PDE: separation of variables problem

eckiller
Messages
41
Reaction score
0
I am to reduce the following PDE to 2 ODEs and find only the particular solutions:

u_tt - u_xx - u = 0; u_t(x,0) = 0; u(0,t) = u(1,t) = 0

I guess u = X(x)T(t), and plug u_tt, u_xx into PDE and divide by u to get:

T''/T = X''/X + 1 = K

I solve X'' + (1-K)X = 0 first.

From characteristic equation, r = +- sqrt(1-K)

Becausse of boundary conditions, we must have 1-K < 0

So r = +- i sqrt(K-1)

(I think I make an error somewhere around here...)

==> u = c1 cos(sqrt(K-1)x) + c2 sin(sqrt(K-1)x)

From boundary conditions, c1 =0, and sin(sqrt(K-1)) = 0 => sqrt(K-1) = n*pi

=> K-1 = n^2*pi^2
=> K = n^2*pi^2 + 1

Correct so far?

Now for T:

T'' - KT = 0

T'' - (n^2*pi^2 + 1)T = 0

r = +- sqrt(n^2*pi^2+1)

I think my K = n^2*pi^2 + 1 is wrong because it is strictly positive and I don't think an e^t solution will satisfy the initial condition.

Please help.

The book's answer is cos( sqrt(n^2*pi^2-1)t) * sin(n*pi*x)
 
Physics news on Phys.org
When you separate variables and equate to a separation constant, usually need to check the boundary conditions against a positive, negative, and zero separation constant. For now, these types of equations just have a negative separation constant. That is:

\frac{T^{&#039;&#039;}}{T}=\frac{X^{&#039;&#039;}}{X}+1=K=-\lambda;\quad \lambda&gt;0

So, need to solve:

\frac{X^{&#039;&#039;}}{X}+1=-\lambda

or:

X^{&#039;&#039;}+X(1+\lambda)=0

Solving for the roots:

m=\pm \sqrt{-(1+\lambda)}

Thus have:

X(x)=A_1Cos(\sqrt{1+\lambda}x)+A_2Sin(\sqrt{1+\lambda}x)

Substituting the boundary conditions yield:

\lambda=(n\pi)^2-1;\quad n&gt;0

and:

X(x)=A_2Sin(n\pi x)

Can you do T now?

Edit: Also, I think you need another constraint to obtain a particular solution, you know, wave equations usually specify the initial velocity as well.
 
Last edited:
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...
Back
Top