KoFEM v0.5.0

A guided first simulation, public validation examples, and corrections to loads
and shell/solid joints.

Added

• Try a sample simulation opens a wall bracket as CAD geometry and guides you
  through meshing, supports, loading and solving.
• Example and validation pages at /examples/, with models you can open in
  KoFEM, reference calculations, convergence studies and limitations. Older
  /validation/ links redirect to the corresponding example pages.
• Feedback offers problem reports and suggestions, with optional diagnostics
  that omit geometry, files and results.
• The engine starts loading when the app opens. The status bar shows progress,
  failures and Retry.
• Opt-in workflow analytics, with Cookie settings in the app when analytics
  are enabled.

Changed

• Saved analyses retain their linear or quadratic element order. Older analyses
  without an order still open as linear.
• The cantilever example opens ready to solve and compares its result with a
  hand calculation.
• Validation cases use the same loads as the app. Published results are now
  0.47358 mm axial extension, 2.7750 plate-with-hole stress concentration, and
  24.042 dimensionless Cook's membrane deflection. The benchmark models and
  loads were corrected; acceptance tolerances are unchanged.

Fixed

• Clicking inside a selected curved region deselects it.
• The status bar counts face and shell edge loads.
• Solid surface loads warn when some selected faces cannot be matched to the
  mesh. Malformed force and traction vectors are rejected with clear errors.

Breaking

• Shell/solid joints change when re-meshed. Shell meshing retains three wall
  thicknesses of solid material at planar joints and connects the shell through
  its physical footprint. Re-mesh and re-solve existing shell models to obtain
  these transitions. In tested tension and bending cases, comparisons with an
  all-solid model on the same mesh stayed within 0.36% for peak junction stress
  and 0.34% for the block stress field.
• Coupled and shell surface-load results can change. Total forces distribute
  by area or edge length, traction acts per unit area, and pressure is applied.
  One graded-face test changes its moment from 257.142857 to 240 N·mm; pressure
  previously omitted contributes the intended −120 N. Re-solve affected
  analyses, including coupled topology optimizations. Saved results do not
  update automatically.
• Conflicting displacement constraints stop the solve. A component cannot
  be fixed at zero and prescribed to a nonzero displacement, or prescribed to
  two different values. Remove the conflicting group or make its values agree.
  Identical duplicates remain valid.

Known limitations

• Only one body per model can be a Shell.
• The planar shell/solid transition checks do not establish junction-stress
  accuracy for every model. Generic distributing couplings retain their
  existing junction-stress limitations.
• Coarse linear tetrahedra can make thin-walled solid models too stiff. In a
  reported 3 mm bracket, displacement was 0.0275 mm with linear elements versus
  0.0560 mm with quadratic elements. Use quadratic elements or an appropriate
  Shell idealisation and check convergence.
• A reported meshing crash on complex STEP/IGES models remains unresolved and
  is deferred to 0.6.0.