|April 27, 2010, 10:55||
Fluent 6.3.26 vs 12.1 and partition method
Join Date: Mar 2009
Posts: 52Rep Power: 10
Thank you for reading my post!
I have some questions concerning fluent 6.3.26 and fluent 12.1. I have tried to find the answers in the manual and on the forum but without satisfying result so far.
First off all I will explain my project setup. I am doing a cfd research on 3D bluff bodies. I use Fluent 3ddp, the realizable k-e model with non equilibrium wall functions, a velocity inlet and a pressure outlet.
I use a hybrid mesh, where the boundary layers are resolved by hexagonal elements, the remainder of the domain is filled with tetras. The hex and tetra elements are one to one connected by pyramids. The mesh is thus conformal.
Fluent is installed on a cluster and runs in parallel. For the calculation 1 node and 8 processors are used.
I had the same mesh, problem setup, boundary conditions, etc run on the cluster in fluent 6.3.26 and 12.1. When I checked the results they differ (drag coefficient f.e.) although the problem setup was the same.
I noticed that the partition method of both versions differs. Fluent 6.3.26 uses the principle axes method while 12.1 uses the metis method.
From the manual I understood that the method should not influence the final result. Although it can influence the computational time. Is this correct? Or does the partition method influences the result (drag value)? Is this due to differences in the interface between the different partitions? Does anyone knows where to find information on the influence of the partition method on the final result?
My second question concerns the differences between both versions of Fluent. Has there been an update in the turbulence models? Or the non equilibrium wall functions? Does anyone know where I could find this information?
In short I am looking for the reasons why both results differ. And mostly whether the partition method can have an influence on the result. I imagine that there have been several updates from 6.3.26 to 12.1 in fluent influencing the results obtained.
Any help or link will be greatly appreciated!
Last edited by Anorky; April 27, 2010 at 11:42.
|Thread||Thread Starter||Forum||Replies||Last Post|
|How to partition cartesian grid for MPI in fortran||CF||Main CFD Forum||3||January 17, 2008 05:38|
|how to do hex mesh in fluent gambit by partition?||john||FLUENT||0||January 16, 2006 05:23|
|Two Processor Fluent Prob||zxaar||FLUENT||0||November 26, 2004 01:25|
|Parallel Fluent on Linux clusters||Chengwang Lei||FLUENT||3||November 24, 2004 03:05|
|Ignore cells on partition boundary||Karl||FLUENT||7||May 11, 2002 22:12|