[lanran214]

Hi, everyone:

For the welding simulation, I want to apply a Gaussian heat flux on the free surface cells. I modified the rhoe(ijk) density*specific of the free surface cells in the qsadd.F90 subrutine, the temperature could increased correctly at first several seconds, but when the temperature exceeded the melting point, the solver run into an error, the time step and stability limit was not stable, and the fluid even disappeared in the 3-D postprocess interface, what's wrong with it?

Sometimes it shows the error message: heat transfer/conduction flux exceeded stability limit.

Is the error caused by the frees surface cells? and why the error occurred when the temperature exceed the melting point exactly?

Thank you.

[xuesong]

I think you can try a two-fluid simulation. The one fluid with void is less physical, maybe there's some uncovered technical details.
---- 2020.12.31-----
I find i didn't activate the surface tension model. Now it works with one fluid with surface tension model on. There's no negative temperature error occurred.