Task 219: RK4 Precision Validation for Gravity Capture

Status: DONE

Human directive: Phase 21, item 24

Problem

The project uses fixed-dt RK4 for various orbital transfer calculations. For ballistic trajectories near gravity wells (gravity capture, hyperbolic flyby, periapsis passage), the gravitational acceleration changes rapidly, and fixed-dt may not capture these dynamics accurately.

The adaptive RK45 (Dormand-Prince) and symplectic Störmer-Verlet already exist in propagation.rs, but systematic validation for gravity-capture scenarios is missing.

Plan

Phase 1: Analytical test cases (TDD — write tests first)

1. Two-body hyperbolic flyby: Given V∞ and periapsis distance, compute turning angle analytically. Compare RK4/RK45 at various dt.
2. Gravity capture: Apply known ΔV at periapsis of hyperbolic orbit → compute resulting bound orbit analytically. Compare integrator results.
3. Kepler equation validation: Propagate known elliptical orbit, compare position at arbitrary true anomaly using Kepler equation.
4. SOI transition: Test accuracy at sphere-of-influence boundaries where gravity field switches.

Phase 2: Identify failure modes

1. Run RK4 at dt=60s, 600s, 3600s for gravity-capture scenarios
2. Measure position/velocity error vs analytical solution
3. Document dt thresholds where RK4 accuracy degrades
4. Compare with RK45 adaptive results

Phase 3: Fix affected analyses

1. If specific episode analyses use inadequate dt, switch to RK45 or reduce dt
2. Update cross-validation if needed
3. Document findings in analysis reports

Files

• crates/solar-line-core/src/propagation.rs — integrators
• crates/solar-line-core/tests/oracle_propagation.rs — oracle tests
• cross_validation/validate.py — scipy cross-validation