[mgg]

Hallo everyone,

I got a problem of sampling surface. I am running a pipe flow with cyclic bc (DNS). When the flow is fully developed turbulent, I did a run-time surfacesampling of inlet. The sample function in the controlDict is as following described. The problem is that the number of sampled points/values (4846) does not match the number of actual faces of the patch (6400). This 4846 values are the results from 128 cores parallel computation. And with other core numbers, this number will also change. I also tried with 2 cores, then the number is 6400. My decompose methode is scotch. I also tried other methode, with simple the sampling function can not sample anything. Could anyone explain that?

        surfaceSampling
    {

        type surfaces;

        functionObjectLibs ("libsampling.so");
        enabled         true;
        outputControl timeStep;
        outputInterval 10;
        interpolationScheme cellPointFace; 

        surfaceFormat    foamFile;  
        fields
        (
            U
            
        );
        surfaces
        (
           inlet
          {
        type            patch;
        patches         ( inlet );
        interpolate     false;    
        triangulate     false;
          }
        );
    }
  }

[wyldckat]

Greetings mgg,

This sounds like a bug. A few questions:

1. Which OpenFOAM version are you using?
2. Are you able to reproduce this bug with a simple case, perhaps one based on OpenFOAM's tutorials?

Best regards,
Bruno

[mgg]

Originally Posted by wyldckat

Greetings mgg,

This sounds like a bug. A few questions:

1. Which OpenFOAM version are you using?
2. Are you able to reproduce this bug with a simple case, perhaps one based on OpenFOAM's tutorials?

Best regards,
Bruno

Hallo Bruno,

thank you for your reply. I attached a pic to descreibe my problem. I sampled the inlet patch of of a cylinder, with has periodic bc between inlet and outlet. You see, there are a lot of mesh values missing. I can figure it out.

[wyldckat]

Hi mgg,

Unfortunately your response does not answer my questions, therefore I'm unable to help you, simply because there is no clear way for me to try and reproduce the same problem.

Best regards,
Bruno

[mgg]

Originally Posted by wyldckat

Hi mgg,

Unfortunately your response does not answer my questions, therefore I'm unable to help you, simply because there is no clear way for me to try and reproduce the same problem.

Best regards,
Bruno

Hallo Bruce,

I solve the problem with another way. I use cutting plane in sample Dict and disable triangulation. Then I got exactly the same cell number as it should be. Because I only need fully developed turbulent flow, it does not matter where I cut it. So I cut in the center of the pipe instead of boundary patch. (Is that beacuse patch is in the boundary, so it is hard for the sampling utility to get the value?) I do not know why sample patch did not work properly, but anyway I got what I want. Thank you for your help anyway.

My OpenFOAM is 2.2.2.