CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Continuity Convergence problem (http://www.cfd-online.com/Forums/main/111840-continuity-convergence-problem.html)

crispysalt January 16, 2013 10:57

Continuity Convergence problem
 
Hi,

If all my residuals converge except continuity what can I do to make it converge?

Thanks.

wxmoore January 16, 2013 13:46

Continuity Convergence Problem
 
Check your overall mass balance between the inlets and outlets of the flow domain, the net mass imbalnce should be below about 0.5%. Also check mesh quality to verify it is within recommended software code guidelines for skewness, orthogonality etc. and that you have sufficient mesh resolution especially in areas of high gradients. If you are using enhanced wall treatment and k-omega SST then make sure you have enough inflation layers and the y+ value is near 1. Also try using the coupled solver instead of segregated solver and try switching to second order discretization. Also it may take more iterations before momentum equation begins to converge. As a last resort lower momentum under-relaxation and allow to run for several hundred more iterations. Finally there may be transient flow features that the steady state method cannot resolve, requiring a non steady solution method for convergence.


Good Luck,
William

crispysalt January 28, 2013 04:16

I am simulating a fluidized bed reactor using Eulerian multiphase model, I set two phases in my simulation: solid and liquid. I did patch the initial volume fraction before running. I am currently trying out the laminar flow in transient simulation first. For the boundary condition i set inlet as velocity inlet and outlet as pressure outlet with gauge pressure = 0pa. When i try to calculate, the solution is unable to converge under steady state or transient state. I tried reducing the under-relaxation factors to 0.2 ,reduce the time step and solve using the first order solution method but it still diverge. Are there other ways to make it converge?

monkeytron January 28, 2013 16:54

Have you checked your maximum velocity in your domain to see if it is within expected limits? Often low mesh quality can cause high velocities which can be looked for and can show you where you might need to improve mesh quality.


All times are GMT -4. The time now is 20:54.