AUV POTENTIAL-FLOW STUDY / 2026-10-06

Steady 3-D source-panel calculation at 0.30 m/s, zero incidence, unbounded
incompressible water. Inviscid and irrotational. Motors off. This is a preliminary
external-envelope calculation, not a viscous wake/propeller simulation or a
measurement. No drag or stability claim follows from this model.

Reproduce with Python 3.12 and the versions in requirements.txt:
  python geometry.py          # voxelize original CAD into calculation meshes
  python panel_solver.py      # independent analytical-sphere benchmark
  python study.py             # three panel meshes + 1,000 fixed fluid probes
  python study.py --integrate # solved throughflow streamlines
  python verify_paths.py      # streamline step-halving / segment clearance

The input CAD is in metres: X nose to tail, Y up, Z lateral. Input CAD was used
only to generate a 2 mm closed voxel envelope. Cavities and openings are
simplified/capped, and small ties and fasteners suppressed. The full CAD is not
the calculation boundary. Do not use the envelope meshes to manufacture parts.

results.json has conditions, numerical checks, exact fixed probes and limitations.
surface-results.csv has one row per panel at its centroid. Pressure change is
Cp * 44.91 Pa, relative to freestream pressure at the same height. The HTML
viewer clips only the colours (speed 0..0.60 m/s, Cp -2..1), not these data.
Sharp-corner surface peaks vary with resolution: no peak-pressure claim.
probes-comparison.csv contains Cartesian velocities for the same 1,000 fluid
probes on all three meshes; relative changes are normalized by 0.30 m/s.
streamlines.npz holds positions (m), speed (m/s), travel time (s), and seed IDs.
Only paths staying outside the mesh with positive streamwise velocity are
displayed. See streamline-checks.json for filtering and integration details.
solution-*.npz contains meshes, source strengths, surface and probe velocities.

Sphere verification uses 915 off-surface probes at 1.05R, 1.25R and 2R.
Numerical verification/refinement is not experimental validation.

References:
https://web.mit.edu/fluids-modules/www/potential_flows/LecturesHTML/lec09/lecture9.html
https://ntrs.nasa.gov/citations/19910009745
