HEAT solver convergence

dharmatejapdharmatejap Member Posts: 6


I am simulating ohmic heating wire example.

I changed the wire material to Silicon with a resistivity of 2e5 S/m. If I change the thickness of wire to 0.11um from 1um (example value), the simulation fails to converge even at 1V.

When I apply mesh constraint to the volume of wire and set max edge length as 0.01um, the simulation is successful (NOTE: for max edge length 0.1um it still fails). However, it takes a lot of time understandably due to small mesh. It takes forever to simulate for 2V.

Is there a better solution to this so that I can sweep voltages? I am happy with a solution of tolerable accuracy but the simulation isn't completing to get any solution.


  • kghaffarikghaffari Posts: 80Ansys Employee


    In the new setup you describe, the area of the conducting material is much larger, so it makes sense for the simulation time to increase significantly. As you suggested, the best strategy is to work with the mesh size to identify the best value for ideal simulation-time/trade off accuracy (while avoiding convergence errors). You might find that the simulation can still converge with a larger max edge length (e.g. between 0.01 and 0.1 um) which will result in a faster simulation.


Sign In or Register to comment.