Heat equation - separation of variables

Click For Summary

Homework Help Overview

The discussion revolves around the heat equation, specifically the problem of solving the equation du/dt = d²u/dx² with given boundary conditions and an initial condition involving sin²(x). Participants are exploring the separation of variables method and its application to the problem.

Discussion Character

  • Exploratory, Conceptual clarification, Mathematical reasoning

Approaches and Questions Raised

  • Participants discuss the separation of variables approach and the resulting expressions for u(x,t). There is a focus on how to satisfy the initial condition u(x,0) = sin²(x) and the implications of Fourier series in this context. Questions arise regarding the relationship between the derived expressions and the initial condition.

Discussion Status

Some participants have clarified the distinction between the two problems presented, while others are examining the implications of boundary conditions on the solution form. There is ongoing exploration of Fourier coefficients and their role in satisfying the initial condition.

Contextual Notes

Participants are navigating the complexities of applying boundary conditions and initial conditions in the context of Fourier series, with some expressing uncertainty about how to reconcile the derived expressions with the initial condition. The discussion reflects a mix of understanding and confusion regarding the setup of the problem.

Gekko
Messages
69
Reaction score
0

Homework Statement



du/dt=d2u/dx2, u(0,t)=0, u(pi,t)=0

u(x,0) = sin^2(x) 0<x<pi

Find the solution

Also find the solution to the initial condition:

du/dt u(x,0) = sin^2(x) 0<x<pi


The Attempt at a Solution



From separation of variables I obtain

u(x,t) = B.e^(-L^2t).sin(Lx)

For the boundary condition u(pi,t)=0, u(x,t) = Bm . e^(-m^2t) . sin(mx)

Finally for u(x,0) = sin^2x

u(x,0) = 4(cos(m.pi)-1).e^(-m^2t).sinmx / pi(m^3-4m)
found through superposition and Fourier sum of sines

What I don't see is how to solve the initial condition because when I diff. above wrt t and set t-0 I obtain

du/dt u(x,0) = -m^2 4(cos(m.pi)-1).e^(-m^2t).sinmx / pi(m^3-4m) = sin^2x

How can this be set to sin^2x by choosing a value of m?
 
Physics news on Phys.org
I'm not sure what you are talking about here. You understand that these are two different problems, don't you? You solution "for u(x,0) = sin^2x

u(x,0) = 4(cos(m.pi)-1).e^(-m^2t).sinmx / pi(m^3-4m)" is a sum isn't it?

That is
u(x, t)= \sum_{m=1}^\infty e^{-m^2t}(4(cos(m.pi)-1).e^(-m^2t).sinmx / pi(m^3-4m)).

Now, with initial condition u_t(x, 0)= sin^2(x) you must have
u_t(x, 0)= \sum_{n=1}^\infty -m^2 B_m sin(mx)= sin^2(x)

Find Fourier sine coefficients for sin^2(x) and solve for B_m.
 
Gekko said:
From separation of variables I obtain

u(x,t) = B.e^(-L^2t).sin(Lx)

For the boundary condition u(pi,t)=0, u(x,t) = Bm . e^(-m^2t) . sin(mx)
Just to be clear, separation of variables gives you

u(x,t) = e^{-L^2t}(A\cos Lx + B\sin Lx)

Then applying the boundary condition u(0,t)=0 gives you A=0, and applying the condition u(π,t)=0 requires L be an integer.
 
Understood now. Thanks a lot for clearing that up
 

Similar threads

  • · Replies 11 ·
Replies
11
Views
3K
  • · Replies 15 ·
Replies
15
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
5
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 5 ·
Replies
5
Views
1K
Replies
1
Views
2K
Replies
8
Views
1K