[lhlh]

I am working on a simulation about hydrodynamics of gas-particle system in a riser, and the solid volume fraction profile is an emphasis. My question is: is the default residual setting of 1*10-3 adequate to my project, if not, what are generally they?

Another question about the setting of residua is:

I try to solve a transient problem using the setting of residual at 1*10-4. Since the problem is transient, the beginning of solution is very hardly converged with this setting.

Is it possible that I can change the setting of residual to 1*10-3, and let the solution converged, once the main variable such as velocity or solid concentration is stable (no change in further iterations), and then change the residual setting back to 1*10-4. After monitoring the change of main variable, if the variable is again stable and converged with this new setting, can the solution obtained in this way be counted as a converged solution with the residual setting of 1*10-4?

Thanks for any reply,

[LuckyTran]

It is better to determine convergence by monitoring flow variables. If you have already determined that the flow variables are converged, there is no need to switch back to checking residuals. Your solution is already converged so why check again?

[lhlh]

It is because that I want to get a converged solution at the level of residual of 10-4, not just only 10-3. Smaller convergence criterion might have a better accuracy for the simulation result.