Cluster and local computer giving different results for reactive flow
 

Hello
I am running a reacting flow of a premixed hydrogen-air mixture into a combustion chamber to simulate a flame in Ansys Fluent. This simulation is steady-state in a 2D-axisymetric geometry. The species its mixture properties are imported through a CHEMKIN file, using the Konnov (2019) mechanism.

When running the simulation through a cluster the results start to differ compared to running on a local computer. When ran through the cluster the flame profile start to get unrealistic and incorrect (and sometimes blow off happens where the flame should be stable, or the flame is stable where flashback should happen), but when running the exact same file on the local computer the flame is realistic and correct.

I attached images showing the difference. The temperature contours are on the same scale and the heat of reaction contours are also on the same scale. The cluster simulations where the images are taken from is simply continued on a previous done converged simulation on the local computer (converged local column in image) for about 20.000 more iterations.

Important to know is that I do not change anything in the computational domain. The boundary conditions, species, mesh etc are all the same. I simply just read the data file of the previous simulation and run it for more iterations. The way the CHEMKIN files are read are the same, I checked this by comparing the reaction properties in Ansys Fluent for the cluster and the local computer.

The version of Ansys Fluent and the precision is also the same.

Additionally I have tried:
-Using different amount of CPU’s (1, 4 and 16), which also does not help. The local computer also uses 4 CPU’s.
-Starting from the beginning. So first a cold flow on the cluster till it was converged and afterwards adding an ignition. This also did not make a difference.
-Trying a simple combustion, in order to make sure that there isn’t something wrong with the cluster itself. The results for cluster and local were the same.

The cluster runs on Linux, while the local computer runs on Windows.

I was wondering if anybody knows why the problem occurs, or where I have to pay more attention to in order to find the source of this problem. And, if I am very lucky, if someone knows the solution to this problem :)

as you are doing everything on cluster, you should have journal file to perform simulation
try to run same file on local machine

are you reading files as input data somewhere, that could be the reason also

I've never used Chemkin, you may try to find documentation about parallezation of that software

frankly speaking, there should be some difference but tiny. I've never seen something similar to your case if everything was OK with input files.