Turbomachinery Convergence Issue
 

Hi All,

I have been working on Centrifugal Compressor project and I have an important questions about it. I hope the experts in this field will help me, thanks in advance.

1- When I just use TurboGrid default mesh about 200.000 it will quickly convergence and give some reasonable results but if I check y+ value its about 700 and it is not right for cfx or fluent suggested range , I tried for finer mesh and it shows that more good quality resolution around boundary layer its going to more unstable and diverge after few iteration. I can't decrease y+ value without increase mesh count. So how could I solve this problem ? Here is some things that I tried..

* I searched for it and a lot of people say that you should decrease physical time step but it did not work.

* First I solve problem with coarse mesh then I use flow field variables as an initial condition for fine mesh but it did not work too.

* I use some expert parameters and try to decrease under relaxation factors but it did not solve the problem either.

So what should I do for get y+ about 70-100 range ?

Here is case conditions that I use ;

Inlet : Stationary Frame Total Pressure
Outlet : Mass Flow Outlet ( Design Conditions)
Others are classic TurboMode boundaries (Periodicity,wall..etc)

Material : Air Ideal Gas
Turbulance Model : k-epsilon (I did not want to solve boundary completely so ı just use this)