convergence about SST model
 

hello,everybody! recently i have compute the flow and temperature field about a U-bend channel (using the CFX-11.0). the turbulence model is SST. if i the mesh is coarse, the compute converged very well. but if i refin the mesh. it can't converge now. i have read the previous message about this problem. some people advised to larger the timescale. but i don't know how to do that.can anybody help me?

my setting is like this: the timescale control is autotimescale control the length scale option is conservation timescale factor is 1.0 max iteration is 500

Re: convergence about SST model
 

sorry, my setting is like this: the timescale control is autotimescale control; the length scale option is conservation; timescale factor is 1.0; max iteration is 500.

Re: convergence about SST model
 

Hi,

This issue has been discussed several times recently. Do a search for plenty of comments.

In short use a larger physical timescale.

Glenn Horrocks

Re: convergence about SST model
 

have a try of seting "length scale option" to "aggressive". it will make your calcualtion more robust.

Re: convergence about SST model
 

Hi, Horrocks. thank you for your answer! Because i have never use the "timescale control" option of "Physical Timescale". i stil use the "auto timescale". so i don't know which value of the "Physical Timescale" i should choose. could you tell me how to set it? 1s or 2s? thank you!

Re: convergence about SST model
 

thank you! frank. I will have a try as your suggestion. then, i will tell you the result. Regards!

Re: convergence about SST model
 

Hi,

Start off with the timescale the auto timescale uses. Then start increasing it. After the initial few iterations you can usually increase it lots (10 times is normal, 100 or even 1000 is possible).

As discussed in the documentation, a guide for timestep size is set it to be equal to the average residence time of fluid in the domain.

Glenn Horrocks

Re: convergence about SST model
 

Hi, frank. I have a try using the "length scale option" of "aggresive". but the effect is not clear. the calculation can't convergent still:(