I am facing a peculiar problem. i start my cold flow simulation with a segregated steady state solver and then switch to a coupled solver for mixing-reaction simulation. due to this shift and the fact that i am using steady-unsteady route, residual definitions change and all residuals go northwards. i have been monitoring all primitive variables and find that they are under control and varying slowly. my intermediate results are also physically meaningful. to advance calculation in time i have STOPPED MONITORING ALL residuals and allowing calculation to advance to next time step with a pre-specified number of iterations per time step. is it OK to turn off all residual monitors and judge convergence based solelu upon variation of primitive variables? thanx for any advice/suggestion in this matter.

Swarup.

That's what I do for my case, bubbly pipe upflows. I do not trust the residuals alone. I monitor relevant flow quantities and wait for them to level off. But I cannot be sure whether this will apply to your case as well.

i am treating a case which is unsteady for entire flow time, i.e. it never reaches a steady state during the flow time. so, i don't think i will be able to see "levelling off" of any flow quantities in my case. the problem is i do not have good experimental data to compare predictions and have to rely solely upon understanding of physical phenomena inside the reactor to verify results. swarup.

I spoke too soon without reading your steady-unsteady shifting. Sorry.

Oh, and if you're not aware, here is a route to aca-support. They may be able to help. (Monday to Thursday, 9:00 PM to 10:30 PM, EST)

http://im.chatanywhere.net/chatanywhere.php?u=1603

Hope that helps...

Best Regards, Srinath Madhavan