Power Field Function Missing?

[kitamuraa24]

When the ' power_field_function_on ' argument is set to True, under the lbs options, it seems that no such field gets exported to the PVTU.

[ragusa]

@andrsd David, can you take a look as of why the power is not one of the fields in the vtu? if you need to test, you can use a fast-running case from the test suite (in Keigen). Thanks

[andrsd]

GetScalarFieldFunctionList only returns field functions with flux moments.
If you need power density, call phys.GetPowerFieldFunction and export that into VTU files

[andrsd]

I did not try yet on my machine and there is no tests as far as i can see that would exercise this API, so probably need to build one...

[ragusa]

I did not try yet on my machine and there is no tests as far as i can see that would exercise this API, so probably need to build one...

a test can be made by picking a Keigen problem that runs fast and adding what is needed.

[ragusa]

GetScalarFieldFunctionList only returns field functions with flux moments. If you need power density, call phys.GetPowerFieldFunction and export that into VTU files

power is a scalar field. so that scalar field terminology is not conducive to a clear understanding. @wdhawkins should GetScalarFieldFunctionList be renamed so that it is clear it only recovers the flux moments?

[wdhawkins]

Just a quick FYI, I don't think that will work since, iirc, we only have ExportMultipleToVTU and it expects a list which GetPowerFieldFunction doesn't supply.

[andrsd]

I tried my "theory" and:

1. apply this patch on keigenvalue_transport_1d_1g.py:

diff --git a/test/python/modules/linear_boltzmann_solvers/transport_keigen/keigenvalue_transport_1d_1g.py b/test/python/modules/linear_boltzmann_solvers/transport_keigen/keigenvalue_transport_1d_1g.py
index 3e54ce9e7..d43ab5e96 100644
--- a/test/python/modules/linear_boltzmann_solvers/transport_keigen/keigenvalue_transport_1d_1g.py
+++ b/test/python/modules/linear_boltzmann_solvers/transport_keigen/keigenvalue_transport_1d_1g.py
@@ -17,6 +17,7 @@ if "opensn_console" not in globals():
     from pyopensn.mesh import OrthogonalMeshGenerator
     from pyopensn.xs import MultiGroupXS
     from pyopensn.aquad import GLProductQuadrature1DSlab
+    from pyopensn.fieldfunc import FieldFunctionGridBased
     from pyopensn.solver import DiscreteOrdinatesProblem, NonLinearKEigenSolver
 
 if __name__ == "__main__":
@@ -81,6 +82,8 @@ if __name__ == "__main__":
             "use_precursors": use_precursors,
             "verbose_inner_iterations": False,
             "verbose_outer_iterations": True,
+            "power_field_function_on": True,
+            "power_normalization": 1.
         }
     )
     k_solver = NonLinearKEigenSolver(
@@ -90,3 +93,7 @@ if __name__ == "__main__":
     )
     k_solver.Initialize()
     k_solver.Execute()
+
+    ff_pd = phys.GetPowerFieldFunction()
+    vtk_base_name = "test"
+    FieldFunctionGridBased.ExportMultipleToPVTU([ff_pd], vtk_base_name)

Note that we grab a single field function, but we pass python list into ExportMultipleToPVTU (this is expected behavior).

2. Run the input file, open test.pvtu, I see:

where power_generation is the expected field function name. IDK, if this result is correct, but the code produces something (hopefully correct).

[andrsd]

I think that if one wants everything in a single file, just add the power generation field function to the field function list passed into ExportMultipleToPVTU.

[andrsd]

a test can be made by picking a Keigen problem that runs fast and adding what is needed.

Here is a sketch of adding this to the test suite. However, the power_normalization parameter should be something that makes sense in that test, rather then 2.0 (which I just used to see that the normalization does something)