Register Blogs Members List Search Today's Posts Mark Forums Read

 April 21, 2003, 10:26 PLEASE HELP _ FREE CONVECTION #1 DORDIKASH Guest   Posts: n/a Hi please help me . I want to model free convection between two vertical plates with diffrent temprature . It doese not converge . and after too many itteration it seens that it could not get the flow field. and it has bad answers . I use fluent V5.2.3 .

 April 21, 2003, 13:22 Re: PLEASE HELP _ FREE CONVECTION #2 emre Guest   Posts: n/a If the problem is momentum equations which are not solved, you should "trigger" the flow. You could do this in many ways. The one I prefer is to give a y-velocity(assuming -y is the gravity direction) which is very close to zero like 0.001m/s, at the initialisation panel. If you are already doing this, start with lower gravity and increase it gradually.

 April 24, 2003, 00:20 Re: PLEASE HELP _ FREE CONVECTION #3 DORDIKASH Guest   Posts: n/a DEAR FRIEND : I started with lower gravity about .009806 , after some itteration it did converged and gave me a good field of flow , but when I increased the gravity gradually , it began to diverge , and with lower under_relaxation factors I could get no improvment . I should say that my problem contain radiation , conduction in solid and convection ,and it seems that I should use too many cells . What do you think about methods I should use in my solution . segregated or coupled ,implicit or explicit and ... . please help me .

 April 24, 2003, 02:08 Re: PLEASE HELP _ FREE CONVECTION #4 emre Guest   Posts: n/a I would use segregated solver. You should have a mesh which is fine enough, but having too many cells may result in instabilities. Also note that changing the gravity changes the Ra number. Having Ra higher than 1e9, you should switch to turbulence. Regards