Difference in Converge while using CFX and FLUENT
 

Hello everyone
I am writing this post because I am seeing a discrepancy in convergence while carrying out my simulation. So I want to discuss what could the reasons.

I have a simple geometry using hexahedral mesh and having 2 million nodes.
Velocity inlet and static pressure condition outlet is used. Reynolds number is 30000 and Mach number 0.11

My numerical setting in fluent
k-e model with scalable wall functions
Pressure based solver > First order upwind

My simulation diverges in the following case. Even when the relaxation factors are reduced the simulation doesnt converges.


When I used CFX using same mesh and same turbulence model. The solution easily converges to 10^-6 for all residuals.

Is the flow initialized the same way? With same values for both cases?

There are dozens of different settings between CFX and Fluent. The AMG parameters in Fluent have also changed from version to version. The pressure interpolation scheme in Fluent is second order. Unfortunately I've never used CFX to tell you what settings are different, even when they're similar.