Problem with the convergence for CFX solve
 

Hi.
I'm architect learning a bit about aerodynamic and Ansys.
I want to simulate the wind along a street with a small barrier (wall) in the middle using CFX. So, following some tutorials and posts found here, I built a domain, the mesh, set up the boundary condition (wall no slip for ground, front plane and black plane, inlet, outlet, and opening for the sky) and ran a simulation with a good result. Here I have some screenshots of the mesh and results:
http://imgur.com/a/2lEQn
http://imgur.com/a/ZChY3

The problem is that I wanted to change the inlet wind (originally a normal velocity of 6m/s) for a parabolic wind velocity profile from the inlet face in the domain using these expressions:
Dt= 2[m]
Ue= 6[m s^-1]
Vel= Ue*(y/Dt)^0.36
http://imgur.com/a/kMxzf

But now the CFX solve doesn't have convergence.
http://imgur.com/a/vJ1or

Someone understand what should I change or improve?

You simulation is converging, it's just doing so in a slower rate. Let it run for a few hundred iterations (not just the 100 CFX uses as default).

You might also want to play with the timestep you're using. Set a physical timescale to a fraction of lenght_scale/velocity_scale, which for your case I suppose would be 0.1/3 to 1/3 s.

Thanks for your response, brunoc.
I'll try your suggestions and will publish here the results.

Hi, I followed the suggestions but the solver doesn't reach a convergence, yet.
http://imgur.com/a/zlVKX

Originally, with a inlet wind of 6m/s (flat velocity profile) 50 iterations were enough to get a convergence. I only changed the speed wind parameter for a expression to get a parabolic wind velocity profile, but now I don't have convergence at all.

Any other suggestions?

Have you also tried changing the timestep?

This FAQ gives a range of suggestions: http://www.cfd-online.com/Wiki/Ansys...gence_criteria

brunoc, ghorrocks, thanks for you help. I fixed it the problem. It was my fault, because just now I'm learning Ansys. I had put the expression Vel in a wrong field in the Inlet's details.
Now, I have a good convergence, again.

Thanks