Add Hertz contact rebound benchmark

[SueHeir]

Summary

• Adds examples/bench_hertz_rebound/ — a quantitative benchmark that validates Hertzian contact mechanics by dropping a single sphere onto a rigid flat wall
• Sweeps 4 impact velocities (0.1, 0.5, 1.0, 2.0 m/s) × 4 COR values (0.5, 0.7, 0.9, 0.95) = 16 test cases
• Validates coefficient of restitution, contact duration (vs Hertz theory t_c = 2.87(m²/RE*²v₀)^{1/5}), and peak overlap

Test plan

• cargo build --no-default-features --example bench_hertz_rebound passes
• cargo test --no-default-features passes (exit 0)
• Single-case run produces correct rebound results
• Full parameter sweep (16 cases) completes successfully
• validate.py passes: 49/49 checks pass
• COR ≤ 3% error for COR ≥ 0.7; ≤ 12% for COR < 0.7 (known Hertz nonlinearity)
• Contact duration ≤ 10% error vs analytical Hertz prediction
• Peak overlap within tolerance of elastic Hertz prediction

🤖 Generated with Claude Code