| name | diffraction-analysis-unit-conversion-traps |
| description | Sub-skill of diffraction-analysis: Unit Conversion Traps (+3). |
| version | 1.0.0 |
| category | engineering |
| type | reference |
| scripts_exempt | true |
Unit Conversion Traps (+3)
Unit Conversion Traps
| Source | Mass Unit | Inertia Unit | Density | Convert to spec.yml |
|---|
WAMIT .frc | te | te·m² | 1 (te/m³) | ×1000 for kg, kg·m² |
| OrcaWave | te | te·m² | 1 (te/m³) | ×1000 for kg, kg·m² |
| AQWA | kg | kg·m² | 1025 (kg/m³) | Direct |
| spec.yml | kg | kg·m² | kg/m³ | — |
Trap: When porting from WAMIT .frc to spec.yml, multiply inertia by 1000 (te·m² → kg·m²). Do NOT multiply twice — the ISSC TLP case 2.5 had a 1000x error from double conversion.
Phase Correlation for Zero-Magnitude DOFs
Fixed DOFs (e.g. heave/roll/pitch on TLPs) produce zero RAO magnitudes. Phase is undefined (atan2(0,0) noise), so:
multi_solver_comparator.py: overrides phase correlation to 1.0 when peak_mag < 1e-10validate_owd_vs_spec.py: skips DOFs with max_diff < 1e-6 in pass/fail verdict
Benchmark Report Geometry Schematics
The build_mesh_schematic_html() in benchmark_plotter.py renders interactive 3D Plotly mesh visualizations. It requires mesh_path in solver_metadata. When using build_orcawave_metadata_from_yml(), propagate mesh_path from build_solver_metadata() base metadata — the enriched YAML metadata doesn't resolve mesh paths.
Mesh Quality Assessment
- Adjacent panel ratio (between neighbors) is the key metric — not global min/max area ratio
- Gradual size variation is acceptable even with large global ratio
- Report uses
NOTE: prefix (informational) not WARNING: for area ratio
- Advanced metrics (Jacobian, aspect ratio, skewness) available via gmsh meshing tool
