SteveDoyle2
commented
2 years ago So maybe first off, that blog post is referencing a version that was never released. The blog release was supposed to be timed with 1.4, so when I found out it was released much later, I was too busy and not ready for a release. So...1.4 is the dev version and I highly recommend that for doing SMT. It's a lot more streamlined.
Am I correct in assuming the i vector is the vector along which to take section cuts?
No. i is the axis of the output data (e.g., torque is about the i axis). coord_out defines the march direction and we march in the i axis of that coordinate system. In the 1.4 code, that corresponds to iaxis_march, but doesn't exist in the 1.3 code (it's just coord_out.i).
At least conceptually, I think you're big point of confusion is the projection method. The default is method='Z-Axis Projection' and you're using method='CORD2R' options. I find "Z-Axis Projection" to be much easier, so that's what I use. The code below shows an example of each.
from pyNastran.bdf.bdf import BDF
from pyNastran.op2.tables.ogf_gridPointForces.smt import smt_setup, plot_smt
from pyNastran.bdf.mesh_utils.cut_model_by_plane import get_element_centroids, get_stations
import numpy as np
deck = BDF()
## Starting and Ending Points
# (Exact numbers made up for this thread)
p1 = np.array([500.0, 0.0, 9.0]) # Start at the nose
p3 = np.array([0.0, 0.0, 0.0]) # March towards origin
# defines the XZ plane for the shears/moments
p2 = p1 + np.array([1., 0., 1.]) # point C
# Either method is fine, but depends how you want to define z-axis
if 0:
zaxis = p1 + np.array([0., 0., 1.]) # point B
xyz1, xyz2, xyz3, i, k, coord_out, iaxis_march, x_stations_march = get_stations(
deck, p1, p2, p3, zaxis, nplanes=20,
method='CORD2R',
)
else:
zaxis = np.array([0., 0., 1.]) # point B
xyz1, xyz2, xyz3, i, k, coord_out, iaxis_march, x_stations_march = get_stations(
deck, p1, p2, p3, zaxis, nplanes=20,
method='Z-Axis Projection',
)
print(f'xyz1 = {xyz1}')
print(f'xyz2 = {xyz2}')
print(f'xyz3 = {xyz3}')
print(f'i = {i}')
print(f'{coord_out}')
print(f'iaxis_march = {iaxis_march}')
print(f'x_stations_march = {x_stations_march}; n={len(x_stations_march)}')
xyz1 = [500. 0. 9.]
xyz2 = [501. 0. 10.]
xyz3 = [0. 0. 0.]
i = [1. 0. 0.]
CORD2R 500. 0. 9. 500. 0. 10.
501. 0. 9.
iaxis_march = [-0.99983804 0. -0.01799708]
x_stations_march = [ 0. 26.32005229 52.64010457 78.96015686 105.28020915
131.60026143 157.92031372 184.240366 210.56041829 236.88047058
263.20052286 289.52057515 315.84062744 342.16067972 368.48073201
394.80078429 421.12083658 447.44088887 473.76094115 500.08099344]; n=20
Czarified
I've been trying to learn pyNastran methods for interface loads, and Load/Moment Diagrams. This has been a very frustrating process, since the documentation for these functions is sparse.
I found an m4 blog post walking through the process, but the function calls seem to be different than in my version (Python: 3.10.4, pyNastran: 1.3.4).
I'm expecting my
ivector to have a large and negative x-component, and a small and negative z-component, but instead I see the opposite:array([-0.01932164, 0. , 0.99981332])Am I correct in assuming the i vector is the vector along which to take section cuts? If so, I would expect the
stationsvariable to be an array of length 20, equally spaced from ~500 to 0.I apologize if I'm just missing something obvious here, but I've been scratching my head on this for several hours for a couple days, with no luck.