How Can Variable Transformation Solve a Non-Homogeneous PDE?

  • Thread starter Thread starter kgal
  • Start date Start date
  • Tags Tags
    Pde
kgal
Messages
41
Reaction score
0

Homework Statement



Find the solution of:

utt-uxx = sin(∏x) for 0<x<1
u(x,0)=0 for 0<=x<=1
ut(x,0)=0 for 0<=x<=1
u(0,t)=0
u(1,t)=0

Homework Equations



utt-uxx = sin(∏x) for 0<x<1
u(x,0)=0 for 0≤x≤1
ut(x,0)=0 for 0≤x≤1
u(0,t)=0
u(1,t)=0

The Attempt at a Solution



I was thinking that I would need to somehow make a change of variables, converting the PDE to a homogenous PDE and continuing from there. I don't really know how I can do this or solve it at all without some assistance!
 
Physics news on Phys.org
kgal said:

Homework Statement



Find the solution of:

utt-uxx = sin(∏x) for 0<x<1
u(x,0)=0 for 0<=x<=1
ut(x,0)=0 for 0<=x<=1
u(0,t)=0
u(1,t)=0


Homework Equations



utt-uxx = sin(∏x) for 0<x<1
u(x,0)=0 for 0≤x≤1
ut(x,0)=0 for 0≤x≤1
u(0,t)=0
u(1,t)=0

The Attempt at a Solution



I was thinking that I would need to somehow make a change of variables, converting the PDE to a homogenous PDE and continuing from there. I don't really know how I can do this or solve it at all without some assistance!

The strategy is the same as for linear ODEs with constant coefficients: Find a particular solution which satisfies the PDE but not necessarily the boundary conditions. Then add a complementary solution (a solution of u_{tt} - u_{xx} = 0) in order to satisfy the boundary conditions.
 
pasmith said:
The strategy is the same as for linear ODEs with constant coefficients: Find a particular solution which satisfies the PDE but not necessarily the boundary conditions. Then add a complementary solution (a solution of u_{tt} - u_{xx} = 0) in order to satisfy the boundary conditions.

So basically something like Autt-Bxx= sin∏x and Cutt-Duxx=0 and then sum the up into a solution u(x,t)?
 
kgal said:
So basically something like Autt-Bxx= sin∏x and Cutt-Duxx=0 and then sum the up into a solution u(x,t)?

No. You have u(x,t) = u_p(x,t) + u_c(x,t) where u_p is any solution of
<br /> u_{tt} - u_{xx} = \sin \pi x<br />
and u_c is the solution of
<br /> u_{tt} - u_{xx} = 0<br />
subject to u_c(x,t) = -u_p(x,t) on the boundary.

But here it is actually simplest to try a solution of the form u(x,t) = f(t)\sin \pi x, given that (\sin(\pi x))&#039;&#039; = -\pi^2 \sin(\pi x). This reduces the problem to an ODE for f(t).
 
pasmith said:
No. You have u(x,t) = u_p(x,t) + u_c(x,t) where u_p is any solution of
<br /> u_{tt} - u_{xx} = \sin \pi x<br />
and u_c is the solution of
<br /> u_{tt} - u_{xx} = 0<br />
subject to u_c(x,t) = -u_p(x,t) on the boundary.

But here it is actually simplest to try a solution of the form u(x,t) = f(t)\sin \pi x, given that (\sin(\pi x))&#039;&#039; = -\pi^2 \sin(\pi x). This reduces the problem to an ODE for f(t).

So you're saying that I need to set u(x,t)= u_tt-u_xx = f(t)sin(∏x)?
How do I find the function f(t)?

What I did was this:
found u_xx, u_tt and plugged them into u_tt - u_xx = sin(∏x). What do I do with the boundary conditions and the initial conditions (how do they factor in)?
 
kgal said:
So you're saying that I need to set u(x,t)= u_tt-u_xx = f(t)sin(∏x)?
How do I find the function f(t)?

What I did was this:
found u_xx, u_tt and plugged them into u_tt - u_xx = sin(∏x).

What do I do with the boundary conditions and the initial conditions (how do they factor in)?

The conditions u(0,t) = f(t) \sin 0 = 0 and u(1,t) = f(t) \sin \pi = 0 are satisfied automatically.

The remaining two conditions require that
u(x,0) = f(0) \sin\pi x = 0
for all 0 &lt; x &lt; 1 and
u_t(x,0) = f&#039;(0) \sin \pi x = 0
for all 0 &lt; x &lt; 1. This gives you initial conditions for f and f&#039;.
 
Great! I got it!
How would I solve this problem using a technique like changing variables instead of separation of variables?
 
Back
Top