Validate ellipsoid nonlinear heave with radiation convolution

[cmccomb]

What changed

• Add the pinned MATLAB Nonlinear_Hydro/ode4/RegularCIC baseline and extend the existing heave mesh mode to convolution radiation.
• Allow the radiation-memory trapezoidal step more fixed-point iterations near the moving waterline; no physical force coefficient is changed.
• Gate mesh buoyancy, Froude–Krylov excitation, drag, radiation convolution, body motion, and PTO output against the pinned MATLAB trajectory.

Paired evidence

On 3,001 saved MATLAB states, radiation convolution agrees within 3e-10 N, nonlinear buoyancy within 6e-9 N, drag within 7e-11 N, and heave excitation within 4e-7 N. The independent 150 s Python trajectory differs by at most 6.39 mm heave, 7.41 mm/s velocity, 8.89 kN PTO force, and 8.98 kW absorbed power. The previous constant-radiation ellipsoid case still passes its gates.

Verification

• Pinned R2025b RegularCIC baseline completed successfully.
• Local paired tests: 2 passed for Regular and 2 passed for RegularCIC.
• Local suite: 90 passed, 93 artifact-dependent tests skipped.
• actionlint and git diff --check passed.

SungJun Won's original commits remain in branch ancestry.