Solving heat equation BACK in time

Click For Summary
SUMMARY

The discussion focuses on solving the one-dimensional heat partial differential equation (PDE) backward in time, specifically ∂u/∂t = -∇²u = -∂²u/∂x², with the goal of determining the initial temperature profile u(x,t=0) that results in a uniform temperature u(x,t1)=c after a diffusion time t1. The boundary conditions are convective, defined by ∂u(L,t)/∂x = -h*u(L,t). The user encounters challenges in implementing a numerical solution using finite difference methods in MATLAB, recognizing the problem as ill-posed and seeking alternative strategies for resolution.

PREREQUISITES
  • Understanding of partial differential equations (PDEs)
  • Familiarity with finite difference methods for numerical analysis
  • Proficiency in MATLAB programming
  • Knowledge of boundary value problems and stability analysis
NEXT STEPS
  • Research techniques for solving ill-posed problems in numerical analysis
  • Explore MATLAB's built-in functions for PDE solving, such as pdepe
  • Learn about regularization methods to stabilize solutions
  • Investigate vector field visualization techniques for better understanding of diffusion
USEFUL FOR

Mathematicians, numerical analysts, and engineers working on heat transfer problems, particularly those interested in backward time analysis and numerical methods for PDEs.

Curl
Messages
756
Reaction score
0
I want to solve the one-dimensional heat PDE backward in time ∂u/∂t = -∇2u = -∂2u/∂x2 , x element of [0,L]

Basically, I want to find what the initial temperature profile u(x,t=0) should be such that after some time t1 of diffusion, I am left with the bar at a uniform temperature u(x,t1)=c for c>0 and boundary conditions are convective, i.e. ∂u(L,t)/∂x = -h*u(L,t)

I am having trouble doing this numerically using finite difference in MATLAB, and I realize this is an ill-posed problem. But it seems to be pretty simple so it should be possible. Is there some trick to solving this, or would I have to resort to a brute-force method of guess/check?

I'd appreciate some insight. Thanks!
 
Physics news on Phys.org
It could help to draw a vector field and see if there is a way back. But I assume that it is not unique.
 

Similar threads

  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 13 ·
Replies
13
Views
3K
  • · Replies 6 ·
Replies
6
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 2 ·
Replies
2
Views
4K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K