FluxSolver results

Numerical methods and mathematical models of Elmer
Post Reply
aminer2k
Posts: 20
Joined: 14 Dec 2009, 00:34
Location: Northern California, USA

FluxSolver results

Post by aminer2k »

Hello,
I am getting results from FluxSolver that I can not improve. I have a very simple example problem where there is fixed temperatures on both sides of a block. The block has cubes of metal on both ends (1x1x1mm) and a block of low thermal conductivity material in the middle. The ends are fixed temperature of 300 and 400C. I compute the temperature field with the Heat Solver, and then compute gradT with the flux solver. The gradT results have large spikes that are not correct, near the boundaries. These spikes do not improve by increasing grid density, decreasing the convergence criteria, changing preconditioning, or by using an indirect solver vs. direct solver. Please see attached. Any idea what is happening? The temperature distribution is fine, but the gradT result is not.
Thank you
Attachments
2d plot of gradTz
2d plot of gradTz
Picture 4.png (12.39 KiB) Viewed 1974 times
3d view of gradTz
3d view of gradTz
Picture 5.png (47.41 KiB) Viewed 1974 times
2d plot of T(z)
2d plot of T(z)
Picture 6.png (11.13 KiB) Viewed 1974 times
Post Reply