Update basicsim.py

devel/openmc/basicsim.py

@@ -19,26 +19,131 @@
 ##                                                                            ##
 ################################################################################
 
-def main():
+import numpy as np
+import openmc
+
+TRUE_VALUES = np.array([5.60481868e+01, 5.75019668e-04, 7.67958251e-01])
+TRUE_ERRORS = np.array([2.58956360e-03, 4.83597510e-06, 1.96171839e-03])
+TRUE_POSITIONS = np.array([0.00872665, 1.15202318, 2.29531971])
+
+def is_interactive():
+    import sys
+    return ( '--interactive' in sys.argv[1:]
+             or '-i' in sys.argv[1:] )
+
+def pencil_beam_model(cfg, E0, N):
+    """Return an openmc.Model() object for a monoenergetic pencil
+     beam hitting a 1 mm sphere filled with the material defined by
+     the cfg string, and compute angular distribution for the 
+     Debye-Scherrer cones"""
+
+    import NCrystal as NC
+
+    # Material definition
+
+    m1 = openmc.Material.from_ncrystal(cfg)
+    materials = openmc.Materials([m1])
+
+    # Geometry definition
+
+    sample_sphere = openmc.Sphere(r=0.1)
+    outer_sphere = openmc.Sphere(r=100, boundary_type="vacuum")
+    cell1 = openmc.Cell(region=-sample_sphere, fill=m1)
+    cell2_region = +sample_sphere & -outer_sphere
+    cell2 = openmc.Cell(region=cell2_region, fill=None)
+    geometry = openmc.Geometry([cell1, cell2])
 
-    import openmc
+    # Source definition
 
-    mat_al = openmc.Material.from_ncrystal('Al_sg225.ncmat')
-    materials = openmc.Materials([mat_al])
+    source = openmc.IndependentSource()
+    source.space = openmc.stats.Point((0, 0, -20))
+    source.angle = openmc.stats.Monodirectional(reference_uvw=(0, 0, 1))
+    source.energy = openmc.stats.Discrete([E0], [1.0])
 
-    sphere0 = openmc.Sphere(r=10, boundary_type='vacuum')
-    cell0 = openmc.Cell(region=-sphere0, fill=mat_al)
-    geometry = openmc.Geometry([cell0])
+    # Execution settings
 
     settings = openmc.Settings()
-    settings.source = openmc.IndependentSource()
-    settings.run_mode = 'fixed source'
+    settings.source = source
+    settings.run_mode = "fixed source"
     settings.batches = 10
-    settings.particles = 1000
+    settings.particles = N
 
-    # Run OpenMC model
-    model = openmc.Model(geometry, materials, settings)
-    model.run()
+    # Tally definition
+
+    wl = NC.ekin2wl(E0)
+    angular_binwdidth = np.radians(1.0)
+    debye_cones = np.array([ 2.0*np.asin(wl/(2.0*hkl.d)) for hkl in NC.createInfo(cfg).hklObjects() if wl < 2.0*hkl.d ])
+    angular_bins = np.sort(np.concatenate(([0,angular_binwdidth], debye_cones-0.5*angular_binwdidth, debye_cones+0.5*angular_binwdidth)))
+    print(angular_bins)
+
+    tally1 = openmc.Tally(name="debye-scherrer cones")
+    tally1.scores = ["current"]
+    filter1 = openmc.SurfaceFilter(sample_sphere)
+    filter2 = openmc.PolarFilter(angular_bins)
+    filter3 = openmc.CellFromFilter(cell1)
+    tally1.filters = [filter1, filter2, filter3]
+
+    tally2 = openmc.Tally(name="angular distribution")
+    tally2.scores = ["current"]
+    filter2b = openmc.PolarFilter(np.linspace(0, np.pi, 180+1))
+    tally2.filters = [filter1, filter2b, filter3]
+    tallies = openmc.Tallies([tally1, tally2])
+
+    return openmc.Model(geometry, materials, settings, tallies)
+
+def check_results(sp_file):
+    with openmc.StatePoint(sp_file) as sp:
+        tal = sp.get_tally(name='debye-scherrer cones')
+        df = tal.get_pandas_dataframe()
+    bin_high = df['polar high [rad]'].values
+    bin_low = df['polar low [rad]'].values
+    values = df['mean'].values/(bin_high - bin_low)
+    errors = df['std. dev.'].values/(bin_high - bin_low)
+    assert np.shape(TRUE_VALUES) == np.shape(values), 'Wrong number of Debye-Scherrer cones'
+    return np.all(values > TRUE_VALUES-TRUE_ERRORS) and np.all(values < TRUE_VALUES+TRUE_ERRORS) 
+
+def plot_tally(sp_file):
+    with openmc.StatePoint(sp_file) as sp:
+        tal = sp.get_tally(name='debye-scherrer cones')
+        df = tal.get_pandas_dataframe()
+    bin_high = df['polar high [rad]'].values
+    bin_low = df['polar low [rad]'].values
+    cone_positions = 0.5*(bin_high + bin_low)
+    cone_values = df['mean'].values/(bin_high - bin_low)
+    cone_errors = df['std. dev.'].values/(bin_high - bin_low)
+
+    with openmc.StatePoint(sp_file) as sp:
+        tal = sp.get_tally(name='angular distribution')
+        df = tal.get_pandas_dataframe()
+    bin_high = df['polar high [rad]'].values
+    bin_low = df['polar low [rad]'].values
+    angdist_bins = bin_high
+    angdist_values = df['mean'].values/(bin_high - bin_low)
+    angdist_errors = df['std. dev.'].values/(bin_high - bin_low)
+
+    import matplotlib.pyplot as plt
+    plt.figure()
+    plt.step(np.rad2deg(bin_high), angdist_values)
+    plt.errorbar(np.rad2deg(cone_positions), cone_values, yerr=cone_errors, fmt='.')
+    plt.errorbar(np.rad2deg(TRUE_POSITIONS), TRUE_VALUES, yerr=TRUE_ERRORS, fmt='.', label='True values')
+    plt.yscale('log')
+    plt.ylabel('density')
+    plt.xlabel('Polar angle [deg]')
+    plt.legend()
+    plt.show()
+
+def main():
+
+    n_particles = 1000000
+    E0 = 0.0045  # eV
+    cfg = 'Al_sg225.ncmat'
+    model = pencil_beam_model(cfg, E0, n_particles)
+    sp_file = model.run()
+
+    if is_interactive():
+        plot_tally(sp_file)
+
+    assert check_results(sp_file), 'Statistically significant deviation from true values'
 
 if __name__ == '__main__':
     main()