Solution Diverges while Parallel Processing
I'm running a single-step chemistry, DO radiation modeling, laminar reacting flow case. when I run the job interactively, I don't have any convergence problems. the moment I switch to the parallel solver (6 processors on an SGI Origin 2000 running irix 6.2: background job using the QSUB batch scheduler), the solution diverges.
initially I thought I had very high energy and density relaxation factors, but changing any of the relaxation factors (pressure, momentum, species ....) does not change the divergence phenomenon. the residuals go up in the same pattern - for the continuity the residuals follow the same values each time the iterations begin.
can anybody please point out what is going wrong. thanks a ton.
|All times are GMT -4. The time now is 22:20.|