3.4 KiB
Spherical Cap Central Point Load Benchmark
This folder contains a displacement-control FemLite benchmark derived from
models.sme.spherical_cap_with_central_point_load.pdf:
generate_spherical_cap.pyspherical_cap_displacement_control.inpspherical_cap_displacement_control_shell4.inpcheck_spherical_cap_results.pyrun_case.bat
The reference model is the NAFEMS spherical cap postbuckling case: radius
a = 10 m, thickness th = 0.203840 m, E = 210 GPa, nu = 0.3, cap angle
pi/4, rim support, and crown displacement control. The Shell-16 baseline keeps
the documented polar topology with a crown S3 fan and S4 meridional bands, uses
the historical full-rim constraint (RIM_DOF_MAX=6), and explicitly requests
SHELL_GEOMETRIC=LINEAR. The Shell-4 finite-strain target uses the same
geometry and material scale, but replaces the crown fan with an
all-quadrilateral square-to-disk S4 remesh so that it can enter the V5.11
SHELL_GEOMETRIC=FINITE_STRAIN production path. Its rim support intentionally
constrains translations only (RIM_DOF_MAX=3) to match the PDF shell
point/rim-support interpretation used by the diagnostic decks; full rotational
rim clamping is not the Shell-4 postbuckling benchmark target because it masks
the observed load drop. Both decks prescribe the crown displacement to
-6.2 m; the reaction is nondimensionalized as
Fn = -F*a/(E*th^3*2*pi), wn = -w/th
FemLite currently checks the displacement-control shell path that its
small-strain shell formulation can support. By default, the checker runs
static-linear Shell-16 samples at the PDF table displacement values and writes
spherical_cap_load_displacement.csv. If the environment variable
FEMLITE_SPHERICAL_CAP_NONLINEAR=1 is set, the checker instead attempts a real
--static-nonlinear Shell-4 direct-displacement path over the bounded V5.11
post-peak window and validates continuation_path.csv, the translational rim
support, path-control diagnostics and a reproducible load drop after the first
peak. The bounded nonlinear route defaults to uz=-3.24 m; it can be changed
with FEMLITE_SPHERICAL_CAP_BOUNDED_UZ.
If FEMLITE_SPHERICAL_CAP_RIKS=1 is set, the checker runs a separate
load-control Shell-4 Riks route. It converts the crown control to a proportional
crown CLOAD, writes spherical_cap_riks_load_path.csv, and validates that
the load parameter reaches a real limit-point neighborhood and decreases after
the peak while displacement continues to grow. This is a method-4 load-type
Riks evidence route for a real post-limit branch segment; broader
branch-following generalization is still deferred.
V5.11 uses this example as a finite-strain S4 capability proof, not as a claim
that the full uz=-6.2 m PDF curve has been strictly reproduced. Full
Shell-4/8/16 digitized-curve matching, later restiffening-window tuning and
late post-peak performance optimization are deferred to the next version.
Run:
run_case.bat
Results are written to results\auto_run.
Important output files:
continuation_path.csv: production accepted-increment direct displacement path, present only when the optional nonlinear diagnostic route completes.spherical_cap_load_displacement.csv: checker-derived nondimensionalwn = -w/thandFn = -F*a/(E*th^3*2*pi)curve.spherical_cap_riks_load_path.csv: optional load-control Riks path, present only whenFEMLITE_SPHERICAL_CAP_RIKS=1completes.