[cfdcheckers]

I am trying to sample the field values on an internal curved surface. The way I managed to sample the values was using the function object surfaceInterpolate in controlDict. For this I created a dummy surface where I needed the sampling and defined it as "empty" physical surface, and correspondingly making it empty in each 0 folder files.

functions
{
    dummySurfaceData
    {
        type            surfaceInterpolate; //surfaceFieldValue;
        libs            ("libfieldFunctionObjects.so");
        writeControl    writeTime;
        writeFields     true;
        surfaceFormat   raw;
        regionType      faceZone; //faceZone;
        name            dummySurface;
        operation       none;
        fields
        (
            (p pNear) (T TNear) (U UNear)
        );
   }
}

If I open the results in paraview I can see the data on the faceZone but I did not find any file written for the dummySurface field values in the case folder. All I see is the pNear, TNear and UNear files in the time folders but the entries for the dummySurface are none in those.


So am wondering if there's any way I can write these sampled field values to a (csv) file?

[piu58]

You may define lines where you get the results. Here an example:

setFormat raw;
surfaceFormat raw;
type    sets;
libs    ("libsampling.so");
interpolationScheme cellPoint;
fields (   U  ); // Fields to sample.


sets ( 
    w0 {  type lineUniform; axis distance; nPoints 500;
       start (  -0.03 0  0 );
       end (  -0.06 0  0 );
    }
    w5 {  type lineUniform; axis distance; nPoints 300;
start (  -0.0295442 0.00520945  0 );
end (  -0.0413619 0.00729324  0 );
    }

    w10 {  type lineUniform; axis distance; nPoints 300;
start (  -0.0281908 0.0102606  0 );
end (  -0.0394671 0.0143648  0 );
    }
}

  );

[cfdcheckers]

Originally Posted by piu58

You may define lines where you get the results. Here an example:

Thanks Uwe for a quick example. I just tried this, in my case it will probably be arcUniform type. However, this method requires me to use start and end angle values. This will not be very convenient when I have changes in the dimensions. I intend to do multiple simulations by changing cylinder diameters. That's why a defined patch name would be very convenient, I thought.

Additionally, I do not want to specify number of cells manually.



Any way out in that case?

[cfdcheckers]

Ok, I tried to parameterise it so that it makes a little more convenient.

But it seems mag(U) is not recognised now. Strange. On the forum it doesn't seem like it's been noticed in recent times.


Update: I realised why. It couldn't recognise the field because it doesn't exist! Found an easy way to tackle this by including #includeFunc mag(U) in controlDict