CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   divergent of a FSI problem (https://www.cfd-online.com/Forums/fluent/121806-divergent-fsi-problem.html)

edooo0 August 5, 2013 17:54

divergent of a FSI problem
 
Hi,

I am now doing a FSI model by Ansys Fluent (14.0). My model is a multi-rigid-body system, linked by rigid joints, in water. The structure part has been written by UDF code, and complied with Fluent fluid Solver. The dynamic mesh feature is conducted.

The problem that I am facing now is about the divergence of the fluid solver. This problem only occurs when the mass of the rigid body is relatively small, for example, the density of the body is similar as the density of water (the mass of body is same as the mass of water in the same volume/area). When the mass of body is much bigger than water, the convergence is very good by any solution method, but when the mass of the rigid body is same as that of water, it is very hard to achieve convergent.

I have tried almost all the solution methods in Fluent for the small mass case. With the UDF code, the dynamic mesh is activated, therefore in the solution methods panel, only “first order implicit” method of Transient Formulation can be selected. Firstly, I tried all four pressure-velocity coupling schemes (SIMPLE, SIMPLEC, PISO, Coupled) with small under-relaxation factors (as small as 0.1), while none of them get convergent. Secondly, I tried to active “Non-Iterative Time Advancement” under the first order implicit transient formulation. In this way, only the “Fractional Step” scheme with small relaxation factors (0.1, 0.1) can make the simulation convergent. However, the accuracy is not very good.

I have tested the smaller time steps, finer mesh, different boundary conditions, while it seems they are not effective for my problem. I believe the “fractional step” scheme can work better along with my UDF code. I think if I can achieve the convergence with a relaxation factor as high as 0.7, 0.7 for pressure and momentum, maybe the results could be better. But one of the questions is that how I can get convergent by the “fractional step” scheme with relative higher relaxation factor. I have read the “advanced solution controls” dialog box in the users guide, and tried to modify some of the advanced settings, but did not improve much. Could anyone can give some advices on my situation?

Look forward for your valuable suggestions and thank you very much for your time.

Best Regards


All times are GMT -4. The time now is 20:44.