Transient Combustion Simulation

[sjohn]

I have been running a transient combustion simulation for the last five days. I had been varying the time step between 1e-07 s to 1e-06 s and monitoring it regularly. Last night, I kept the time step size at 5e-07 and the residual behavior got worse as the case ran overnight. It comppleted 5594 time steps this morning. Can I lower the step size and continue running this case from 5594? I did have the case served at an earlier time step 5280 (prior to residuals getting worse) so I could start from that time step.

[Rob]

Not really as your (re)starting point might not be valid. But... Is the aim to see the equilibrium state or see what happens on the way to equilibrium?

[sjohn]

Hi Rob, The aim is to see the equilibrium state. This case is similar to the case we had discussed yesterday but run at a different inlet temperature. In this case, the species concentration did not reduce to zero, when I had switched from steady state to transient.

[Rob]

OK, thanks. I'd not spotted who'd posted the topic. 

If the solution is still physically sensible you can probably continue on as you're just updating what you currently have with time. IE if it's a bit wrong it'll just take more time steps to be less wrong and at equilibrium, if converged, it'll be right. 

[sjohn]

Before I got to see your reply, I had restarted the simulation with a smaller step size (2e-07). Here is the improvement in the residual behavior.

 

The pressure monitor showed some minor spikes but is stable now.

[Rob]

I'd also monitor fluxes and point data to see how that's behaving.