CH12 BVPs in Rectangular Coordinates
CH12 BVPs in Rectangular Coordinates
Boundary-Value Problems in
Rectangular Coordinates
12-1
Chapter Contents
12.1 Separable Partial Differential Equations
12.2 Classical PDEs and Boundary-Value
Problems
12.3 Heat Equation
12.4 Wave Equation
12.5 Laplace’s Equation
12.6 Nonhomogeneous Boundary-Value
Problems
12.7 Orthogonal Series Expansions
12.8 Higher-Dimensional Problems
12-2
12.1 Separable Partial Differential
Equations
12-3
◼ Linear PDE:
The general form of a linear second-order
PDE is given by
12-4
Example: (Linear Second-Order PDEs)
The PDEs
2u 2 u 2u u
+ 2 = 0 and + = xy
x y
2
x y
2
12-5
◼ Find Solution by Separation of Variables:
To try solution of the form
u(x, y) = X(x)Y(y)
u
= X 'Y
x
u
= XY '
y
2u
= X "Y
x 2
2u
= XY "
y 2
12-6
Example 1: (Separation of Variables)
2u u
Find product solutions of =4 .
x 2
y
Solution:
Let u(x, y) = X(x)Y(y), the PDE becomes
X "Y = 4 XY '
X" Y'
=
4X Y
X" Y'
Let = = −
4X Y
X " + 4 X = 0 & Y ' + Y = 0
12-7
Case 1: ( = 0)
X ” = 0 and Y’ = 0
The solutions are X = c1 + c2x and Y = c3.
u = XY = (c1 + c2 x)c3 = A1 + B1x
where A1 = c1c3 , B1 = c2c3.
12-8
Case 2: ( = −2, > 0)
X” – 42X = 0 and Y’ − 2Y = 0
The solutions are
X = c4 cosh 2x + c5 sinh 2x and Y = c6e .
2y
2 y
u = XY = (c4 cosh 2x + c5 sinh 2x )c6e
2 y 2 y
= A2 e cosh 2x + B2 e sinh 2x
where A2 = c4c6, B2 = c5c6.
12-9
Case 3: ( = 2 , > 0)
X” + 42X = 0 and Y’ + 2Y = 0
The solutions are
X = c7 cos 2x + c8 sin 2x and Y = c9e .
− 2 y
− 2 y − 2 y
u = A3e cos 2x + B3e sin 2x
where A3 = c7c9, B3 = c8c9. ■
12-10
Theorem 12.1.1: (Superposition Principle)
If u1, u2, …, uk are solution of a homogeneous
Linear PDE, then the linear combination
u = c1u1 + c2u2 + … + ckuk
where the ci = 1, 2, …, k are constants, is also a
solution.
12-11
Definition 12.1.1: (Classification of Equations)
The PDE
2u 2u 2u u u
A 2 +B + C 2 + D + E + Fu = 0
x xy y x y
where A, B, C, D, E, and F are real constants,
is said to be hyperbolic if B 2
− 4 AC 0
parabolic if B 2 − 4 AC = 0
elliptic if B 2
− 4 AC 0
12-12
Example 2: (Classifying Linear Second-Order
PDEs)
2u u
Solution: 1) 3 2 = A = 3, B = 0, C = 0
x y
B 2 − 4 AC = 0 : parabolic
2u 2u
2) 2 = 2 A = 1, B = 0, C = −1
x y
B 2 − 4 AC 0 : hyperbolic
2u 2u
3) + 2 = 0 A = 1, B = 0, C = 1
x y
2
B 2 − 4 AC 0 : elliptic ■
12-13
12.2 Classical PDEs and Boundary-Value
Problems
12-14
◼ Three Typical 2nd-order PDEs:
2u u
k 2 = , k 0
x t
2
u 2
u
a 2= 2
2
x t
2u 2u
+ 2 =0
x y
2
12-16
Suppose that:
➢ The flow of heat within the rod takes place
only in the x-direction.
➢ The lateral surface of the rod is insulated, i.e.,
no heat escapes from this surface.
➢ No heat is being generated within the rod.
➢ The rod is homogeneous, i.e., its mass per
unit volume is a constant.
➢ The specific heat and thermal conductivity
K of the material of the rod are constants.
12-17
Two empirical laws of heat conduction:
1) The quantity of heat Q in an element of
mass m is
Q = mu = (Ax )u (2.1)
where u is the temperature of the element.
2) The rate of heat flow Qt through the cross
section is proportional to the area A of the
cross section and the partial derivative with
respect to x of the temperature:
Qt = – K Aux. (2.2)
12-18
From (2.2), heat builds up in the slice at the net rate
− KAu x ( x, t ) − [ − KAu x ( x + x, t )]
= KA[u x ( x + x, t ) − u x ( x, t )]
By differentiating (2.1) w.r.t. t, the net rate is also
given by Qt = Axut
Equating the two expressions
K u x ( x + x, t ) − u x ( x, t )
= ut
x
By letting k = K/gr gives the heat equation
ku xx = ut
12-19
12.3 Heat Equation
12-20
Example: Solve the following heat equation:
2u u
k 2= , 0 x L , t0
x t
u (0, t ) = 0 , u ( L, t ) = 0 , t 0
u ( x, 0) = f ( x ) , 0 x L
12-21
Solution: 1) Using u(x, t) = X(x)T(t) and - as
the separation constant
X T
= = −
X kT
X + X = 0 and T + kT = 0
X ( x ) = c1 + c2 x, =0
X ( x ) = c1 cosh x + c2 sinh x, = −2 0
X ( x ) = c1 cos x + c2 sin x, = 2 0
T (t ) = ce − kt
12-22
2) The boundary conditions yield
u(0, t) = X(0)T(t) = 0 and u(L, t) = X(L)T(t) = 0.
X(0) = X(L) = 0
X + X = 0, X (0) = 0, X ( L) = 0
X ( x ) = 0, if = 0
X ( x ) = 0 if = −2 0
12-23
3) To find a nontrivial solution, it remains the
case of = 2 0 . The general solution is
X ( x ) = c1 cos x + c2 sin x
The boundary conditions yield c1 = 0 and
= n/L, n = 1, 2, 3, ….
n = n2 = (n/L)2, n = 1, 2, 3, … and the
corresponding solutions are
n
X n ( x ) = c2 n sin x, n = 1, 2, 3, ...
L
Tn = c3n e − k ( n 2 2 / L2 ) t
un = X n ( x )Tn (t ) = An e − k ( n / L )t sin n x
2 2 2
L
where An = c2nc3n.
12-24
The superposition principle implies that
n
u ( x, t ) = un = Ane − k ( n 2 2 / L2 ) t
sin x
n =1 n =1 L
satisfies the heat equation and the B.C.
12-26
12.4 Wave Equation
12-27
Example: Solve the following wave equation
2
u 2
u
a 2
= , 0 x L , t0
x 2
t 2
u (0, t ) = 0 , u ( L, t ) = 0 , t0
u
u ( x , 0) = f ( x ) , = g ( x ), 0 x L
t t =0
12-28
Solution: 1) Assume u(x, t) = X(x)T(t)
X T
= 2 = −
X aT
X + X = 0 and T + a 2T = 0
From the B.C., X has nontrivial solution
only for the case of = 2 > 0, > 0. Thus,
in the following, we only consider this case.
12-29
2) X + X = 0 and = 2 > 0
X = c1 cos x + c2 sin x
X(0) = 0 and X(L) = 0
c1= 0 and c2 sin L = 0.
= n/L, n = 1, 2, 3, …
n
n = n / L & X ( x ) = c2 sin x, n = 1, 2, 3, ...
2 2 2
L
na na
T (t ) = c3 cos t + c4 sin t
L L
Let An = c2c3, Bn = c2c4,
na na n
un = An cos t + Bn sin t sin x
L L L
12-30
The superposition principle implies that
na na n
u(x, t) = An cos t + Bn sin t sin x
n=1 L L L
satisfies both the PDE and the B.C.
12-31
u
To determine Bn, we use = g(x), 0 x L
t t=0
u na na na na n
= − An sin t + Bn cos t sin x
t n=1 L L L L L
u
na n
t =0 = g ( x ) = Bn sin x
t n =1 L L
na 2 L n
Bn = g ( x) sin dx
L L 0 L
2 L n
Bn = 0 g(x)sin dx
na L
na na n
u(x, t) = An cos t + Bn sin t sin x
n=1 L L L
where An and Bn are given above.
12-32
na na n
4) Note that, un = An cos t + Bn sin t sin x
L L L
na n
= Cn sin t + n sin x
L L
An Bn
where Cn = An + Bn , sin n = , cos n =
2 2
Cn Cn
is called standing waves or normal modes.
12-33
5) When n = 1, u1(x, t) is called the first
standing wave, the first normal mode or the
fundamental mode of vibration.
The frequency f1 = a/2L of the first normal
mode is called the fundamental frequency
or first harmonic.
12-34
■
12-35
12.5 Laplace’s Equation
12-36
Example: Solve
2u 2u
+ 2 = 0, 0 x a , 0 y b
x y
2
u u
= 0, = 0, 0 y b
x x=0 x x=a
u(x, 0) = 0, u(x, b) = f (x), 0 x a
12-37
Solution: 1) With u(x, y) = X(x)Y(y)
X" Y"
=− = −
X Y
X " + X = 0 & Y "− Y = 0
12-38
3) Now, consider
X + X = 0, X '(0) = 0, X '(a) = 0
Case of = 0,
X = c1, c1 0 is a nontrivial solution.
Case of = −2 < 0, > 0,
No nontrivial solutions.
Case of = 2 > 0, > 0
X = c1 cos x, = n/a, n = 1, 2, 3, ….
and the corresponding eigenvalues are
n = (n/a)2, n = 1, 2, …
12-39
By corresponding 0 with n = 0
The eigenvalues are n = (n/a)2, n = 0, 1, 2, …
and the corresponding eigenfunctions are
n
X = c1 cos x, n = 0, 1, 2, ...
a
12-41
6) To satisfy the final condition u( x, b) = f ( x ), 0 x a
n n
A0b + An sinh b cos x = f ( x )
n =1 a a
which is a half-range expansion of f in a
Fourier cosine series.
If we let A0b = a0/2 and An sin (nb/a)= an
1 a
A0 = f ( x )
ab 0
2 a n
An =
n 0
f ( x )cos xdx
a
a sin b
a
The solution then derived. ■
12-42
Example: (Dirichlet Problem)
Solve the following Dirichlet Problem
2 u 2u
+ 2 = 0, 0 x a, 0 y b
x y
2
u (0, y ) = 0, u (a, y ) = 0
u ( x, 0) = 0, u ( x, b) = f ( x)
12-43
Solution: With the same procedure, it is not
difficult to show that the solution is
n n
u( x, y ) = An sinh y sin x
n =1 a a
2 a n
where An =
n b 0
f ( x )sin xdx
a
■
a sinh
a
12-44
Example: (Superposition Principle)
The solution of the following BVP
2 u 2u
+ 2 =0, 0 xa, 0 yb
x y
2
u (0, y ) = F ( y ) , u (a, y ) = G ( y ) , 0 y b
u ( x, 0) = f ( x) , u ( x, b) = g ( x) , 0 x a
can be divided into two problems, each of
which has homogeneous B.C. on parallel
boundaries. Details are given below:
12-45
1) It is easy to show that the solution of the BVP
is given by u = u1 + u2 , where u1 and u2 are
solutions of the following two problems.
2u1 2u1
Problem 1: 2 + 2 = 0, 0 x a, 0 yb
x y
u1 (0, y ) = 0, u1 (a, y ) = 0, 0 yb
u1 ( x, 0) = f ( x), u1 ( x, b) = g ( x), 0 x a
2u 2 2 u 2
Problem 2: 2 + 2 = 0, 0 x a, 0 yb
x y
u2 (0, y ) = F ( y ), u2 (a, y ) = G ( y ), 0 y b
u2 ( x, 0) = 0, u2 ( x, b) = 0, 0 xa
12-46
12-47
2) u1 and u2 can be shown to be as follows:
n n
n
u1 ( x, y ) = An cosh y + Bn sinh y sin x
n =1 a a a
n n
n
u2 ( x, y ) = Cn cosh x + Dn sinh x sin y
n =1 b b b
2 a n
where An = 0 f ( x )sin xdx
a a
1 2 a n n
n a 0
Bn = g ( x )sin xdx − An cosh b
sinh b a a
a
2 b n
Cn = F ( y )sin ydy
b 0 b
1 2 b n n
Dn =
n b 0 G ( y )sin ydy − Cn cosh a ■
b b
sinh a
b
12-48
12.6 Nonhomogeneous Boundary-Value
Problems
12-49
◼ A typical nonhomogeneous BVP for the heat
equation is
2u u
k 2 + F ( x, t ) = , 0 x L, t 0
x t
u(0, t ) = u0 (t ), u( L, t ) = u1 (t ), t 0
u( x, 0) = f ( x ), 0 x L
12-51
Let
u(x, t) = v(x, t) + (x)
Then the problem can be reduced to solve the
following two problems:
Promble 1: k "+ F ( x ) = 0, (0) = u0 , ( L) = u1
2 v v
k x 2 = t
Promble 2 : v (0, t ) = 0, v ( L, t ) = 0
v ( x, 0) = f ( x ) − ( x )
12-52
2u u
Example 1: Solve k 2 + r = , 0 x L, t 0
x t
subject to u(0, t ) = 0, u(1, t ) = u0 , t 0
u( x, 0) = f ( x ), 0 x 1
12-53
2) The solution of k "+ r = 0 is
r 2
( x) = − x + c1x + c 2
2k
The B.C. u (0, t ) = v(0, t ) + (0) = 0
u (1, t ) = v(1, t ) + (1) = u0
with v(0, t) = 0 and v(1, t) = 0 implies that
(0) = 0 and (1) = u0
c2 = 0 and c1 = r/2k + u0.
r 2 r x
( x ) = − x + + u0
2k 2 k
12-54
3) Finally, the I.C. u(x, 0) = v(x, 0) + (x) = f (x)
v(x, 0) = f (x) – (x).
It remains to solve
2v v
k 2 = , 0 x 1, t 0
x t
v(0, t ) = 0, v(1, t ) = 0, t 0
r 2 r
v( x, 0) = f ( x) + x − + u0 x, 0 x 1
2k 2k
12-55
4) Using the aforementioned procedure, we have
v ( x, t ) = An e − kn 2 2t
sin nx
n =1
An = 2 f ( x ) + x − + u0 x sin nxdx
1 r 2 r
0
2k 2k
A solution of the original problem is
r 2 r
u(x, t) =− x + +u0 x +Ae −kn22t
sinnx
2k
n
2k n=1
5) Note that,
v ( x, t ) → 0 as t → : transient solution
u( x, t ) → ( x ) as t → : steady − state solution ■
12-56
Case 2: (F and the B.C. are Time-Dependent)
Consider
2u u
k 2 + F ( x, t ) = , 0 x L, t 0
x t
u(0, t ) = u0 (t ), u( L, t ) = u1 (t ), t 0
u( x, 0) = f ( x ), 0 x L
12-57
This time we let a solution in the form of
u(x, t) = v(x, t) + (x, t)
2u 2v 2 u v
2 = 2 + 2 and = +
x x x t t t
12-58
If the function is required to satisfy
(0, t ) = u0 (t ), ( L, t ) = u1 (t )
(6.1)
The B.C. for v(x,t) becomes homogeneous.
Clearly, the function
x (6.2)
( x, t ) = u0 (t ) + [u1 (t ) − u0 (t )]
L
satisfies (6.1) and xx = 0.
u ( x , t ) = v ( x, t ) + ( x, t )
x
= v ( x, t ) + u0 (t ) + [u1 (t ) − u0 (t )] (6.3)
L
12-59
Under the choice of u(x,t), the problem becomes
2v v
k x 2 + G ( x, t ) = t , 0 x L, t 0
v (0, t ) = 0, v( L, t ) = 0, t 0 (6.4)
v ( x, 0) = f ( x ) − ( x, 0), 0 x L
where G(x, t) = F(x, t) – t.
12-60
Strategy for solving (6.4):
Make the assumption that the time-dependent
coefficients vn(t) and Gn(t) can be found so that
n
v ( x, t ) = vn (t )sin x
n =1 L
n
and G (x, t ) = Gn (t )sin x
n =1 L
where sin(nx/L), n = 1, … are the eigenfunctions of
X”+ X = 0, X(0) = 0, X(L) = 0
corresponding to the eigenvalues
n = n2 = n22/L2
12-61
Example 2: Solve
2u u
= , 0 x 1, t 0
x 2
t
u (0, t ) = cos t , u(1, t ) = 0, t 0
u ( x, 0) = 0, 0 x 1
12-62
The problem becomes
2v v
+ (1 − x) sin t = , 0 x 1, t 0
x 2
t
v(0, t ) = 0, v(1, t ) = 0, t 0
v( x, 0) = x − 1, 0 x 1
for v(x, t).
12-63
3) Assume that the functions v(x, t) and G(x, t) =
(1 – x) sin t can be written in the following
Fourier sine series expressions:
v( x, t ) = vn (t ) sin nx
n =1
and (1 − x) sin t = Gn (t ) sin nx
n =1
2 1 2
Gn (t ) = (1 − x )sin t sin nxdx = sin t
1 0 n
2
(1 − x) sin t = sin t sin nx
n =1 n
12-64
4) Since v( x, t ) = vn (t ) sin nx
n =1
2v v
x 2
=
n =1
v n ( t )( − n )sin nx and
2 2
= vn (t )sin nx
t n =1
The PDE becomes
2sin t
n =1
vn '(t ) + n vn (t ) sin nx =
2 2
n =1 n
sin nx
2sin t
vn '(t ) + n vn (t ) =
2 2
n
2 n 2 2 sin t − cos t − n 2 2t
vn ( t ) = + Cn e
n n +1
4 4
n 2 2 sin t − cos t − n 2 2t
v ( x, t ) = 2 + Cn e sin nx
n =1 n( n + 1)
4 4
12-65
5) To determine Cn:
Since v( x, 0) = x − 1, 0 x 1
−2
n =1 n( n + 1)
4 4
+ Cn sin nx = x − 1
−2 1 −2
+ Cn = 2 ( x − 1)sinnxdx =
n( n + 1)
4 4 0 n
2 2
Cn = −
n( n + 1) n
4 4
2 n sin t − cos t + e
2 2 − n 2 2t
e
− n 2 2t
v ( x, t ) = − sin nx
n =1 n( n 4 4 + 1) n
12-66
6) Finally, we have
u ( x, t ) = (1 − x )cos t
2 n 2 2 sin t − cos t + e − n t e − n t
2 2 2 2
+ − sin nx
n =1 n ( n + 1)
4 4
n
■
12-67
12.7 Orthogonal Series Expansions
12-68
Example 1: (Using Orthogonal Series Expansions)
2u u
Solve k 2 = , 0 x 1, t 0
x t
u
u (0, t ) = 0, x =1 = − hu (1, t ), h 0, t 0
x
u ( x, 0) = 1, 0 x 1
12-69
2) Consider
X + X = 0, X (0) = 0, X (1) + hX (1) = 0 (7.1)
Nontrivial solutions exist only for
= 2 > 0, > 0. (Example 2 of Sec. 11.4)
X + X = 0 X = c cos x + c sin x.
1 2
X(0) = 0 c1 = 0 X = c2 sin x.
X’(1) + hX(1) = 0
cos + h sin = 0 or tan = − (7.2)
h
which has roots, denoted by n, n = 1, 2, …
The eigenvalues are n = n2 and the
corresponding eigenfunctions
X(x) = c2 sin nx, n = 1, 2, ….
12-70
− k 2n t
3) The solution of T + kT = 0 is T (t ) = c3e
un = XT = An e − k 2n t
sin n x
u( x, t ) = An e − k 2n t
sin n x
n =1
12-71
12-72
1 1
5) 0 sin n x dx = 0 (1 − cos 2n x ) dx
1
2
2
1 1
= 1 − sin 2n
2 2n
Since sin 2n = 2sin n cos n and, by (7.2), n cos n = − h sin n
1 1
sin n xdx = ( h + cos2 n )
2
0 2
1
1 1 1
Moreover, sinn xdx = − cos n x = (1 − cos n )
0 n 0 n
2h(1 − cos n )
An =
n ( h + cos2 n )
1 − cos n
u ( x, t ) = 2h e − k n 2t
sin n x ■
n =1 n ( h + cos n )
2
12-73
12.8 Higher-Dimensional Problems
12-74
◼ Heat and Wave Equations in Two Dimensions:
u u u
2 2
x y t
12-75
Example 1: (Temperatures in a Plate)
Find the temperature u(x, y, t) in the plate if
the initial temperature is f (x, y) and if the B.C.
are held at temperature zero for time t > 0.
Solution: 1) The problem is to solve
2u 2u u
k 2 + 2 = , 0 x b, 0 y c, t 0
x y t
subject to u (0, y, t ) = 0, u (b, y, t ) = 0, 0 y c, t 0
u ( x, 0, t ) = 0, u ( x, c, t ) = 0, 0 x b, t 0
u ( x, y, 0) = f ( x, y ), 0 x b, 0 y c
12-76
2) Let u(x,y,t) = X(x)Y(y)T(t)
X Y T
k ( X YT + XY T ) = XYT or =− +
X Y kT
X" Y" T '
Let =− + = −
X Y kT
Y T
X + X = 0 and Y = kT + (8.1)
Moreover, we let
Y" T'
= −, + = −
Y kT
Y "+ Y = 0, T '+ k ( + )T = 0 (8.2)
12-77
3) Now the homogeneous conditions
u(0, y, t ) = 0, u(b, y, t ) = 0 X (0) = 0, X ( b) = 0
u( x, 0, t ) = 0, u( x, c, t ) = 0 Y (0) = 0, Y ( c) = 0
We have two problems, one in X and one in Y:
X + X = 0, X (0) = 0, X (b) = 0
Y + Y = 0, Y (0) = 0, Y ( c) = 0
The eigenvalues and eigenfunctions are
m 2 2 m
m = 2 and X m ( x ) = sin x, m = 1, 2, 3, ...
b b
n 2 2 n
n = 2 and Yn ( y ) = sin y, n = 1, 2, 3, ...
c c
12-78
4) The solution for T '+ k ( m + n )T = 0 are
− k [( m / b )2 +( n / c ) 2 ]t
T (t ) = C5e
− k [( m / b )2 + ( n / c )2 ]t m n
umn ( x, y, t ) = Amn e sin x sin y
b c
Using superposition principle, we have
m m
u(x, y, t) = Amne −k[(m/b)2+(n/c)2 ]t
sin xsin y
m=1 n=1 b c
12-79
5) At t = 0, we have
m n
u ( x, y, 0) = f ( x, y ) = Amn sin x sin y
m =1 n =1 b c
4 c b m n
Amn = 0 0 f (x, y)sin xsin y dxdy
bc b c
The problem is then solved. ■
12-80