I have a laminar premix combustion problem going. the steady solution converges easily (the energy part converges slowly though). I start the unsteady simulation with the following paramters and I get recirculation at the exit:

second order pressure; SIMPLE pressure-velocity coupling; second order upwind for energy, velocities (2D), species, density and DO (discrete ordinates) radiation

the recirculation continues to grow (number of faces at the exit) and some of the residuals start going up (energy, y-velocity, methane) and every time-step some more faces have recirculation. I've tried adaption, first order schemes, lower time steps, lots of iterations per time step, but everything fails eventually.

guys, any suggestions? it's kind of urgent as my advisor is breathing down my neck. I've absolutely no idea what the problem is. I've shortened the combustor by 3 inches and nothing else (boundary conditions, flow conditions etc) has changed. the original grid gives me near perfect (!) results, but this modified grid refuses to yield any solution at all. the steady solution seems to be good because it matches the steady solution part for the first grid.

HELP!

Did you try to use the coupled solver of fluent6?

I am surprised that you got converged results for a prmixed combustion with the segregated solver. I always failed. The best hint then was to use the coupled solver. Maybe it 'll help you too even for unsteady simulations..

volker:

I've never been able to work with the coupled solver. never got the residuals to come down, especially the temperatures, both steady and unsteady.

I've always been able to work with the segregated solver.

thanks anyway

prateep

May I suggest that you run an unsteady, non-reacting flow.. This should decouple radiation and reaction complications. Try to start simple. Progressively add components! Good luck.

hi alammar!

the unsteady solution finally worked! I had to increase the mass flow rate first to eliminate the recirculation at the exit and then slowly reduced the mass flow (it's made up of high frequency sine waves). the recirculation came back, but was restricted to fewer faces near the exit.

it has dissapeared now and I'm able to get good convergence for all the residual values.

thanks for the suggestions

prateep