How to fix my case file

[elinlarsson744-ux]

Hi, I have a Chanel geometry, where zone 2, zone 5 should be inlet outlet and rest should be no slip walls, but when I run the solve with this case file I always get float point exception, can someone suggest a fix in my case file. Thanks in advance

[timfelle]

Could you post the case file in a code block instead of a screenshot? That way, one can copy it and try it to help debug this.
Enclose it in:
```
Paste here.
```

[elinlarsson744-ux]

Okay, here is the case file

{
  "version": 1.0,
  "case": {
    "mesh_file": "chanel_mesh.nmsh",
    "output_at_end": true,
    "output_boundary": true,
    "output_checkpoints": false,
    "nsamples": 10,
    "numerics": {
      "polynomial_order": 5,
      "time_order": 3,
      "dealias": true
    },
    "fluid": {
      "scheme": "pnpn",
      "Re": 5600.0,
      "boundary_conditions": [
        {
          "type": "velocity_value",
          "value": [3.0, 0.0, 0.0],
          "zone_indices": [2]
        },
        {
          "type": "no_slip",
          "zone_indices": [3, 4, 6, 7]
        },
        {
          "type": "outflow+dong",
          "zone_indices": [5]
        }
      ],
      "initial_condition": {
        "type": "uniform",
        "value": [3.0, 0.0, 0.0]
      },
      "velocity_solver": {
        "type": "cg",
        "preconditioner": {
          "type": "jacobi"
        },
        "projection_space_size": 0,
        "absolute_tolerance": 1e-7,
        "max_iterations": 800
      },
      "pressure_solver": {
        "type": "gmres",
        "preconditioner": {
          "type": "hsmg"
        },
        "projection_space_size": 5,
        "absolute_tolerance": 1e-3,
        "max_iterations": 800
      },
      "output_control": "simulationtime",
      "output_value": 0.1
    },
    "time": {
      "end_time": 1.0,
      "variable_timestep": true,
      "target_cfl": 0.4,
      "max_timestep": 1e-4
    }
  }
}

[timofeymukha]

In addition, please state what version of Neko you are running, and on which hardware (CPU or GPU, etc). Also, please post the trace of the error in the log.

[elinlarsson744-ux]

done there as well thanks. Also for reference, I have used this library https://github.com/Nek5000/Nek5000/tree/master/tools/exo2nek

[timofeymukha]

Not a lot of luck here. I can more or less with 100% certainty confirm that this is a mesh issue. Not in terms of quality, the mesh in nice. But rather there is something going wrong in the conversion, but we do not catch it. I don't have any experience with ICEM mesh conversion. But I think @timfelle has used exo2nek successfully.

What one should do is try to setup a Nek5000 case with this mesh and see if that runs. That would at least identify which conversion went wrong. It's a bit bad timing for me, I gotta leave now and will be away for a week. In case noone else has time to help you out, it'd have to wait until then.

[elinlarsson744-ux]

No problem at all I also think it might be when I export from ICEM it is by default hex8 but in some papers I saw hex27 is preferred, so now i will talk to the ANSYS guys how to do that, meanwhile I will wait maybe @timfelle have some solutions for me. Thanks for the help though. Much appreciated.

[vbaconnet]

I use a python script that converts ICEM mesh to rea which I can share with you. From rea you can then use reatore2 to convert to re2 and then rea2nbin to convert to the neko mesh format. You may try it to convert your ICEM mesh and see if it works with Neko.

Here I attach two files:
mesh_to_rea.py, which is the main file you will execute.
mshconvert.py, which contains the implementation.

Make sure to change the name of the mesh file you want to convert in mesh_to_rea.py, and then you should be good to go with

python mesh_to_rea.py

Disclaimer, this is a very old converted that did not write myself, it works for my case but can't promise it will work for you. And unfortunately I cannot offer any help with it...

[elinlarsson744-ux]

The code really worked and now my case is runing, although the import of numpy is missing, thank you so much for helping.