Hello FLUENT user, In my modelling, the setting conditions are unsteady, laminar flow; I actived energy because I need different working temperatues; two phases air-oil; atomsphric pressure. A simple model but iteration couldn't finish. Residuals increased dramatically then simulation stop. Anybody can tell me what is the problem in my setting? Many thanks in advance.

Anna

Hi!

Try to use the standard approach to the problem of convergence... - Decrease underrelaxation (not under 0.2 or 0.3 - the solution will become unphisical) - try to approach the operating point in a few steps (increase velocity fom 0 to operating velocity) - try to solve steady and then use the solution as initial solution for unsteady - decrease time step (courant number if you are using VoF) -chech the past messages on forum (there has been a lot of discusions on this topic)

Good luck

MATEUS

This could be any manner of things:

1) Poor grid quality, i.e. high aspect ratio cells, cells with very high skew & rapid changes in cell size.

2) Specifying incorrect / non realistic initial boundary conditions and or physical properties.

3) Too high under-relaxation factors.

4) Differencing Scheme issues, i.e. false diffusion.

5) If you're using a near wall (Wall Function turbulence model i.e. K-e) make sure the Y+ is in the correct range. i.e. 30 < Y+ < 500. With 11.63 as an absolute minimum.

6) Trying to solve an unsteady / time dependant problem using a steady state solution method.

7) Using too large a time step when carrying out transient computations.

8) Positioning boundaries (especially outflow boundaries) too close to regions with high flow gradients / recirculation regions.

9) Pressure correction issues.

The list goes on! All the best in obtaining a converged solution, time and patience and you'll track the problem down.

Regards James

Hi

I just want to tell you that even though you obtained divergence check the properties contours and x-y plots.

In other words learn from your case where the flow is giving you unexpected results.

New CFD users tend to play with the computational parameters too much, without looking a previous analysis of the results.

I sugguest that you save the results until the point where the simmulation was convergence and then after 50 or 100 iterations. This approach will provide enough information to tell you if you have a periodic flow.

Best regards

Alex Munoz

Hi thank you guys very helpful suggestions.