Cos@4 p xD A1 + €€€€€€€ - €€€€€€ Cos@4 p xD b3 ,
1 2 16 2 13 2 b3 1
3 9 9 32 32
y1,4 @xD ® €€€€4€€ - €€€€€€€ Cos@4 p xD b4 + b a A2 - b a Cos@4 p xD A2 ,
b 1 •2 •2
32 32
y1,5 @xD ® b a A2 - b a Cos@4 p xD A22 + €€€€€€
2 b5 1
€ - €€€€€€ Cos@4 p xD b5 ,
32 32

y1,6 @xD ® €€€€6€€ - €€€€€€€ Cos@4 p xD b6 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• •

• !!!
b 1 16 b a Sin@2 p xD A1 A2
€€€€€€€€€€€€€€€€€€€ ,
32 32 3 3

y1,7 @xD ® - €€€€ p 2 x b7 + €€€€ p 2 x2 b7 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

•

• !!!
1 1 16 b a Sin@2 p xD A1 A2
€€€€€€€€€€€€€€€€€€€ ,
4 4 3

y1,8 @xD ® - €€€€ p 2 x b8 + €€€€ p 2 x2 b8 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

•

• !!!
1 1 16 b a Sin@2 p xD A2 A1
€€€€€€€€€€€€€€€€€€€ ,
4 4 3

y1,9 @xD ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

• !!! €€€€€€€€€€€€€€€€€€ + €€€€
16 b a Sin@2 p xD A1 A2 b9 1
€€ - €€€€€€€ Cos@4 p xD b9 ,
3 3 32 32

y1,10 @xD ® €€€€€€€€

b10 1 1 •2 1 •2 2 •2
- €€€€ Cos@2 p xD b10 + €€€€ b a A3 - €€€€ b a Cos@2 p xD A3 + €€€€ b p x a Sin@2 p xD A3,
8 8 3 3 3
y1,11 @xD ® €€€€ b a A23 - €€€€ b a Cos@2 p xD A23 + €€€€ b p x a Sin@2 p xD A23 + €€€€€€€€ - €€€€ Cos@2 p xD b11,
1 1 2 b 11 1
3 3 3 8 8
y1,12 @xD ® €€€€€€€€ - €€€€€€ Cos@4 p xD b12 + b a A4 - b a Cos@4 p xD A4,
b12 1 •2 •2
32 32

y1,13 @xD ® b a A4 - b a Cos@4 p xD A24 + €€€€€€€€ - €€€€€€€ Cos@4 p xD b13>

2 b13 1
32 32

Then, according to yForm, we pick up the spatial function forms for uSymbolv as

bait1 =
8a_. x n_. :> x n •; FreeQ@a, xD, a_ f_@b_ xD :> f@b xD •; FreeQ@a, xD, a_ x f_@b_ xD :> x f@b xD<;

funcList = Table@
Prepend@Cases@yForm@@i, 2DD, #D & •ž bait1 •• Flatten •• Union, 1D, 8i, Length@yFormD<D

881, x, x2 , Cos@2 p xD<, 81, Cos@2 p xD, Cos@4 p xD<, 81, Cos@2 p xD, Cos@4 p xD<,
81, Cos@4 p xD<, 81, Cos@4 p xD<, 81, Cos@4 p xD, Sin@2 p xD<, 81, x, x2 , Sin@2 p xD<,
81, x, x2 , Sin@2 p xD<, 81, Cos@4 p xD, Sin@2 p xD<, 81, Cos@2 p xD, x Sin@2 p xD<,
81, Cos@2 p xD, x Sin@2 p xD<, 81, Cos@4 p xD<, 81, Cos@4 p xD<<
Chapter9.nb 321

We assume the solution for y1, i as a linear combination of funcList[[i]], substitute it into yForm[[i]], equate both sides,
solve for the undetermined coefficients, and obtain the solution for y1, i as

psi1Sol = Table@
form = funcList@@kDD;
cList = Table@ci, 8i, Length@formD<D;
sol = cList.form; algExpr =
sol - yForm@@k, 2DD •. b k_ -> Integrate@Expand@y¢ @xD D@sol, xD •. yRuleD, 8x, 0, 1<D;
algEq = Append@Coefficient@algExpr, Rest@formDD,
algExpr •. Thread@Rest@formD -> 0DD == 0 •• Thread;
y1,k -> Function@x, sol •. Solve@algEq, cListD@@1DD •• Expand •• EvaluateD,
8k, Length@yFormD<D

Solve::svars : Equations may not give solutions for all "solve" variables.

Solve::svars : Equations may not give solutions for all "solve" variables.
322 Chapter9.nb

48 b p 2 x a A1@T1 , T2 D A1@T1 , T2D

:y1,1 ® FunctionBx, - €€€€€€€ b a A1 @T1, T2 D A1@T1 , T2 D + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
•
32 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ -

48 b p 2 x2 a A1 @T1 , T2D A1 @T1, T2 D

3 3 + p2

€€€€€€€€€€€€€ + €€€€€€€ b a Cos@2 p xD A1@T1 , T2D A1 @T1, T2D -

•
32 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

48 b p x a A2@T1 , T2 D A2@T1 , T2D 48 b p 2 x2 a A2@T1 , T2 D A2@T1 , T2D

3 +p 2 3
2 • •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ -

12 b p 2 x a A3@T1 , T2 D A3@T1 , T2D 12 b p 2 x2 a A3@T1 , T2 D A3@T1 , T2D

3 + p2 3 + p2
• •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€ -

48 b p 2 x a A4@T1 , T2 D A4@T1 , T2D 48 b p 2 x2 a A4@T1 , T2 D A4@T1 , T2D

3 + p2 3 + p2

€€€€€€€€€€€€€ F, y1,2 ® FunctionAx,

• •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 + p2 3 + p2
€€€€ b a A1 @T1, T2 D - €€€€€€€ b a Cos@2 p xD A1@T1 , T2D + €€€€ b a Cos@4 p xD A1 @T1, T2 D E, y1,3 ®
4 • 2 16 • 2 4 • 2

9 9 3
FunctionAx, €€€€ b a A1 @T1, T2 D - €€€€€€ b a Cos@2 p xD A1@T1 , T2D2 + €€€€ b a Cos@4 p xD A1 @T1, T2 D2E,
4 2 16 4
9 9 3
y1,4 ® FunctionAx, €€€€ b a A2 @T1 , T2D - €€€€ b a Cos@4 p xD A2 @T1 , T2 D E,
4 • 2 4 • 2

3 3

y1,5 ® FunctionAx, €€€€ b a A2 @T1 , T2D - €€€€ b a Cos@4 p xD A2 @T1 , T2 D2E,

4 2 4

16 b a Sin@2 p xD A1 @T1 , T2D A2 @T1, T2 D

3 3

€€€€€€€€€€€€€€€€€€€€€€€€€ F,
• •

3 • !!!
y1,6 ® FunctionBx, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

16 b a Sin@2 p xD A1 @T1, T2 D A2@T1 , T2 D

3

€€€€€€€€€€€€€€€€€€€€€€€€€ F,
•

• !!!
y1,7 ® FunctionBx, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

16 b a Sin@2 p xD A2 @T1, T2 D A1@T1 , T2 D

3

€€€€€€€€€€€€€€€€€€€€€€€€€ F,
•

• !!!
y1,8 ® FunctionBx, - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

16 b a Sin@2 p xD A1 @T1 , T2D A2 @T1, T2 D

3

€€€€€€€€€€€€€€€€€€€€€€€€€ F,
• !!!
y1,9 ® FunctionBx, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 3

y1,10 ® FunctionAx, -c2 + Cos@2 p xD c2 + €€€€ b p x a Sin@2 p xD A3 @T1, T2D E,

2 • 2

3
y1,11 ® FunctionAx, -c2 + Cos@2 p xD c2 + €€€€ b p x a Sin@2 p xD A3 @T1, T2D2 E,
2
3
y1,12 ® FunctionAx, €€€€ b a A4 @T1, T2D - €€€€ b a Cos@4 p xD A4@T1 , T2 D E,
4 • 2 4 • 2

3 3
y1,13 ® FunctionAx, €€€€ b a A4 @T1, T2D - €€€€ b a Cos@4 p xD A4@T1 , T2 D2E>
4 2 4
3 3

We note that all of the spatial dependent functions satisfy the boundary conditions.

The sol2Formv1 consists of two parts: a part is related to the secular terms, which consists of terms proportional to E±I w1 T 0
and E±I w2 T 0 and the other part consists of terms that are not related to the secular terms. The positions of the former are

pos@1D =
HPosition@# sol2Formv1 •. expRule1@1D •. Exp@_. + _ T0D -> 0, a_ •; a =!= 0, 1D •• Flatten ••
RestL & •ž 9E-I w1 T0 , EI w1 T0 =

8811<, 810<<
Chapter9.nb 323

pos@2D =
HPosition@# sol2Formv1 •. expRule1@2D •. Exp@_. + _ T0D -> 0, a_ •; a =!= 0, 1D •• Flatten ••
RestL & •ž 9E-I w2 T0 , EI w2 T0 =

8811<, 810<<

Since pos[1] and pos[2] are the same, we simply define

8pos1, pos1cc< = pos@1D •• Flatten

811, 10<

Next, we augment y1,pos1 and h1,pos1 by a linear combination of all possible homogeneous solutions. The unknown
coefficients can be determined by the orthogonality conditions. (y1, pos1cc and h1, pos1cc are simply the complex conjugates of
y1, pos1 and h1, pos1, respectively.) First, we express their solution forms using psi1Sol and eta1Sol as

STFormv =
8y1,pos1@xD + Sum@c1,j Fj @xD, 8j, 2<D, h1,pos1@xD + Sum@c1,j I wj Fj @xD, 8j, 2<D< •. eta1Sol •.
psi1Sol •. ci_ -> 0 •. modeshapes •. params •• Expand

• !!!
:$ %%€€€€%%% c1,1 - $ %%%
€€€€ % Cos@2 p xD c1,1 + 2 Sin@2 p xD c1,2 + €€€€ b p x a Sin@2 p xD A3@T1 , T2D2 ,
2 2 2
3 3 3

%%%% p Cos@2 p xD c1,1 + 2 I • !2! p Sin@2 p xD c1,2 +

2 I $ %%€€€€%%% p c1,1 - 2 I $ %€€€€
2 2
3 3

€€€€ I b p a A3 @T1, T2 D2 - €€€€ I b p a Cos@2 p xD A3 @T1, T2D2 + €€€€ I b p 2 x a Sin@2 p xD A3 @T1, T2 D2 >
2 2 4
3 3 3

Demanding that STFormv be orthogonal to the adjointC[[i]], we obtain the unknown coefficients

c1Rule =
Solve@int@adjointC@@#DD.STFormv, 8x, 0, 1<D == 0 •. modeshapes •. params •. intRule1 ••.
intRule2 •. int -> Integrate, c1,# D & •ž 81, 2< •• Flatten

b p a A3 @T1, T2 D2
:c1,1 ® 0, c1,2 ® - €€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€ >
• !!!
3 2

Then, we adjust psi1Sol with

psi1Sol@@pos1DD = y1,pos1 -> Function@x, STFormv@@1DD •. c1Rule •• EvaluateD

y1,11 ® FunctionAx, - €€€€ b p a Sin@2 p xD A3@T1 , T2D2 + €€€€ b p x a Sin@2 p xD A3 @T1, T2 D2E
1 2
3 3

psi1Sol@@pos1ccDD =
y1,pos1cc -> Function@x, STFormv@@1DD •. c1Rule •. conjugateRule •• EvaluateD

y1,10 ® FunctionAx, - €€€€ b p a Sin@2 p xD A3@T1 , T2D + €€€€ b p x a Sin@2 p xD A3 @T1, T2 D E

1 • 2 2 • 2

3 3
324 Chapter9.nb

Ÿ Horizontal Modes

The sol2Formh consists of two parts: a part is related to the secular terms, which consists of terms proportional to E±I w3 T 0
and E±I w4 T 0 and the other part consists of terms that are not related to the secular terms. The positions of the former are

pos@3D =
HPosition@# sol2Formh •. expRule1@3D •. Exp@_. + _ T0D -> 0, a_ •; a =!= 0, 1D •• Flatten ••
RestL & •ž 9E-I w3 T0 , EI w3 T0 =

882<, 83<<

pos@4D =
HPosition@# sol2Formh •. expRule1@4D •. Exp@_. + _ T0D -> 0, a_ •; a =!= 0, 1D •• Flatten ••
RestL & •ž 9E-I w4 T0 , EI w4 T0 =

88<, 8<<

Hence, we assume that the solution for u 2, 2 and v2, 2 consists of two parts corresponding to the parts in sol2Formh. The
part, which is not related to the secular terms, is

sol2Formha = Delete@sol2Formh, pos@3DD

8E-I T0 w1 -I T0 w3 Sin@p xD, EI T0 w1+I T0 w3 Sin@p xD, E-I T0 w1-I T0 w4 Sin@2 p xD,
EI T0 w1 -I T0 w4 Sin@2 p xD, E-I T0 w1+I T0 w4 Sin@2 p xD, EI T0 w1+I T0 w4 Sin@2 p xD<

We define the corresponding undetermined coefficients as

uSymbolha = Table@y2,i, 8i, Length@sol2FormhaD<D;

vSymbolha = uSymbolha •. y -> h;

Hence, the general solution in terms of the sol2Formha can be written as

sol2ha = 8u2,2 -> Function@8x, T0 , T1 , T2<, uSymbolha.sol2Formha •• EvaluateD,

v2,2 -> Function@8x, T0 , T1 , T2<, vSymbolha.sol2Formha •• EvaluateD<;

The part of sol2Formh, which is related to the secular terms, is

sol2Formhb = sol2Formh@@pos@3D@@1DDDD

8EI T0 w1 -I T0 w3 Sin@p xD<

and define the undetermined coefficients as

uSymbolhb = Table@ys2,i, 8i, Length@sol2FormhbD<D;

where the undetermined coefficients for v2, 2 are related to the uSymbolhb due to the orthogonality condition. Therefore, the
general solution in terms of the sol2Formhb can be written as
Chapter9.nb 325

sol2hb = 8u2,2 -> Function@8x, T0 , T1 , T2<, uSymbolhb.sol2Formhb •• EvaluateD,

v2,2 -> Function@8x, T0 , T1 , T2<, -I w3 uSymbolhb.sol2Formhb •• EvaluateD<;

The total solution is a combination of sol2ha, sol2hb, and the complex conjugate of sol2hb.

Substituting sol2ha into order2Eqm[[{3,4}]] and equating the coefficients of sol2Formha on both sides, we have

algEqa =
Flatten@Coefficient@Subtract žž # •. sol2ha, sol2FormhaD & •ž order2Eqm@@83, 4<DDD == 0 ••
Thread;

Solving for the uSymbolha and the vSymbolha yields

symbolaRule = Solve@algEqa, uSymbolha~Join ~vSymbolhaD@@1DD •. params

2 b a A1 @T1 , T2 D A3@T1 , T2D 2 b a A1 @T1 , T2 D A3@T1 , T2D

:y2,1
• •

• !!! • !!!
® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , y2,2 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

16 b a A1 @T1 , T2 D A4@T1 , T2D 16 b a A1@T1 , T2D A4 @T1, T2D

3 3
• • •

•!! • !!!
y2,3 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , y2,4 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

16 b a A4@T1 , T2D A1 @T1, T2D 16 b a A1@T1 , T2D A4 @T1, T2D

3 3 3
•

• !!! • !!!
y2,5 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , y2,6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ ,

• !!! •!!
-2 I 3 b p a A1 @T1 , T2D A3 @T1, T2 D, h2,2 ® 2 I 3 b p a A1@T1 , T2D A3 @T1, T2 D,
3 3 3
• •
h2,1 ®
64 I b p a A1 @T1 , T2D A4 @T1, T2 D 64 I b p a A1 @T1, T2D A4 @T1, T2 D
€€€€€€€€ >
• •

• !!! •!!
h2,3 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ , h2,4 ® 0, h2,5 ® 0, h2,6 ® €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 3 3 3

Substituting sol2hb into either of order2Eqm[[{3,4}]], equating the coefficients of sol2Formhb on both sides, and solving for
the uSymbolhb, we have

symbolbRule =
Solve@Coefficient@Subtract žž order2Eqm@@3DD •. sol2hb, sol2FormhbD == 0 •• Thread,
uSymbolhbD@@1DD •. params

4 b a A1@T1 , T2D A3 @T1, T2D

:ys2,1 €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ >
•

• !!!
® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3

Substituting symbolaRule and symbolbRule into sol2ha and sol2hb yields

sol2haForm = 8u2,2@x, T0 , T1 , T2D, v2,2 @x, T0 , T1, T2 D< •. sol2ha •. symbolaRule;

sol2hbForm = 8u2,2@x, T0 , T1 , T2D, v2,2 @x, T0 , T1, T2 D< •. sol2hb •. symbolbRule;
326 Chapter9.nb

sol2h = 8u2,2 ® Function@8x, T0 , T1 , T2<,

sol2haForm@@1DD + sol2hbForm@@1DD + Hsol2hbForm@@1DD •. conjugateRuleL •• Expand ••
EvaluateD, v2,2 ® Function@8x, T0 , T1, T2<, sol2haForm@@2DD +
sol2hbForm@@2DD + Hsol2hbForm@@2DD •. conjugateRuleL •• Expand •• EvaluateD<

2 b EI T0 w1 +I T0 w3 a Sin@p xD A1@T1 , T2 D A3@T1 , T2D

:u2,2 ® FunctionB8x, T0 , T1 , T2<, €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
•!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ +

16 b EIT0 w1+I T0 w4 a Sin@2 p xD A1 @T1, T2D A4 @T1, T2 D

3

•!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ -

4 b E-I T0 w1+I T0 w3 a Sin@p xD A3 @T1, T2D A1 @T1, T2 D

3 3
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ -

16 b E-I T0 w1+I T0 w4 a Sin@2 p xD A4 @T1, T2D A1 @T1, T2 D

3
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€ -

4 b EIT0 w1-I T0 w3 a Sin@p xD A1 @T1 , T2D A3 @T1, T2 D

3
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ +

2 b E-I T0 w1-I T0 w3 a Sin@p xD A1 @T1, T2D A3 @T1, T2 D

3
• •

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€ -

16 b EIT0 w1-I T0 w4 a Sin@2 p xD A1 @T1, T2D A4 @T1, T2 D

3
•

•! !
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€ +

16 b E-I T0 w1-I T0 w4 a Sin@2 p xD A1 @T1, T2D A4 @T1, T2 D

3

€€€€€€€€€€€€€€€€€€€€€€ F,
• •

3 • !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

v2,2 ® FunctionB8x, T0 , T1 , T2<, 2 I • !3! b EI T0 w1 +I T0 w3 p a Sin@p xD A1@T1 , T2D A3 @T1, T2D +

3

64 I b EI T0 w1+I T0 w4 p a Sin@2 p xD A1 @T1, T2 D A4@T1 , T2 D

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -

4 I b E-I T0 w1+I T0 w3 a Sin@p xD w3 A3@T1 , T2 D A1@T1 , T2D

3 3
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€ +

4 I b EI T0 w1-I T0 w3 a Sin@p xD w3 A1@T1 , T2 D A3@T1 , T2D

3
•

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€ -

• !!!
2 I 3 b E-I T0 w1-I T0 w3 p a Sin@p xD A1 @T1, T2 D A3@T1 , T2 D -
3
• •

64 I b E-I T0 w1-I T0 w4 p a Sin@2 p xD A1 @T1, T2 D A4@T1 , T2 D

€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ F>
• •

3 • !3!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€

Ÿ 9.2.3 Solvability Conditions

Substituting sol1, sol2v, and sol2h into eqEps[[3]] yields

Chapter9.nb 327

order3Eq =
IlinearSys •. u_ k_,1 -> u k,3 M == IIlinearSys •. u_ k_,1 -> u k,3 M - H#@@1DD & •ž eqEps@@3DDL •.

sol1 •. sol2v •. eta1Sol •. psi1Sol •. sol2h •. intRule1 ••. intRule2 •.

modeshapes •. yRule •. int -> Integrate •• ExpandM •• Thread; •• Timing

843.242 Second, Null<

Collecting the terms that may lead to secular terms, the terms proportional to EI wi T0 , from the equations governing the
vertical modes, we have

ST2v = CoefficientA#@@2DD & •ž order3Eq@@81, 2<DD •. expRule1@#D, EI w# T0 E & •ž 81, 2<;

Collecting the terms that may lead to secular terms from the equations governing the horizontal modes, we have

ST2h = CoefficientA#@@2DD & •ž order3Eq@@83, 4<DD •. expRule1@#D, EI w# T0 E & •ž 83, 4<;

Demanding that ST2v[[i]] be orthogonal to the adjointC[[i]], we obtain the solvability conditions
328 Chapter9.nb

SCond2v =
SolveAHint@adjointC@@#DD.ST2v@@#DD, 8x, 0, 1<D •. intRule1 ••. intRule2 •. modeshapes •.
params •. int -> IntegrateL == 0,
H0,1L
A# @T1 , T2 DE@@1DD & •ž 81, 2< •• Flatten •• ExpandAll;
SCond2v •. displayRule

I EIT1 s1+I t1 Ù0 HP1@xD - Cos@2 p xD P1 @xDL â x

:D2 A1 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
1

• !!!
2 •
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - A1 m1 + €€€€ I p 3 a A21 A1 -
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 6 p 3
• •
224 2 2• 256 I b2 p a2 A21 A1 64 I b2 p 3 a2 A21 A1 2 •
2
€€€€€€€€€ I b p a A1 A1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ + €€€€ I E2 I T1 s2 p 3 a A22 A1 -
€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 3 + p2 9 + 3 p2 3
352 • 2 • 352 •
€€€€€€€€€ I b2 E2 IT1 s2 p a2 A22 A1 + €€€€ I E2 I T1 s4 p 3 a A24 A1 - €€€€€€€€€ I b2 E2 I T1 s4 p a2 A24 A1 +
9 3 9
2 2 • •
4 • 256 • 256 I b p a A1 A2 A2 64 I b2 p 3 a2 A1 A2 A2
€€€€ I p 3 a A1 A2 A2 - €€€€€€€€€ I b2 p a2 A1 A2 A2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ +
3 9 3 + p2 9 + 3 p2
• •
1 • 4 • 64 I b2 p a2 A1 A3 A3 16 I b2 p 3 a2 A1 A3 A3
€€€€ I p 3 a A1 A3 A3 - €€€€ I b2 p a2 A1 A3 A3 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ +
3 3 3 + p2 9 + 3 p2
2 2 • •
4 • 256 • 256 I b p a A1 A4 A4 64 I b2 p 3 a2 A1 A4 A4
€€€€ I p 3 a A1 A4 A4 - €€€€€€€€€ I b2 p a2 A1 A4 A4 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ ,
3 + p2 9 + 3 p2
I EIT1 s1-I T1 s2+I t1 Ù01 Sin@2 p xD P1 @xD â x
3 9

• !!!
4 •
D2 A2 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€ - A2 m1 + €€€€ I p 3 a A1 A2 A1 -
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 2 p 3
•
1024 • 64 I b2 p 3 a2 A1 A2 A1 2 •
€€€€€€€€€€€€ I b2 p a2 A1 A2 A1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ + €€€€ I E-2 I T1 s2 p 3 a A21 A2 -
9 3+p 2 3
•
352 2 -2I T1 s2 2 2 • 3 2 • 32 2 2 2• 64 I b2 p 3 a2 A22 A2
€€€€€€€€€ I b E p a A1 A2 + 6 I p a A2 A2 + €€€€€€ I b p a A2 A2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€ +
9 3 3 + p2
• 32 •
2 I E-2 I T1 s2 +2I T1 s4 p 3 a A24 A2 + €€€€€€€ I b2 E-2 I T1 s2+2 I T1 s4 p a2 A24 A2 +
3

€€€€€€€€€€€€€€ >
2 3 2 • •
• 16 I b p a A 2 A3 A3 • 64 I b2 p 3 a2 A2 A4 A4
I p 3 a A2 A3 A3 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ + 4 I p 3 a A2 A4 A4 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 + p2 3 + p2

Demanding that ST2h[[i]] be orthogonal to the adjointC[[i+2]], we obtain the solvability conditions
Chapter9.nb 329

SCond2h =
SolveAHint@adjointC@@# + 2DD.ST2h@@#DD, 8x, 0, 1<D •. intRule1 ••. intRule2 •. modeshapes •.
params •. int -> IntegrateL == 0,
H0,1L
A #+2 @T1 , T2 DE@@1DD & •ž 81, 2< •• Flatten •• ExpandAll;
SCond2h •. displayRule

:D2 A3 ® -A3 m2 + €€€€ I p 3 a A1 A3 A1 - €€€€€€€€€ I b2 p a2 A1 A3 A1 +

2 • 136 •
3 3
• •
32 I b2 p 3 a2 A1 A3 A1 • 32 I b2 p 3 a2 A2 A3 A2 3 •
€€€€€€€€€€€€€€ + 2 I p 3 a A2 A3 A2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ + €€€€ I p 3 a A23 A3 -
3+p 2 3+p 2 4
• 2 3 2 2• • 2 3 2 •
4 8 I b p a A3 A3 32 I b p a A3 A4 A4
€€€€ I b2 p a2 A23 A3 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€ + 2 I p 3 a A3 A4 A4 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ ,
3 + p2 3 + p2
I EIT1 s1-I T1 s4+I t2 Ù01 Sin@2 p xD P2 @xD â x
3

• !!!
4 •
D2 A4 ® - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€€€€€€€€€€€ - A4 m2 + €€€€ I p 3 a A1 A4 A1 -
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
4 2 p 3
•
1024 • 64 I b2 p 3 a2 A1 A4 A1 •
€€€€€€€€€€€€ I b2 p a2 A1 A4 A1 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ + 4 I p 3 a A2 A4 A2 -
9 3 + p2
• •
64 I b2 p 3 a2 A2 A4 A2 3 • 16 I b2 p 3 a2 A3 A4 A3
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€ + I p a A3 A4 A3 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€ +
3 + p2 3 + p2
2 • 352 • •
€€€€ I E-2 I T1 s4 p3 a A21 A4 - €€€€€€€€€ I b2 E-2 I T1 s4 p a2 A21 A4 + 2 I E2 I T1 s2 -2 I T1 s4 p 3 a A22 A4 +
3 9

€€€€€€€€ >
•
32 • • 32 • 64 I b2 p 3 a2 A24 A4
€€€€€€€ I b2 E2 IT1 s2-2 I T1 s4 p a2 A22 A4 + 6 I p 3 a A24 A4 + €€€€€€ I b2 p a2 A24 A4 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
3 3 3 + p2

We can rewrite SCond2v and SCond2h in compact form by collecting the coefficients of the same form as follows:

collectForm = TableA9Ai@T1 , T2 D EI wi T0 , Ai @T1, T2 D E-I wi T0 =, 8i, 4<E •• Flatten;

•

cubicTerms = Nest@Outer@Times, collectForm, #D &, collectForm, 2D •• Flatten •• Union;

cubicST@i_D := E-I wi T0 cubicTerms •. expRule1@iD •. Exp@_ T0 + _.D -> 0 •• Union •• Rest

330 Chapter9.nb

moduEq = MapIndexed@Hterm = cubicST@#2@@1DDD; coef = Coefficient@#1@@2DD, termD;

#1@@1DD == H#1@@2DD •. Thread@term -> 0DL + coef.termL &, Join@SCond2v, SCond2hDD;
moduEq •. displayRule

I EIT1 s1+I t1 Ù0 HP1@xD - Cos@2 p xD P1 @xDL â x

:D2 A1 == - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
1

• !!!
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€ -
4 6 p
i2 64 I b2 p 3 a2 y 2 •
A1 m1 + j j €€€€ I p 3 a - €€€€€€€€€ I b2 p a2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A1 A1 +
256 I b2 p a2
k3 9 + 3 p2 {
224

i2 y • i2 y
3 3 + p2

E2 I T1 s2 j €€€€ I p 3 a - €€€€€€€€€ I b2 p a2 z A22 A1 + E2 I T1 s4 j €€€€ I p 3 a - €€€€€€€€€ I b2 p a2z A24 A1 +

k3 { k3 {
352 352 •

i
j y
j €€€€ I p 3 a - €€€€€€€€€ I b2 p a2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
9 9

z A1 A2 A2 +
2 2 2 3 2
p a p a •
k3 {
4 256 256 I b 64 I b

i 16 I b2 p 3 a2 y
9 3 + p2 9 + 3 p2
j
j €€€€ I p 3 a -
64 I b2 p a2
€€€€€ z
z A1 A3 A3 +
•
k3 {
1 4
€€€€ I b2 p a2 - €€€€€€€€€€€€€€€€€€€€€€€€
€ € - €€€€€€€€€€€€€€€€
€€€€€€€€

i 2 3 2y
3 3+ p 2 9+3p 2

j
j €€€€ I p 3 a - €€€€€€€€€ I b2 p a2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A1 A4 A4,
•
2 2

k3 {
4 256 256 I b p a 64 I b p a
3+p 2 9+3p 2

I EIT1 s1-I T1 s2+I t1 Ù0 Sin@2 p xD P1 @xD â x

9
1

• !!!
D2 A2 == - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ - A2 m1 +
4 2 p
i4 64 I b2 p 3 a2 y
j z A1 A2 A1 + E-2 I T1 s2 i
j €€€€ I p 3 a - €€€€€€€€€€€€ I b2 p a2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z j €€€€ I p 3 a - €€€€€€€€€ I b2 p a2 y z A21 A2 +
k3 {
• •
k3 {
1024 2 352

i 64 I b2 p 3 a2 y 2 •
3 + p2

z A2 A2 + E-2I T1 s2+2 I T1 s4 i j2 I p 3 a + €€€€€€ I b2 p a2y

9 9
j
j6 I p 3 a + €€€€€€€ I b2 p a2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z z A24 A2 +
k {
•
k {
32 32

i 16 I b2 p 3 a2 y • i 64 I b2 p 3 a2 y
3 3 + p2 3
j
jI p 3 a - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A2 A3 A3 + j
j4 I p3 a - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A2 A4 A4 ,
•
k { k {
i y i 32 I b2 p 3 a2 y
3 + p 2 3 + p 2

D2 A3 == -A3 m2 + j j €€€€ I p a - €€€€€€€€€ I b p a + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z zA A • j €€€€€€€€€€€€€ z

1 3 A1 + j2 I p a - €€€€€€€€€€€€€€€€ z A2 A3 A2 +
2 3 2
p a •
k { k {
2 3 136 2 2 32 I b 3

i y i y
3 3 3 + p 2 3 + p 2

j
j €€€€ I p 3 a - €€€€ I b2 p a2 - €€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A23 A3 + j j2 I p 3 a - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A3 A4 A4 ,
2 3 2 2 3 2
p a • p a •
k4 { k {
3 4 8 I b 32 I b
3 + p2 3 + p2
I EIT1 s1-I T1 s4+I t2 Ù0 Sin@2 p xD P2 @xD â x
3
1

• !!!
D2 A4 == - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€€
€€€€€€€€€€€€€€€€€€€€€€€€€€ - A4 m2 +
4 2 p
i
j 64 I b2 p 3 a2 y • i 64 I b2 p 3 a2 y
j €€€€ I p 3 a - €€€€€€€€€€€€ I b2 p a2 + €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A1 A4 A1 + j j4 I p 3 a - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z
z A2 A4 A2 +
•
k3 { k {
4 1024

i 2 p 3 a2 y
3+p 2 3 +p 2

j z A3 A4 A3 + E-2I T1 s4 i
jI p 3 a - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z j €€€€ I p 3 a - €€€€€€€€€ I b2 p a2 y
9

z A21 A4 +
k3 {
• •
k {
16 I b 2 352

i 64 I b2 p 3 a2 y 2 •
3 + p2

E2 I T1 s2 -2 I T1 s4 i
j2 I p 3 a + €€€€€€€ I b2 p a2 y
9

z A22 A4 + j j6 I p 3 a + €€€€€€€ I b2 p a2 - €€€€€€€€€€€€€€€€€€€€€€€€€€€€€ z

z A4 A4 >
k {
•
k {
32 32
3 3 3 + p2

These solvability conditions satisfy the symmetry properties because

formList = TableAAj@T1 , T2 D TableAAi @T1, T2D Ai @T1, T2 D, 8i, 4<E, 8j, 4<E;
•

formList •. displayRule

88A21 A1, A1 A2 A2 , A1 A3 A3 , A1 A4 A4<, 8A1 A2 A1 , A22 A2, A2 A3 A3 , A2 A4 A4 <,

• • • • • • • •

8A1 A3 A1 , A2 A3 A2, A23 A3 , A3 A4 A4<, 8A1 A4 A1 , A2 A4 A2 , A3 A4 A3, A24 A4 <<

• • • • • • • •
Chapter9.nb 331

notationList = Table@si,j, 8i, 4<, 8j, 4<D

88s1,1 , s1,2, s1,3 , s1,4<, 8s2,1 , s2,2, s2,3 , s2,4<, 8s3,1 , s3,2, s3,3 , s3,4 <, 8s4,1 , s4,2 , s4,3, s4,4 <<

coefRule = MapThread@Rule, 8notationList,

Coefficient@H2 w # IL moduEq@@#, 2DD, formList@@#DDD & •ž Range@4D<, 2D •• Flatten;

testList = Table@si,j - sj,i, 8i, 4<, 8j, i, 4<D •• Flatten •• Union •• Rest

8s1,2 - s2,1 , s1,3 - s3,1 , s2,3 - s3,2 , s1,4 - s4,1 , s2,4 - s4,2 , s3,4 - s4,3 <

testList •. coefRule •. params •• Factor

80, 0, 0, 0, 0, 0<
 329

REFERENCES

1. Arafat, H. N., Nayfeh, A. H., and Chin, C-M., 1998, "Nonlinear Nonplanar Dynamics of
Parametrically Excited Cantilever Beams," Nonlinear Dynamics, 15, 31-61.
2. Benedettini, F., Rega, G., and Alaggio, R., 1995a, "Experimental Analysis of the Finite
Dynamics of a Suspended Cable," Proceedings of the 15th ASME Biennial Conference on
Mechanical Vibrations and Noise, DE Vol. 84-1, 543-552.
3. Benedettini, F., Rega, G., and Alaggio, R., 1995b, "Non-Linear Oscillations of a Four-
Degree-of-Freedom Model of a Suspended Cable under Multiple Internal Resonance
Conditions," J. Sound and Vibration, 182, 775-798.
4. Chin, C-M. and Nayfeh, A. H., 1996, "Bifurcation and Chaos in Externally Excited Circular
Cylindrical Shells," J. Applied Mechanics, 63, 565-574.
5. Chin, C-M. and Nayfeh, A. H., 1997, "Three-to-One Internal Resonances in Hinged-Clamped
Beams," Nonlinear Dynanmics, 12, 129-154.
6. Chin, C-M., Nayfeh, A. H., and Lacarbonara, W., 1997, "Two-to-One Internal Resonances in
Parametrically Excited Buckled Beams," Proceedings of the AIAA Structures, Structural
Dynamics and Materials Conference, 1, 213-223
7. Chu, H-N. and Herrmann, G., 1956, "Influence of Large Amplitudes on Free Flexural
Vibrations of Rectangular Elastic Plates," J. Applied Mechanics, 23, 532-540.
8. Crespo da Silva, M. R. M., 1978, "Flexural-Flexural Oscillations of Beck's Column Subjected
to a Planar Harmonic Excitation," J. Sound and Vibration, 60, 133-144.
9. Crespo da Silva, M. R. M. and Glynn, C. C., 1978a, "Nonlinear Fle xural-Flexural- Torsional
Dynamics of Inextensional Beams, I. Equations of Motion," J. Structural Mechanics, 6, 437-
448.
10. Crespo da Silva, M. R. M. and Glynn, C. C., 1978b, "Nonlinear Flexural-Flexural- Torsional
Dynamics of Inextensional Beams, I. Forced Motion," J. Structural Mechanics, 6, 449-461.
11. Lacarbonara, W., Nayfeh, A. H., and Kreider, W., 1998, "Experimental Validation of
Reduction Methods for Nonlinear Vibrations of Distributed-Parameter Systems: Analysis of a
Buckled Beam," Nonlinear Dynamics, 17, 95-117.
12. Malkin, I. G., 1956, Some Problems in the Theory of Nonlinear Oscillations, GITTL,
Moscow.
13. McIvor, I. K., 1966, "The Elastic Cylindrical Shell under Radial Impulse," J. Applied
Mechanics, 33, 831-837.
14. Nayfeh, A. H., 1975, "Nonlinear Propagation of a Wave Packet in a Hard-Walled Circular
Duct," J. Acoustical Society of America, 57, 803-809.
15. Nayfeh, A. H., 1981, Introduction to Perturbation Techniques, Wiley, New York.
16. Nayfeh, A. H., 1998, "Reduced-Order Models of Weakly Nonlinear Spatially Continuous
Systems," Nonlinear Dynamics, 16, 105-125.
17. Nayfeh, A. H. and Asfar, K. R., 1986, "Response of a Bar Constrained by a Non-Linear
Spring to a Harmonic Excitation," J. Sound and Vibrations, 105, 1-15.
18. Nayfeh, A. H. and Bouguerra, H., 1990, "Non-Linear Response of a Fluid Valve," Int. J. Non-
Linear Mechanics, 25, 433-439.
19. Nayfeh, A. H., Kreider, W., and Anderson, T. J., 1995, "Investigation of Natural Frequencies
and Mode Shapes of Buckled Beams," AIAA J., 33, 1121-1126.
 330

20. Nayfeh, A. H. and Lacarbonara, W., 1997, "On the Discretization of Distributed-Parameter
Systems with Quadratic and Cubic Nonlinearities," Nonlinear Dynamics, 13, 203-220.
21. Nayfeh, A. H. and Lacarbonara, W., 1998, "On the Discretization of Spatially Continuous
Systems with Quadratic and Cubic Nonlinearities," JSME Int. J., C41, 510-531.
22. Nayfeh, A. H., Lacarbonara, W., and Chin, C-M., 1999, "Nonlinear Normal Modes of
Buckled Beams: Three-to-One and One-to-One Internal Resonances," Nonlinear Dynamics,
in press.
23. Nayfeh, A. H. and Mook, D. T., 1979, Nonlinear Oscillations, Wiley, New York.
24. Nayfeh, A. H., Nayfeh, J. F., and Mook, D. T., 1992, "On Methods for Continuous Systems
with Quadratic and Cubic Nonlinearities," Nonlinear Dynamics, 3, 145-162.
25. Nayfeh, S. A. and Nayfeh, A. H., 1993, "Nonlinear Interactions Between Two Widely Spaced
Modes - External Excitation," Int. J. Bifurcation and Chaos, 3, 417-427.
26. Nayfeh, A. H. and Nayfeh, S. A., 1995, "Nonlinear Normal Modes of a Continuous System
with Quadratic Nonlinearities," J. Vibration and Acoustics, 117, 199-205.
27. Nayfeh, A. H., Nayfeh, S. A., and Pakdemirli, M., 1995, "On the Discretization of Weakly
Nonlinear Spatially Continuous Systems," in Nonlinear Dynamics and Stochastic Mechanics,
W. Kliemann and N. Sri Namachchivaya, eds., CRC Press, Boca Raton, Florida, 175-200.
28. Pai, P. F. and Nayfeh, A. H., 1990a, "Non-Linear Non-Planar Oscillations of a Cantilever
Beam under Lateral Base Excitations," Int. J. Non-Linear Mechanics, 25, 455-474.
29. Pai, P. F. and Nayfeh, A. H., 1990b, "Three-Dimensional Nonlinear Vibrations of Composite
Beams- I. Equations of Motion," Nonlinear Dynamics, 1, 477-502.
30. Pakdemirli, M., Nayfeh, S. A., and Nayfeh, A. H., 1995, "Analysis of One-to-One
Autoparametric Resonances in Cables - Discretization vs. Direct Treatment," Nonlinear
Dynamics, 8, 65-83.
31. Raouf, R. A. and Nayfeh, A. H., 1990a, "Nonlinear Axisymmetric Response of Spherical
Shells to a Radial Harmonic Excitation," Int. J. Non-Linear Mechanics, 25, 475-492.
32. Raouf, R. A. and Nayfeh, A. H., 1990b, "One-to-One Autoparametric Resonances in
Infinitely Long Cylindrical Shells," Computers & Structures, 35, 163-173.
33. Rega, G., Lacarbonara, W., Nayfeh, A. H., and Chin, C-M., 1999, "Multiple Resonances in
Suspended Cables: Direct versus Reduced-Order Models," Int. J. Non-Linear Mechanics,
accepted for publication.
34. von K↔rm↔n, T., 1910, "Festigkeitsprobleme in Maschinenbau," Encyklopadie der
Mathematischen Wissenscha ften, P. R. Halmos, ed., American Mathematical Society, 3, 211-
385.
35. Yamamoto, T. 1957, "On the Vibrations of a Rotating Shaft," Mem. Fac. Eng. Nagoya Univ.,
9, 19-115.
36. Yamamoto, T. 1960, "Response Curves at the Critical Speeds of Subharmonic and Summed
and Differential Harmonic Oscillations," Bull. JSME, 3, 397-403.