Discussion Forum

Hi all

As a morning quiz (for me) I want to submit a little example of diffusion equation solving (HT for the example below in 4.3a) and overshoots.
I see often people claiming COMSOL is giving an error, or has a bug, (ther might still be but I do not know today of any, and for me this is no bug) because they get un-physical results such as negative concentrations or negative temperatures.
Personally I call it a numerical feature, because it is perfectly explainable and proves rather a lack of understanding of the user who is setting up the model and not validating carefully the results.

The example is a simple bar of aluminium initial temp 0°C held at two fixed temperatures hot 100°C and cold 0°C. Its calculated in the time domain with a coarse mesh on purpose to show the mesh and initial effects due to the initial strong gradient at the Thot side.

I have also included an energy balance estimation

It's a simple model but that contains quite some elements one need to consider for a diffusion equation (erf() solution types) case, and I hope I haven't missed something important, but you should be my warrant to correct that ;)

Now I leave it to you to find the answers (note this is a Results/post-processing effect, but it appears also during solving, my example is not sever enough to block the solver to find valid solutions, so it converges in this case)

--
Good luck
Ivar

ps: I seem to have trouble uploading files from Opera onto the COMSOL site ;)