[lililii124]

Adds HJC concrete to the solid mechanics workflow: pressure- and rate-dependent strength, a continuous three-region compaction EOS, irreversible damage, and a predictor/corrector step that restarts each stage from saved material histories. It reuses the existing Jaumann stress-rate and stress-momentum equations, including the out-of-plane normal stress for plane strain.

Includes a plane-strain block impact and a 3D Taylor impact example. The latter launches a 10 mm diameter, 20 mm long cylinder at 30 m/s against a frictionless rigid plane, with a 0.5 mm initial gap. Its default 0.5 mm spacing gives 12,640 particles, evolved to 60 microseconds. A local normal penalty equation supplies the wall contact. Both examples export particle fields and reproducible response plots.

Taylor impact: damage (top) and equivalent stress (bottom) at 20, 40 and 60 microseconds. The half-cylinder view exposes the interior; colours and positions are unfiltered.

Validation:

• Material tests compare the update with closed-form shear softening and analytical single-particle wall contact/rebound. They also cover compaction/unloading, EOS continuity, tensile cutoff, compiled 3D shear and finite rotation, and stage-history handling. Both impact cases are included in the standard example smoke tests.
• CPU/OpenMP Taylor runs complete at 1 mm and 0.5 mm spacing. At 0.5 mm, halving the time step gives peak-normalized RMS differences of 0.0253% in force, 0.0148% in mean velocity and 0.0133% in mean damage. The final particle damage difference is 0.00208 RMS and 0.0232 maximum. Density remains positive, damage stays bounded and nondecreasing, and momentum changes agree with the wall impulse.
• Refining spacing from 1 mm to 0.5 mm changes those curves by 5.59%, 4.95% and 2.98%. At fixed 1 mm spacing and time step, doubling wall stiffness changes the force peak from 34.66 to 47.57 kN. These sensitivities are documented; neither the force peak nor local damage is presented as a calibrated or spatially converged prediction.

The documentation includes the block-impact plots, reproduction commands, integration conventions and model references: Holmquist, Johnson and Cook (1993), Meyer (2011, ARL-TR-5556), and Taylor (1948) for the cylinder-impact method. Damage is local; fully damaged particles retain confined strength. Erosion, fracture-energy regularization and thermal effects are outside this change. GPU execution is not validated; Taylor history output requires a single process.