Learning Forum

[moranlouie]

I am trying to simulate a Concentrated Solar Power configuration using the Rosseland Ray-Tracing method. The geometry on the solar receiver and dish looks like this:

[Keyur Kanade]

Can you please upload some images of your mesh. nDid you use hex dominent mesh method in meshing? nPlease use tetrahedron method to mesh the geometry. nRegards,nKeyurnGuidelines for Posting on Ansys Learning ForumnHow to access ANSYS help linksnn

[moranlouie]

Thank you, Keyur, for answering. Here is the images of the mesh.nI used a sweep and sizing for the tubes and the more regular geometry in the simulation. I used a multi-zone mesh on the dish. I followed your advice in my previous question to remove sharp corners in the cavity domain. A tetrahedron mesh is used in the irregular faces where the tubes exit.nn

[Rob]

There look to be a few jumps in cell size. However, the interface zones look to be giving all the warnings so I'd start by working though those and checking the definitions. I've never used the Rosseland model, I tend to use S2S or DO for most things, and probably should look at the new(er) Monte Carlo method. n

[moranlouie]

Now it's doing this:999999:mpt_accept:error:accept failed:No such file or directory

[moranlouie]

I found the answer to my last question. I was setting Fluent to solve with 4 cores and 1 graphics processor. I know that our laboratory has an NVIDIA graphics unit but Fluent can't seem to access that and use it to solve the simulation. Setting graphic processor units back to 0 removes the 999999:mpt_accept:error:accept failed:No such file or directory error.n

[Rob]

The Student licence doesn't have a gpu process on it. Otherwise there should be a more useful error just above the 9999etc line which tends to explain what has failed. The 99999etc line is the nodes crashing following the initial event. n