As I work more trying to figure out this whole Fipy thing, I thought I would try to reproduce some results from an ANSYS example on heat conduction.
I used paramaters in the ANSYS example for 2D conduction from here: http://www.me.cmu.edu/academics/courses/NSF_Edu_Proj/ThFlEngr_ANSYS/htm/2D_1_generic.htm And here is my corresponding Fipy code: http://pastebin.com/f41cebac8 I only have one problem with that example, and its related to the same problems I mentioned in my previous email. In my Fipy code, the only way I can achieve the same temperature distribution as in the ANSYS example, is if I set the diffusivity coefficient equal to the thermal conductivity of the material (i.e. 20. W/(m-K) ) when in reality, I think it be given a value of k/(ro-Cp) (~1.172 e-5 for steel). If I do this, and run the simulation out to 100 seconds (100,000 iterations), I get a fully saturate temperature profile at Tmax. However, if I set the alpha coefficient to 20., my answer agrees with the ANSYS example, and only takes ~200 iterations to reach steady state. That's the equivalent of ~20ms given my step size. That seems far too fast for this particular problem. Could it be that I am missing something, or that perhaps the ANSYS example is erroneous in that they didnt define the density and specific heat of the material? Thanks for any pointers! Cheers, Chris
