Validation 12.1 vs 6.3, Difference in Reported Inlet Total Pressure
 

We are switching over from Fluent 6.3 to Fluent 12.1 in some of our large design projects. Before doing this we want to validate and compare results obtained with Fluent 12.1 with results previously obtained with Fluent 6.3.

When comparing mass-flow averaged total pressure loss (P0in - P0out)/P0in for an old design we noticed a difference between 6.3 and 12.1. Looking further at this we discovered that the mass-flow average total pressure on the outlet was almost identical, but the inlet value was different! Normally you would expect the opposite - the outlet might be different due to modifications of models/schemes/numerics, but the inlet should be identical!

Looking further at this we tried to make an iso-clip just at the inlet. Computing mass-flow-averaged total pressure on this iso-clip at the inlet gives exactly the same reported total-pressure from Fluent 6.3 and 12.1 (including all decimals)! Hence, Fluent 6.3 and 12.1 uses the set inlet total-pressure boundary condition profile in the same way, but still reports different total pressure on the inlet plane.

My conclusion is that Fluent 6.3 and Fluent 12.1 extrapolates the inlet boundary values differently. If we compute area-averaged values instead the difference is halfed, but it is still there. My interpretation from that is that the error affects both the inlet mass-flow and the inlet total-pressure extrapolation.

Anyone had any similar problems? I am not sure if Fluent 6.3 or Fluent 12.1 reports the correct value. We will have to go back and look at some old validation cases with measurements to judge that and compare results for a more clean case.

Hi Jonas,
Thanks for the post. It was nice to read. I did a validation a while ago but I use a jet impacting a particle bed (multiphase flow, turbulent flow). I found that there was not a significant difference.

One thing that has really bothered me is that even with the change from 6.3 to 12.1 none of the problems with the multiphase flow solver were corrected. For example, the solver calculates particles moving too fast along the axisymmetric boundary in version 6.3, this STILL hasn't been fixed.

I think when Ansys bought fluent, they pretty much just slapped a new number on it and added a few extra percs, but I HATE the new gui and refuse to use it.

Casey