CHAPTER – 2

NON-HOMOGENOUS PARTIAL DIFFERENTIAL

EQUATIONS:
F (D1, D2) Z = f (x, y) ---------- (1)

Is the general form of non-homogenous equation.

If the operator is factorized reducible or if the operator is not factorized

irreducible. In this case the solution consists of two parts the complementary
function (C.F) and the particular integral (P-I). The complementary function (C.F)
is a solution of the corresponding homogenous equation. F (D 1, D2) z = 0 is the
case of reducible and irreducible equation. The particular Integral: Unlike the
standard methods available for various standard functions for the ordinary
equation the partial differential equation do not obey any standard or set laws for
such an evaluation. The reason is simple. On the experimental side, these
equations are obtained or occur in the working out of functions which not only
depend on many variables. But also on so many boundary or initial conditions
which the particular integral has to fulfill. Whether the conditions are of the
Dirichlet type, the Neumann type, or any other, the resulting function, very often,
is worked out by an association and a certain knowledge of the nature of the
problem concerned. For elementary evaluation of “P.I” for a given partial
differential equation we assume a certain composite function to be the solution and
then, by building up the corresponding differential equation, evaluate the constants
for the first order partial differential equations.

2.6 To find particular integral we have the following four cases:

Case (1) If f(x,y) = , then z = =

Case (2) if f(x,y) = Sin (ax +by) then P.I. is given by

27
z= Sin (ax+by) = Sin (ax+by)

i.e. replace , D1D2 and by –a2, -ab and b2

Case (3) If f(x,y) = xm yn then P.I. is given by

z= xm yn = [F (D1, D2)]-1 xmyn

Case (4) if f(x,y) = V(x,y) then P.I. is given by

z= V(x,y) = V(x,y)

Equation (1) can be written as

z = P.I =

Particular integral when right hand side of equation (1) (i.e. f(x,y)) is a polynomial

in x,y. In this case we write. P.I = – (2)

Where is inverse operator of F (D1, D2) since the operator possesses the

algebraic properties we may be able to carry out indicated division of (2) by

F (D1, D2) at least to a certain number of stages and then a possible operator on
f(x,y).

Example (1) Solve the P.D.E = 3x + 2y, ( –4 ) z = 3x + 2y.

C.F is the solution of ( –4 ) z = 0 , (D1 – 4 ) (D1 + 4 )z = 0

root are not repeated, C.F , Z = f1 (y + 2x) + f2 (y-2x)

For particular integral , Z = ,=

28
P.I = , =

= =

Complete solution is , Z = C.F + P.I , Z = f1 (y+2x) + f2 (y-2x) +

Example (2) Solve the P.D.E = 3x2y

(3 – 4D1D2 + )z = 3x2y , C.F is the solution of (3 – 4D1D2 + ) z = 0,

(D1 – D2)(3D1 – D2) Z= 0 , Factors are not repeated

C.F is Z = f1 (y + x) + f2 (3y + x) ,

P.I = =

P.I = =

= = The complete solution is

Z = C.F + P.I , Z = f1 (y + x) + f2 (3y +x) +

Example (3) Solve the P.D.Es i) = 2 (ii)

Sol (i) = 2 , For C.F. ( - )z = 0

(D1+ D2)(D1- D2) Z =0

C.F is the solution of (D1- D2) (D1+ D2) Z = 0

29
Factors are not repeated the solution is Z = f1 (y + x) + f2 (y- x)

P.I = = =

P.I = = = x2

Complete solution is Z = f1 (y + x) + f2 (y- x) + x2

ii)

(D1D2) z = 4y , For C.F. (D1D2) z = 0 , (D1) z = 0 sol is Z = f1 (y)

and (D2) z = 0 sol is Z = f2 (x) , Z = f1 (y) + f2 (x)

P.I. z = = = = 2xy2

Complete solution is Z = f1 (y) + f2 (x) + 2xy2

Example (4) Solve the P.D.E

Sol: - (2 +5 ) z = x +3t , Put = D1, = D2

(2 +5 )z = ( D1 + i D2) ( D1 - i D2) z = 0

Factors are not repeated For C.F , ( D1 + i D2) ( D1 - i D2) z = 0

Solution is z = f1 ( t+ x) + f2 ( t- x) ,

For P.I. Z= =

= = =

Complete solution is Z = f1 ( t+ i x) + f2 ( t- i x) +

30
Example (5) Solve

Sol. ( +3D1D2 + 2 ) z = x +y , For C.F ( + 3D1D2 + 2 ) z =0

(D1 + D2) (D1 + 2D2) z = 0 , Solution is Z = f1 (y - x) + f2 (y – 2x)

P.I = =

= =

= =

Complete solution is Z = f1 (y – x) + f2 (y – 2x) +

Example (6) Solve , Sol: z = x3y3

for C.F z =0 C.F Solution is

Z = f1 (y + x) + f2 (y + x) + f3 (y +2x)

Where  is one of the cube root of unity , For P.I Z=

= =

= ,=

= , hence the complete solution is

31
Z = f1 (y + x) + f2 (y + x) + f3 (y + 2x) +

Example (7) Solve

Sol: ( + 4D1D2-5 )z= y2 , For C.F ( + 4D1D2-5 )z=0

(D1-D2) (D1+5D2) z = 0 , C.F Solution is z = f1 (y + x) + f2 (y - 5x)

P.I z= =

= =

Complete solution is = f1 (y + x) + f2 (y - )+

Example (8) Solve

Sol: - ( - )z = x2 +y2

For C.F ( - ) z = 0 , (D1 – D2) (D1 + D2) z = 0

Solution is z = f1 (y + x) + f2 (y- x) For P.I. Z=

= =

= =

32
P.I = Complete solution is Z = f1 (y + x) + f2 (y- x) +

Example (9) Solve

Sol. (2 +D1D2 – ) z = 1 , For C.F (2 + D1D2 - )z =0

(2D1 - D2) (D1 + D2) z = 0 Solution is Z= f1 (2y + x) + f2 (y – x)

P.I = =

= = =

Complete solution is Z = f1 (2y + x) + f2 (y – x) +

Example (10) Solve

For C.F ( + 2D1D2 + ) z =0

Repeated Roots (D1 + )2 z = 0 , C.F Z = f1 (y – x) + xf2 (y – x)

For P.I Z = a = 3, b = 2

= =

Complete solution Z = f1 (y - x) + xf2 (y – x) +

Example (11) Solve

Sol. ( - 2D1D2 + ) z = Sin (2x +3y)

C.F (D1 - D2)2 z = 0 Roots are repeated so we have

Solution is Z = f1 (y + x) + xf2 (y + x) , Z=

33
In Z = = (Imag part)

= (Imag part) = (- imag part) of e(2x+3y)i

P.I = - Sin (2x+3y)

Complete solution is Z = f1 (y + x) + xf2 (y + x) - Sin (2x+3y)

Example (12) Solve = Cosmx Cosny + 30(2x+y)

For C.F , ( + )z = 0 , (D1 + iD2) (D1 - iD2) z = 0

Solution is Z = f1 (y + ix) + f2 (y – ix) P.I= Cosmx Cosny + 30(2x + y)

= Cosmx Cosny + 30(2x + y)

= (Real Part of)

= (Real Part)

= P.I solution is

Z=

Complete solution

34
i) Z = f1 (y+ix) + f2 (y-ix)

35
Exercise 2.3
Solve

Q.1 = y2+x

Q.2. = x+y

Q.3. = x2

Q.4. = e2x+y

Q.5. = ex+2y

Q.6. = 2e2x + 3x2y

Q.7. ( –4 D2+4D1 ) z = 4 sin (2x+y)

Q.8. = sinx cos2y

Q.9. = 12 xy

Q.10. = E sinpt

Q.11. (2 – 5D1D2+5 ) z = 5 sin (2x + y)

Q.12. ( + 2D1D2+ ) z = x2 + xy + y2
Q.13. (D1 – 2D2) (D1 – D2)3 z = ex+y

Q.14. = 2 sin (3x + y)

Q.15. ( -7D1 -6 ) z = Sin (x+2y) + e2x+y

Q.16. = Cosx Cos2y

Q.17. = Cos (2x+y)

36
Q.18. ( -2 D2 – D1 +2 ) z = ex+y

Q.19. ( -6D1D2 + 9 ) z = 6x + 2y

Q.20. ( + 2D1D2 + ) z = 2 cosy – x siny

Q.21. = e2x + x3y

Q.22. = e3x + x3y2

Q.23. = x sin (2x-3y)

Q.24. = x2 + xy2 + y3

Q.25. = e2x+y

Q.26. = ex cos2y

Q.27. = ex+3y

Q.28. = 2e2x +3x2y

Answers 2.3

Q.1 z = f1(y+x) + f2 (y-5x) +

Q.2 z = f1(y+2x) + f2(y-3x) +

Q.3 Z = f1(y+ax) + f2 (y-ax) +

Q.4 z = f1(y+2x) + xf2 (y+2x) + ½ x2 e2x+y

2.7 Conical form second order Linear Partial Differential Equations with
variable co-efficients:

Substitution Method: Consider a general equation

37
A -------- (1)

A uxx + 2B uxy + Cuyy = f (x, y, u, u x, uy) Where A, B, and C are either constants
or functions of x, y we choose independent variables to new variable say by r = r
(x, y) and s = s (x, y) whose functional form will be determined, so that the given
equation is reduced to a canonical form which may be readily solved.

Consider the equation. A uxx + 2B uxy + Cuyy = f (x, y, u, ux, uy) ------- (2)

a) If AC – B2 > 0, the equation (2) is said to be Elliptic.

b) If AC – B2 = 0, the equation (2) is said to be Parabolic.

c) If AC – B2 < 0, the equation (2) is said to be Hyperbolic.

A parabolic equation of the form -------- (3)

is said to be a three dimensional diffusion equation, c is the diffusivity and

u (x, y, z, t) denote the temperature at any point (x, y, z) of a solid at time t.

one dimensional diffusion equations are and

respectively. In steady heat flow = 0 and equation (3)

reduces to =0 which is known as Laplace Equation.

This is an elliptic equation. A hyperbolic equation of the form

c being a constant is said to be a three dimensional

wave equation. To solve (1) in each case we form the auxiliary equation. A

------- (4) Solve by quadratic formula

38
AQ2 – 2BQ + c = 0, where Q = To find Q = and then solving the two first
order differential equations.
i) If the given equation is Elliptic i.e. AC – B2 > 0

Then equation (4) has distinct roots from which we obtain

 (x, y) = c1 and  (x, y) = c2 the substitution

r =  (x, y) and s =  (x, y) will reduce the given P.D.E. to conical form,

ii) If the given equation is Parabolic i.e. AC – B2 = 0

Then equation (4) has equal roots from which we obtain

 (x, y) = c, in this case the substitutions

r = x and s =  (x, y) will reduce the given P.D.E. to the conical form.

If can be integrated twice w.r. to r.

iii) If the given equation is Parabolic, the equation (4) has the complex roots
(i.e. AC – B2 < 0) by substitution.

r + is =  (x, y) and r – is =  (x, y) will be found to be more convient.

39
Example (1):Using the indicated transformation solve the partial differential
equation.

x (V = x, z = xy)

Sol: x = 0 ----- (1)

A (u, x, y, ux, uy) ------ (2)

Comparing (1) with (2) we get A = 0, 2B = x, c = –y

A=0 B= , c = –y , AC - B2 = 0. (–y) – <0

So the given equation (1) is Hyperbolic and the auxiliary equation

A + c = 0 , –2. –y = 0

(Integrating) , In xy = In c => xy = c

Thus we use substitution r = x, s = xy , using chain rule

------- (a)

= .1 , =

,=

------ (b)

40
Also = = 

------ (c) Substituting (a) (b) and (c) in equation (1) we have

=0 , = 0 (Integrating w.r. to s) = f1 (r) , u = f1 (r) + f2 (s)

u = f1 (x) + f2 (xy)

Example-2 Solve the partial differential equation by using the indicated

transformation. = 0 ----- (1) (V = x, z = x + y)

A = f (u, x, y, ux, yy) ----- (2)

Comparing (1) and (2) we get. A = 1, B = –1, C = 1.

So that AC-B2 = (1) (1) – (–1)2 = 1 – 1 = 0 AC – B 2 = 0 so the given equation is

parabolic the auxiliary equation is. A + c = 0 , has equal roots.

+ 1 = 0, Q = –1, = –1, dy = –dx (Integrating)

x+y=c r = x, s = x + y By chain rule , =

= =

----- (a)

41
 , ,

Also =

----- (c)

Putting (a) (b) and (c) in (1) we get =0 Integrating w.r. to r, = f1 (s)

integrating . w.r. to r again, u = r f1 (s) + f2 (s) u = x f1 (x + y) + f2 (x + y)

is the solution of the P.D.Eq. (1)

Example (3) Solve the P.D.E. by using the indicated transformation.

=0 (1) (r = x + y, s = 2x – y)

Comparing (1) with A uxx + 2B uxy + c uyy = f (x, y, u, ux, uy)

A = 1, 2B = 1, B = 1/2 C = –2 ,

AC – B2 = 1 (–2) – (1/2)2 = –2 – = < 0 so AC – B2 < 0 and the given

equation (1) is Hyperbolic.

The auxiliary equation is A + c = 0 We get Q2 – Q – 2 = 0

The roots are Q = 2, Q = –1 , = 2, = –1

dy = 2dx and = –1 , y – 2x = c1 dy = –dx, 2x – y = c y + x = c1

We will use the substitution r = 2x – y and s = x + y consider u is a function (x, y)

, ux =

42
ux = ur + 2 us ,

= =

uxx= -------(a)

= =

= uxy = –2 ---- (b)

, ,

uyy = ----- (c)

Putting (a), (b) and (c) in (1) we get , uxx + uxy - 2 uyy = 9 =0

=0 Integrating w.r. to r , = f1 (s) , u = f1 (s) + f2 (r)

u = f1 (s) + f2 (r) , u = f1 (x + y) + f2 (2x – y)

Example (4) A string is stretched between the fixed points (0, 0) and (L, 0) and

released at rest from positive u = A Sin . Find u (x, t) = ?

Sol: The boundary conditions are

u (o, t) = u (L, t) = 0 and the initial conditions are

43
u (x, o) = A Sin , =0 when t = 0

u (x, t) = Bn Cos

Where Bn = dx , Bn = dx

n = 1, B1 = dx = dx

B1 = A Hence the particular solution of the wave equation under given condition

is u (x, t) = A Cos

Example (5) If a string of length L is initially at rest in equilibrium position and

each of its points is given the velocity.

= b Sin3 . Find the displacement u (x, t)

Sol: The solution of the wave equation governing vibration of a string of length
L under given boundary condition is

u (x, t) =

Since the string was at rest initially, we have

u (x, o) = 0 Hence u (x, o) = = 0 Bn = 0

Hence u (x, t) =

Diff. w. r. to t,

Now at t = 0. Applying the initial condition.

44
b Sin3

or

Equation co-efficients of Sin

----- (2)

Hence from (1) & (2)

u (x, t) =

Example (6) Show that the Tricomi equation.

y =0 is of mixed type (elliptic in the upper half-plane and

hyperbolic in the lower half-plane).

Sol: y = 0 ---- (1) Comparing (1) with the equation.

A = f (x, y, u, ux, uy)

A = y, B = 0, C = 1 , AC – B2 = y (1) – 0 = y

When y > 0 the equation (1) is elliptic in the upper half-plane and if y < 0 then
the equation (1) is hyperbolic in the lower half-plane.

Example (7) State the nature of the following Eulers equations i.e. whether
parabolic, hyperbolic or elliptic.

45
 (a) = 0 (b) = 0 (c)

=0

Sol: (a) = 0 ----- (1)

A uxx + 2B uxy + C uyy = f (x, y, u, ux, uy) ----- (2) , Comparing (1) with (2)

A = 1, 2B = 4, B= 2 , C= -1, AC – B2 = 1 – (–1)2 = –1 – 4 = –5 < 0

Hyperbolic

(b) uxx - uxy + uyy = 0 (A) comparing equation (A) with (2) we get

A= 1, 2B= -2, B= -1, C= 1 , AC – B2 = 1 – (–1)2 = 1 – 1 = 0 . Parabolic

(c) 4 uxx + uyy = 0 (A) comparing (A) with equation (2) we get

A = 4 B = 0 C = 1 , AC – B2 = 4(1) – (0)2 = 4 > 0 Equation (A) is

Elliptic.

Example (8) State the nature of the following Equation

2 uxx – uxy – 3 uyy = 0 (1) comparing with equation (1) we get

A = 2, 2B= -2, B = -1, C = -3 , AC – B2 = 2 (–3) –

=–6– < 0 , so the equation (1) is Hyperbolic

Example (9) Show that (a) Laplaces equation uxx + uyy = 0 is elliptic

(b) The heat equation ut = c2 uxx is parabolic (c) The wave equation
utt = c2 uxx is hyperbolic

Sol: (a) Laplaces equation =0

Comparing with equation , A = f (x, y, u, ux, uy)----(1)

46
We get A = 1, B = 0 , C = 1 , AC – B2 = (1) (1) – (0)2 = + 1 > 0

So the Laplaces equation is elliptic. (b) The heat equation , c2

= 0, Comparing with (1) , we get A = c2 , B = 0, C = 0, AC – B2 = 0

So the heat equation is parabolic. (c) The wave equation , c2 = 0

comparing with (1) we get A = c2, C = –1, B = 0 , AC – B 2 = c2 (–1) – (0)2 = – c2

< 0, So the wave equation is Hyperbolic

Example (10) Vibration of beam:

It can be shown that the small free vertical vibrations of a uniform beam Fig. are
governed by the fourth-order equation.

= 0 ----- (1),When c2= (E=young’s modulus of electricity,

I = moment of inertia of the cross-section with respect to the y – axis in the figure.
= density A = cross - sectional area).

Substitution , u = F (x). G (t) , uxxxx = F(4)(x) G(t) ----- (a)

utt = F(x) (t) ------- (b) , Putting (a) and (b) in equation (1)

f(x) (t) + c2 F(4)(x) = 0 , By variable separable

= 4 = Constant

We get = 4 (c) and = –4 ----- (d)

Solving differential equation (c) and (d) We get

F(x) = A Cos x + B Sin x + c Cos hx + D Sin hx

and G(t) = a Cos c 2t + b Sin c2t. u = F(x) G(t)

47
u = (A Cos x + B Sin x + C Cos hx + D Sin hx) (a Cos c2t
+ b Sin c2t)
u = A a Cos x Cos c2t + aB Sin x Cos c2t + aC Cos hx Cos c2t

+D a Sin hx Cos c2t A b Cos x Sin c2t + b Sin x Sin c2t

+ bC Cos hx Sin c2t + D b Sin hx Sin c2t

u = A Cos x (a Cos c2t + b Sin c2t) + B Sin x (a Cos c2t + b Sin c2t)

+ C Cos hx (a Cos c2t + b Sin c2t)

48
Exercise 2. 4
Using the indicated transformations, solve the following equations.
Q.1: uxy – uyy = 0 (v = x, z = x + y)
Q.2: uxx + 2 uxy + uyy = 0, (v = x, z = x – y)
Q.3: uxx – 4 uxy + 3 uyy = 0, (v = x + y, z = 3x + y)
Using (14) sketch a figure of the deflection u(x, t) of vibrating string (Length L =
1, ends fixed c = 1) starting with initial velocity zero and the following initial
deflection f(x), where K is small say K = 0.01
Q.4: f(x) = kx (1 – x) Q.5: f(x) = k Sin 2x,
Q.6: f(x) = k Sin2 x Q.7: f(x) = k (x2 – x4)
Q.8: f(x) = k (x – x3)
Q.9: Show that c is the speed of the two waves given by (4)
Q.10: If a steel wire 2 meters in length weighs 0.8 nt (about 0.18 lb) and is
stretched by a tensible force of 200 nt (about 45 lb). What is the
corresponding speed c of transverse waves.
Q.11: If the equation A uxx + 2B uxy + uyy = f (x, y, u, ux, uy) is hyperbolic it can be
transformed to the normal form uvz = F* (v, z, u, uv, uz)
by setting V =  (x, y), z =  (x, y), where  = constant and  = constant
are the solutions y = y (x) of the equation A y2 – 2 By/ + C = 0
Show that in the case of wave equation ,  = x + ct,  = x – ct
Q.12: If the equation (1) in prob. (11) is parabolic, the substitution
v = x, z =  (x, y) with  = x – ct reduces it to the normal form
uvv = F* (v, z, u, uv, vz) verify this result for the equation, uxx + 2uxy + uyy= 0

Q.13: Find the solutions un =

Fn(x) Gn(t) of example (10) equation (1) corresponding to zero initial

49
 velocity and satisfying the boundary conditions , u(o, t) = 0, u (L, t) =
0 , (ends simply supported for all times t)

uxx (o, t) = 0, uxx (L, t)= 0,(zero moments hence zero curvature, at the ends)

Q.14: Find the solution of equation (1) in example (10) that satisfies the condition
in Prob. (13) and the initial condition. u (x, o) = f(x) = x (L – x)

Q.15: What are the boundary conditions of the beam is clamped at both ends?

Q.16: Show that F(x) in Example (10) satisfies the condition in Prob. (25) if L is
a root of the equation Cos hL Cos L = 1 ----- (1)

Also determine the solution of equation (1)

Q.17: If the beam is clamped at the left and free at the other end (See Fig.) the
boundary conditions are u(o, t) = 0,ux (o, t) = 0, uxx (L, t) = 0, uxxx (L, t) = 0

Show that F(x) in Example (10) satisfies these conditions if L is a root of

the equation. Cos hL Cos L = –1 ----- (2)

Also find the approximation solutions of equation (2)

50