Add phi check for section properties validation

[jchkoch]

To avoid a RuntimeError that a argument of type "float | None" cannot be assigned to parameter "phi" of type "float" in function fea.principal_coordinate" (line 318 in section.py)

[robbievanleeuwen]

Happy to make this change, however curious as to how you got this error as phi should have been calculated earlier in the method calculate_centroidal_properties(). Do you have a file that can reproduce this runtime error?

[jchkoch]

I encountered this when retrieving stresses via the get_stress_at_points method on the following code snippet:

from sectionproperties.pre.library import primitive_sections
from sectionproperties.analysis.section import Section

geom = primitive_sections.rectangular_section(d=300, b=79)  # in mm

geom.create_mesh(mesh_sizes=[5])
sec = Section(geometry=geom)
sec.calculate_geometric_properties()
sec.calculate_warping_properties()
sec.display_results()

case1a = {'n': 2.860 * 1000, 'vx': 0.467, 'vy': -3.174, 'mzz': 1.362 * 1e6, 'mxx': -2.279 * 1e6, 'myy': -5.644 * 1e6, 'm11': 0, 'm22': 0}
case1b = {'n': -2.256 * 1000, 'vx': 0.076, 'vy': 7.002, 'mzz': -1.567 * 1e6, 'myy': 1.686 * 1e6, 'myy': 0.890 * 1e6, 'm11': 0, 'm22': 0}
case2 = {'n': -7.994 * 1000, 'vx': -1.431, 'vy': -0.187, 'mzz': 0.211 * 1e6, 'mxx': -10.747 * 1e6, 'myy': 2.018 * 1e6, 'm11': 0, 'm22': 0}

load_cases = [case1a, case1b, case2]

pt1 = (0., 0.); pt2 = (0., 300.); pt3 = (79., 300.); pt4 = (79., 0.)
points = [pt1, pt2, pt3, pt4]

for idx_lc, lc in enumerate(load_cases):
    print(f"Load case {lc}:")
    stress = sec.calculate_stress(**lc)

    results = sec.get_stress_at_points(pts=points, **lc)

[robbievanleeuwen]

Thanks @jchkoch, I'll have a look into this tomorrow, there shouldn't be a RuntimeError with this script, so maybe something else is going on.

[jchkoch]

yes, maybe there is something else going on here. My fix seemed to remove the error message I was getting from my script which was my immediate concern to get my stress results for my project I am working on.

It might be relevant to add that I needed also to add 'm11': 0, 'm22': 0 to my load cases in order for all of the error messages to disappear. Perhaps there is an issue with the input argument **lc in get_stress_at_points.