Numerical instabilities introduced after printout
I'm running a simulation with intermittent printouts using TOUGH3. I noticed that, after some of those printouts, numerical instabilities are seemingly introduced that stall the remainder of the run. When I remove these printout times, and the simulation proceeds uninterrupted, they appear to go away. Has anyone seen something like this before? And are there any hunches as to the cause of this?
I think this problem has always been there for TOUGH2. It seems you have a very simulation that is very sensitive to time steps. Is it due to phase change? You probably need to check if some of the new MOMOP options help to make the runs more robust, or some solvers options in PetSc.