Merge pull request #308 from GEOUNED-org/dev

Merge dev into main for new release

Author: AlvaroCubi

docs/example_of_use.rst

@@ -0,0 +1,107 @@
+Example of use
+==============
+
+The following is an example of how to use GEOUNED to convert a CAD file to a CSG file 
+for use in MCNP. It includes the methodology and rationale behind the process.
+
+.. warning:: It is assumed that GEOUNED has been properly installed as described in the 
+    `Quick install guide <./quick_install_guide.html>`_.
+
+1. `CAD simplification`_
+2. `Export as STEP`_
+3. `Conversion to CSG`_
+4. `Suspicious solids and debug`_
+5. `Void check verification`_
+
+.. _CAD simplification:
+
+CAD simplification
+~~~~~~~~~~~~~~~~~~
+
+The CAD model of the system should undergo a process of simplification. In this process,
+features that do not affect the radiation transport should be removed. The level of 
+detail is usually decreased. Surfaces that are not allowed like splines are substituted
+by simpler surfaces like planes or cylinders.
+
+.. _Export as STEP:
+
+Export as STEP
+~~~~~~~~~~~~~~
+
+Once the CAD is simplified, it is exported as a STEP file to be used as input by 
+GEOUNED.
+
+.. _Conversion to CSG:
+
+Conversion to CSG
+~~~~~~~~~~~~~~~~~
+
+The below code is run in a Python environment with GEOUNED installed. In the comments
+of the code there is an explanation of the parameters used.
+
+.. code-block:: python
+
+    import geouned
+
+    settings = geouned.Settings(
+        debug=True,  # This will create a folder with the CAD bodies and their decomposition
+        startCell=12001,  # The first cell id
+        startSurf=12001,  # The first surface id
+        compSolids=False,  # False so the different solids of a component 
+                           # are not combined into a single MCNP cell
+        voidGen=True,  # Generate the void cells 
+    )
+
+    geo = geouned.CadToCsg(settings=settings)
+    geo.load_step_file(filename=f"Geometry/model.stp")  # Path to the STEP file
+    geo.start()
+    print("Start exporting")
+    geo.export_csg(
+        geometryName="My model",  # Name of the model to appear in the MCNP file
+        outFormat=("mcnp",),  # The output format
+        UCARD=50,  # Makes the model a filler universe with filler id 50
+        volCARD=True,  # The volume card is included in the cell definition
+        volSDEF=True,  # Generate a SDEF card
+    )
+    print("Exporting finished")
+
+.. _Suspicious solids and debug:
+
+Suspicious solids and debug
+~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Running the above code will generate a debug folder that contains each of the converted 
+solids, both the original and the decomposed ones. Those bodies can be opened and 
+examined in a CAD software to compare them to the original model. If the decomposed 
+model doesn't look right, it is possible to manually decompose a specific solid and 
+redo the conversion.
+
+Also, it is possible that a folder called *Suspicious solids* is created. This folder 
+contains the same solids as the debug folder but only the ones that have been flagged as
+potentially problematic. This can be useful to quickly identify the problematic solids.
+
+After applying any fix necessary and exporting to STEP, the conversion can be run again.
+
+.. _Void check verification:
+
+Void check verification
+~~~~~~~~~~~~~~~~~~~~~~~
+
+Once the conversion is finished, it is recommended to stocastically check the volume
+of the solids and to make some plots.
+
+In this example, the *UCARD* was used to make the model a filler universe. For these 
+initial checks the *U=50* cards should be commented out as we want to test the 
+standalone model. Plots with the MCNP plotter should be done to check that the model 
+looks as expected. 
+
+Also, a void check simulation should be run. The *volSDEF=True* card automatically 
+generated a spherical inward-oriented source with the correct weight to calculate 
+volumes stochastically. The tallies for the solid MCNP cells were also automatically 
+generated. The user can run the simulation directly and the results of the tallies 
+should all be close to 1.0. If a MCNP cell tally shows a result far from 1.0 it would 
+mean that the volume of the solid is not correct.
+
+.. warning:: The void check simulation should be run with a high number of particles to 
+    achieve an acceptable statistical error. An usually number is 1e9 particles. Always
+    check that the statistical error of the tallies is below 0.1.

docs/index.rst

@@ -11,6 +11,7 @@ Supported codes include OpenMC, PHITS, Serpent and MCNP.
    :maxdepth: 2
 
    quick_install_guide
+   example_of_use
    users_guide/index
    python_api
    developers_guide

docs/version_switcher.json

@@ -23,5 +23,10 @@
         "name": "1.4.0",
         "version": "1.4.0",
         "url": "https://geouned-org.github.io/GEOUNED/1.4.0"
+    },
+    {
+        "name": "1.5.2",
+        "version": "1.5.2",
+        "url": "https://geouned-org.github.io/GEOUNED/1.5.2"
     }
 ]

src/geouned/GEOReverse/Modules/MCNPinput.py

@@ -389,6 +389,34 @@ def getTransMatrix(trsf, unit="", scale=10.0):
             0,
             1,
         )
+    elif len(trsf) == 9:
+        if unit == "*":
+            coeff = tuple(map(math.radians, trsf[3:9]))
+            coeff = tuple(map(math.cos, coeff))
+        else:
+            coeff = trsf[3:9]
+
+        axis = FreeCAD.Vector(coeff[0:3]).cross(FreeCAD.Vector(coeff[3:6]))
+        coeff = coeff + (axis.x, axis.y, axis.z)
+
+        trsfMat = FreeCAD.Matrix(
+            coeff[0],
+            coeff[3],
+            coeff[6],
+            trsf[0] * scale,
+            coeff[1],
+            coeff[4],
+            coeff[7],
+            trsf[1] * scale,
+            coeff[2],
+            coeff[5],
+            coeff[8],
+            trsf[2] * scale,
+            0,
+            0,
+            0,
+            1,
+        )
     else:
         if unit == "*":
             coeff = tuple(map(math.radians, trsf[3:12]))
@@ -965,6 +993,7 @@ def points_to_coeffs(scf):
     # coeff [0:3] a,b,c plane parameters
     # coeff [3]   d plane parameter
     # normalization is d set to one if origin is not in the plane
+    return coeff
 
 
 def get_parabola_parameters(eVal, eVect, T, U):
@@ -1250,6 +1279,7 @@ def gq2params(x):
 
     Dinv = np.where(abs(eVal) < 1e-8, eVal, 1 / eVal)  # get inverse eigen value where eigen< 1e-8
     zero = (abs(eVal) < 1e-8).nonzero()  # index in eigen value vector where eigen < 1e-8
+    zero = zero[0]  # nonzero return a tuple with array containing the nonzero indexes
     TD = -XD * Dinv  # Translation vector in diagonalized base
 
     k = np.matmul(TD, XD) + x[9]