Chapter3-8

3.8 Laplace’s Equation in Rectangular Coordinates

Figure 3.8-1 A thin rectangular plate with insulated top and bottom surfaces

The two-dimensional steady state heat equation for a thin rectangular plate shown in Figure 3.8-1 is the Laplace’s equation

= 0(3.8-1)

The Laplace’s equation is an elliptic equation for a finite domain that has the following boundary conditions

u(0,y) = u(a,y) = 00 < y < b

u(x,0) = 00 < x < a

andu(x,b) = f2(x)0 < x < a

We assume that u(x,t) can be separated into X(x), a function of x alone,and Y(y), a function of y alone.

u(x,t) = X(x) Y(y)

From the following relations

= T = T

= T = T

Eq. (3.8-1) becomes

= 0

Divide the above equation byXY to obtain

=  = 2 = constant

Since the LHS depends on x only, and the RHS depends on y only, they must equal to a constant 2. The constant must be negative for non-trivial solution. The separation constant should be negative for the direction with both homogeneous boundary conditions. For this problem, x-direction has the homogeneous boundary condition at both ends.

The boundary conditions on u(x,t) are changed to the boundary condition on X(x)

u(0,y) = 0 X(0) Y(y) = 0 X(0) = 0

u(a,y) = 0 X(a) Y(y) = 0 X(a) = 0

The ODE with respect to x is

= 2XX = C1cos(x) + C2sin(x)

The constant C1 can be determined from the boundary condition at x = 0

At x = 0, X = 0 = C1(1) + C2(0) C1 = 0

At x = a, X = 0 = C2sin(a)

To avoid the trivial solution, C2 0 and

sin(a) = 0 a = nn =

The solution with respect to x is

X = Xn= C2sin, n = 1, 2, 3, …

The ODE with respect to y is

= 2 Y = An’cosh(ny) + Bn’sinh(ny)

The constant An’ can be determined from the boundary condition at y = 0,

Y(0) = 0 = An’Y = Bn’sinh(ny)

The product solution is then

un(x,t) = Bnsinsinh

The general solution for the steady two dimensional heat equation is

u(x,y) = Bnsinsinh

The constants Bncan be determined from the boundary condition at y = b

u(x,b) = f2(x) = Bnsinsinh

The coefficients Bn sinh are then obtained from the Fourier sine series expansion of f2(x)

Bn =

Non-homogeneous Boundary Conditions

Figure 3.8-2 A thin rectangular plate with all non-homogeneous boundary conditions

The separation of variables method can be applied directly to an elliptic equation with all homogeneous boundary conditions except one. If there are two or more non-homogeneous boundary conditions, the original problem will need to be decomposed into sub problems that will have all homogeneous boundary conditions except one as shown in Figure 3.8-2.

The two-dimensional steady state heat equation for a thin rectangular plate shown in Figure 3.8-2 is the Laplace’s equation

= 0(3.8-1)

The heat equation for this case has the following boundary conditions

u(0,y) = g1(y),u(a,y) = g2(y),0 < y < b

u(x,0) = f1(x),u(x,b) = f2(x),0 < x < a

The original problem with function u is decomposed into four sub-problems with new functions u1, u2, u3, and u4. The boundary conditions for the sub-problems are given in Figure 3.8-2 where only one non-homogeneous boundary condition exits for each sub-problem. For example, the function u1 has u1(x,0) = f1(x) as the only non-homogeneous boundary condition. The original function u is related to the new functions by

u =u1 + u2 + u3 + u4

The function u2 is already evaluated in the previous example where

u2(x,y) = Bnsinsinh

The constants Bncan be determined from the boundary condition at y = b

u(x,b) = f2(x) = Bnsinsinh

The coefficients Bn sinh are then obtained from the Fourier sine series expansion of f2(x)

Bn =

The function u4 is the same as u2 when y is interchanged with x and b is interchanged with a. The boundary condition f2(x) is replaced by the boundary condition g2(y). Therefore

u4(x,y) = Dnsinsinh

The constants Dncan be determined from the boundary condition at x = a

u(a,y) = g2(y) = Bnsinsinh

The coefficients Dn sinh are then obtained from the Fourier sine series expansion of g2(y)

Dn =

u1 is the same as u2 with y replaced by b-y

u1(x,y) = Ansinsinh

The constants Ancan be determined from the boundary condition at y = 0

u1(x,0) = f1(x) = Ansinsinh

The coefficients An sinh are then obtained from the Fourier sine series expansion of f1(x)

An =

u3 is the same as u4 with x replaced by a-x

u3(x,y) = Cnsinsinh

The constants Cncan be determined from the boundary condition at x = 0

u(0,y) = g1(y) = Cnsinsinh

The coefficients Cn sinh are then obtained from the Fourier sine series expansion of g1(y)

Cn =

Thus the original solution can be constructed from the solutions of the sub-problems.

1