4

Statically indeterminate structures

Figure 4.1 shows two extremely simple structures. In both cases it is possible to take
either of the bar forces to be arbitrary and still satisfy the equations of equilibrium.
The equilibrium equations thus have many solutions and for that reason both
structures are said to be statically indeterminate. (See Chapter 2.)

FIGURE 4.1 Two simple indeterminate structures

104 Statically Indeterminate Structures [Ch. 4

Statically indeterminate structures are a principle topic of the theory of

structures and the subject of this chapter. The fact that the equilibrium equations in
this case have no unique solution is annoying to the extent that a given load should
produce a unique response from a linear system. The problem is, of course, that the
equations of statics do not completely define a structure; to complete the description
it is necessary to introduce member stiffnesses into the formulation. Once that has
been done the expected unique system response follows directly.
The role of member stiffness is obvious in the two cases shown in Figure
4.1. Intuitively, stiffer members in a structure carry more of the load. Put another
way, as the area of bar 2 goes to zero in each case bar 1 is forced to carry the entire
load P; the argument is symmetric to the extent that as bar 1 goes to zero bar 2 must
carry the load. There are thus simple bounds on the manner in which these systems
can respond.
What eliminates the lack of uniqueness found in the equilibrium equations
is the fact that the pieces of a deformed structure must fit together. That fact will be
enforced here in two different ways and thus give rise to two different methods of
analysis, a force method and a displacement method. For example, if the lower
support is removed in the case of the two bars in series, the structure becomes
statically determinate and the displacement at this support can be computed by
summing the length changes of both bars

F2 + P F2
∆1 + ∆ 2 = +
K1 K2
where

K 1 = A1 E / L1
K 2 = A2 E / L2

Since the support displacement must be zero it follows that

Since a force (F2) is used as the unknown here this type of analysis is called the
force method.
The parallel bars of Figure 4.1 are obvious candidates for another type of
analysis. In this case it is clear that if the displacement δ is known, the bar forces
can be computed as

F1 = K1 δ and F2 = K2 δ

which implies an applied load of

(K1 + K2 ) δ
Sec. 4.1] The Force Method 105

But since the applied load must be P, it follows that

P = (K1 + K2 ) δ ⇒ δ = P / (K1 + K2 )

and

F1 = K1 δ = K1 P / (K1 + K2 ) and F2 = K2 δ = K2 P / (K1 + K2 )

Since a displacement is used as the unknown in this analysis it is called a

displacement method.
106 Statically Indeterminate Structures [Ch. 4

This chapter is concerned with formalizing the force and displacement

methods so that they can be applied to arbitrary structures. The methods are related
to the extent that they both begin by introducing procedures that reduce the structure
to some case that can be dealt with easily (such as a statically determinate
structure). This reduction will be shown to violate the definition of the given
structure. Finally a solution is constructed to repair these violations.
These introductory comments attempt to show that the logic of the analysis
of statically indeterminate structures is simple. The fact that these methods
sometimes appear complex in application is due to the inherent geometrical
complexity of structures which surfaces when displacements or forces must be
computed.

4.1 THE FORCE METHOD

Given the idea of the force method just presented, this section develops a
systematic approach with three steps. It begins with the discussion of a second
single degree of freedom structure which is more practical than the two-bar truss
discussed above. It then moves on to a two-degree of freedom system which finally
leads to a general statement of the force method.

4.1.1 A Single Degree of Freedom System

The force method can easily be described as a sequence of steps:

Step 1. Reduce the structure to a statically determinate structure. That is done in

Example 4.1 by replacing the right-hand horizontal support by a roller. This step
allows the structure to displace where it was formerly fixed. The reaction R which is
temporarily set to zero can be used to define a displacement (discontinuity) δ and its
direction.

Step 2. Compute the value of δ due to the load. The method of virtual work is used
to compute the displacement of the structure at its right support. In this case the
reduced structure is the real structure and the structure marked R = 1 is the virtual
structure. Since this support moves to the right while R and δ have been taken as
positive to the left, δ comes out to he negative.

Step 3. Compute the value of δ due to a unit value of R. Again, virtual work is used
to compute the motion at the support. In this case the structure marked R = 1 serves
as both the virtual structure and the real structure.

Step 4. Solve for the reaction R. R is of course the value of the reaction which is
required in order to push the structure back into place.

Step 5. Compute and plot the final stress resultant diagrams.

Sec. 4.1] The Force Method 107

4.1.2 A Two Degree of Freedom System

The trouble with the single degree of freedom system is that there is an interaction
between redundants which it does not demonstrate. Example 4.2 shows this
interaction clearly. It is again convenient to proceed in steps:

Step 1. Reduce the structure to a statically determinate structure. In the case of

Example 4.2, this requires that two cuts be made thus defining two redundants.
(These redundants are taken to be positive when they place their respective bars in
tension.)

Step 2. Analyze the structure. It will be necessary below to have the bar forces for
three cases of load: the structure under the given load and the structure under unit
values of the redundants R1 and R2.

Step 3. Set up the equations of superposition and compute the coefficients. Two
simultaneous equations will be used to determine the bar forces R1 and R2,

(4.1)

This is now the heart of the force method. Physically it is required to select R1 and
R2 so that the ends of the cut bars are not separated in the final solution. ('The
pieces must fit together.') The interaction mentioned above arises because a change
in R1 causes the ends of bar 2 to separate.
108 Statically Indeterminate Structures [Ch. 4

EXAMPLE 4.2 Continued

Let

The equation

then states that when all the effects are superimposed in the final solution (it is
possible to do this since the equations are linear), the bar separation at R1 must be
zero. The effects to be superimposed are the displacement caused by the load, the
Sec. 4.1] The Force Method 109

displacement caused by a unit value of R1 multiplied by its actual value, and the
displacement caused by a unit value of R2 multiplied by its actual value. The
second of Eqs (4.2) similarly requires that when all effects are superimposed, the
separation at R2 must be zero.
The method of virtual work

is used to compute the coefficients in these equations. In so doing, various

combinations of the three solutions computed in this example are used. The term
δ10 which is the discontinuity at R1 due to the applied load, for example, uses the
zero solution and the one solution in the virtual work expression. In this case FV
→ F1 and F → F0 since the real structure is the zero load case and the virtual
structure corresponds to the one load case. It can similarly be argued that-the
coefficient δij combines the ith and the jth load cases in the virtual work expression.
A tabular form is helpful when computing these coefficients.

Step 4. Solve for the redundants. Given the coefficients it is necessary to solve two
simultaneous equations for the redundants R1 and R2.

Subtracting these equations gives

19.42 (R1 – R2) = 0 ⇒ R1 = R2

Adding the equations gives

29.42 (R1 + R2 ) = 243.2 ⇒ R1 + R2 = 8.27K

or

R1 = R2 = 4.13K

Step 5. Compute the final solution and plot the results. This has been done in the
example.

4.1.3 The General Case

Step 1. Introduce releases to make the structure statically determinate. The basic
idea is to reduce a structure to something that is workable. In this case workable
implies a statically determinate structure for which it is possible to compute both
forces and displacements. In order to do so releases are introduced into the
structure. As remarked in chapter 3, a release is a mechanical device which forces a
particular stress resultant called a redundant to be zero. When a release is inserted
110 Statically Indeterminate Structures [Ch. 4

in a structure it creates, by definition, a discontinuity. In the final solution, the

redundants are selected so that the value of each discontinuity is zero.
With regard to this text, releases will be introduced on a trial and error
basis. When introduced properly, releases can create a statically determinate
structure; when introduced improperly the resulting structure will either be
statically indeterminate or unstable. The number of releases required to make a
structure statically determinate is called the degree of statical indeterminacy, k.
For more comments concerning the degree of statical indeterminacy the reader
should refer to Appendix 6.

Step 2. Analysis. It is necessary to solve the reduced structure for k + 1 loading

conditions: k cases of unit loads corresponding to individual redundants and one
case of the reduced structure under the applied toads.

Step 3. Set up and solve the superposition equations. In general, these equations
have the form

or more simply in matrix form

δ R + δ0 = 0

where δij is the discontinuity at release i due to a unit value of redundant j , δi0 is
the discontinuity at release i due to the applied 'loads' or other external effects such
as temperature and settlement, and Ri is unknown value of the i redundant.
To compute the coefficient δij again requires combining the i and j
solutions in the virtual work expression. As remarked above, each of these
equations requires the discontinuity at a specific release to be zero in the final
solution where all effects are present For a structure which is statically
indeterminate to the kth degree, the force method requires the solution of k
simultaneous equations.

Step 4. Combine solutions and plot the results. As before the final solution is the
combination of all effects,

final solution = zero solution + R1 ⋅ one solution + R2 ⋅ two solution +... +

RK ⋅ kth solution
Sec. 4.1] The Force Method 111
112 Statically Indeterminate Structures [Ch. 4
EXAMPLE 4.3 Cont.
Sec. 4.1] The Force Method 113

EXAMPLE 4.3 Continued

4.2 THE DISPLACEMENT METHOD

There is a less common (manual) method of structural analysis called the

displacement method. In this method constraints are added to the structure until it
becomes workable or falls within the realm of some known solutions. The
displacements associated with the constraints are then selected so that the fictitious
forces associated with the constraints are zero in the final solution. This is most
easily explained through examples.

4.2.1 A structure with a single displacement degree of freedom

Example 4 4 describes an application of the displacement method to a simple rigid
frame. It can be argued in the following manner. In general a rigid frame joint has
three degrees of freedom: a horizontal displacement, a vertical displacement, and a
rotation. In this particular case, if: member length changes are neglected as it is
sometimes common to do, one kinematic degree of freedom remains, the joint
114 Statically Indeterminate Structures [Ch. 4

rotation θ. When θ is specified, it is possible to compute member forces using only

commonly available beam solutions (see Appendix 7).
As a parallel to the force method, the rotation θ can be specified to be zero
(i.e. a constraint can be introduced), but to do so requires the application of a
fictitious external moment F10. A unit value of θ again requires the application of an
external fictitious moment, say F11. In the final solution the value of θ is selected so
that this fictitious external moment is
Sec. 4.2] The Displacement Method 115
zero. Solutions are again superimposed to obtain the final moment diagram.

4.2.2 A structure with two degrees of freedom

Example 4.5 describes a two-degree of freedom structure analyzed by the
displacement method. In this case, a three-span beam is to be solved. This is a two
degree of freedom system, since knowing the two beam rotations at the center
supports means that the structure
116 Statically Indeterminate Structures [Ch. 4

has been reduced to well known solutions for single span beams. In order to
emphasize the idea of a mechanical constraint to which moments may be applied,
fictitious shafts are indicated in the figure (of Example 4.5). As in the case of the
force method, a two-degree of freedom structure requires three analyses, a zero case
which provides for the external load and two cases of unit rotations at the
constraints. Having these analyses it is then possible to write the two superposition
equations which require that the fictitious constraint forces go to zero in the final
solution,

Here δ1 , δ2 are rotations at constraints 1 and 2, F11 is the constraint moment at 1

due to a unit value of δ1 , F12 = F21 is the constraint moment at 1 (joint 2) due to a
unit rotation at joint 2 (joint 1), F22 is the constraint moment at joint 2 due to a unit
rotation at joint 2, F10 is the constraint moment at joint 1 due to load, and F20 is the
constraint moment at joint 2 due to load.
When the load is applied but the interior joints are not allowed to rotate, a
constraint of wL2 / 8 must be applied externally to the first joint as indicated in the
figure (of Example 4.5). It follows that

F10 = - w L2/8 and F20 = 0

If a unit rotation is applied to the first constraint, external moments must

be applied as indicated in the figure (of Example 4.5). It follows that

F11 = 7 EI/L and F21 = 2 EI/L

and from symmetry that

F22 = 7 EI/L and F12 = 2 EI/L

In the case of the example under discussion the superposition equations become
Sec. 4.2] The Displacement Method 117

Adding equations it follows that

Subtracting equations it follows that

118 Statically Indeterminate Structures [Ch. 4

and that

The final solution is shown in the figure.

4.2.3 A three-degree of freedom structure and the general case

It is now possible to formulate a general statement of the displacement method:

Step1. Problem definition. Given a structure, it is first necessary to define the

number of constraints c required to reduce the structure to a workable form. This
number is referred to as the degree of kinematic indeterminacy. It should be pointed
out that while the degree of statical indeterminacy can be shown to be invariant for
a given structure, the degree of kinematic indeterminacy depends upon the type of
analysis to be performed and the types of solutions which are available to work
with.

Step 2. Analysis. For a structure which is kinematically indeterminate to the cth

degree it is necessary to perform c + 1 analyses, one for the loaded structure and c
cases in which unit values of the unknown displacements are applied.

Step 3. Set up and solve the c equations of superposition. Let

The superposition equations then appear as

Sec. 4.2] The Displacement Method 119

This c x c system can be written more compactly as

∑F δ
j
ij j + Fi 0 = 0

or

F δ + Fo = 0

Each of these equations states simply that the fictitious forces associated with each
constraint must be zero in the final solution; the final solution is a superposition of
the fictitious force due to load when all the constraint displacements are zero plus a
term for each unit constraint multiplied by the actual value of that displacement.
The coefficients Fij are computed as part of the analysis of Step 2.

Step 4. Construct the final solution. The final solution is obtained by super position
as

Final solution = zero solution + (one solution) ⋅δ1 + (two solution) ⋅δ2
+
… + (c solution) ⋅δc

Example 4.6 describes the common rigid frame problem of a rigid frame bent
subjected to a uniform load. Kinematically this structure has six degrees of freedom,
two displacements and a rotation at each joint. If the member length changes are
neglected, three degrees of freedom remain, two joint rotations and a horizontal
(sideway) displacement
The superposition equations for a three degree of freedom system are

Using the analysis given in the example these equations become

120 Statically Indeterminate Structures [Ch. 4

Anticipating the problem's symmetry it is convenient to set δ1 = - δ2 in order to

reduce the work of solution. An equivalent procedure is to rewrite the system matrix
as

and
δ1 = - δ2 = - w L3 / ( 72EI)

It is now possible to go back and find the moment diagram given the displacements.

4.2.4 A pile problem

The pile problem of Example 4.7 shows the power of the displacement method. In
this case, bearing piles transfer a load through a soil layer to a bearing stratum
below. It is commonly assumed that the piles act like truss bars, that the pile cap is
rigid, and that the soil does not exert any frictional force on the piles. As the number
of piles increases, this type I of structure can become highly statically indeterminate
The interesting point is that this structure has three kinematic degrees of freedom
represented by the three degrees of freedom of the pile cap independent of the
number of piles used.
The example shows the analysis of a four pile group subjected to three
displacements: a vertical displacement, a rotation, and a horizontal displacement
and considers a simplified case of batter piles with small slope. An analysis is
presented for the simplified problem subjected to a lateral load V. The results are
predictable to the extent that each batter pile carries half the load V but causes a The
example shows the analysis of a four pile group subjected to three resultant moment
within the pile group; this moment is balanced by the couple of the two center piles.
Sec. 4.2] The Displacement Method 121
122 Statically Indeterminate Structures [Ch. 4

EXAMPLE 4.6 Continued

4.3 EXERCISES

1. Compute and plot the moment diagram for the frame shown using:
(a) The force method.
(b) The displacement method (neglecting length change)

Note:
(1) Symmetry can be used to simplify calculations.
Sec. 4.3] Exercises 123

(2) This problem offers the reader an opportunity to check the common
assumption that member length has a negligible effect in the analysis of rigid
frames.

2. Extensions. Both the force and the displacement method can be extended easily
to include such effects temperature, lack of fit, settlement, . . . These effects simply
appear in the 'forcing terms' Fi0 and δi0 and do not affect the remainder of the
analysis. Put another way, these effects produce discontinuities and constraint
forces but they do not affect the manner in which compensation for these
discontinuities and constraint forces is made.
For example, suppose that the upper chord members of the truss of
Example 4.2 are heated (by the sun) 60 oF. This produces a thermal length change
in both bars of

where the coefficient of thermal expansion α is 0.65 x 10-5 per oF for steel. The
term F10 = F20 can now be computed using the virtual work expression

(see Example 4.2). The term 27.6 x 10-4 simply replaces 121.61/E in this
example. It follows that

Once the redundants have been computed, the other bars follow directly.

Exercise 2(a) Solve the case in which only one of the above bars is heated.

Similar remarks hold for the displacement method. For example, suppose that the
left support of Example 4.3 settles by some amount d. Since the settlement
problem for a single span beam is readily available,
124 Statically Indeterminate Structures [Ch. 4

Exercise 2(b) Solve the beam problem of Example 4.4 for the case in which one
of the center supports settles an amount d.
Sec. 4.3] Exercises 125
126 Statically Indeterminate Structures [Ch. 4
Sec. 4.3] Exercises 127

SOLUTION PROBLEM 9 Continued

10. Analyze the structure shown using the force method. Check your solution
using the displacement method.
128 Statically Indeterminate Structures [Ch. 4
Sec. 4.3] Exercises 129

15 Solve the structure shown using the force method.

16 Modify the problem of Example 4.2 and solve the case of that structure
subjected to a temperature change of 100 oF along its upper chord. (Do not include
the 10K force.)

17 Analyze the frame of Exercise 4 above using the force method.

18 Analyze the structure shown using the force method.

I=400 in4
A=20 in2
E=29x106 psi
130 Statically Indeterminate Structures [Ch. 4

19 Solve the truss shown using the force method. Assume L/A=1, E=29x106
psi.

20 Solve the truss shown using the force method. Assume A=1 in2 for all bars.
E=29x106 psi.