How does ANSYS calculate the temperatures?

[MaxTemp2]

Dear all,

I am trying to debug one of my problems (see: Thermal transient analysis with ANSYS), and I need help. Is there a possibility to exactly see the equations, matrices etc. at each node with/from which ANSYS calculates the temperatures at the nodes?

Thx
MaxTemp2

 

[IRstuff]

Unlikely, although there might be something a white paper somewhere. The equations are usually hardcoded and optimized, since they'll form the bulk of the processing latency.

It might be worthwhile putting something like a sheet of diamond of the appropriate thickness, fixing the far side to 150º, and then see what happens at the diamond/steel interface.

TTFN

FAQ731-376

 

[davefitz]

Its been 30 years since I've programmed finite element programs for thermal and stress problems, but the fundamamentals haven't changed. Selection of basis functions, optimizing the inversion of sparse but not well proportioned matrices, and stability criteria based on mesh size and unusual boundary conditions might vary for specific and unusual problems.

My guess is the stability criteria for the case of a convective boundary condition with a small mesh size implies a very small time step is neccesary- you might try the standard trick of comparing the answer for the cases of different time steps, and check to see how the answer converges toward a common answers as the time step is reduced. The stabiklity criteion is a function of the method of integration, and the ANSYS manual should provide that guidance.