Monte Carlo: demo_hop
mission demo_hop · fidelity fast · 200 runs
12.5 %success
8.6–17.8 %95 % interval
81.5 %vehicle recovered
256.6 mCEP50
657.1 mCEP90
939.8 m99 % ellipse, major
Landing dispersion at the target
Local east/north offsets of every touchdown from the
target. Dashed circle: target radius (50 m). Solid ellipse: 99 %
bivariate-normal fit; dotted: 50 %. Mean offset -34.9 m E,
26.7 m N.
Distributions
Metrics (percentiles)
| metric | p1 | p5 | p50 | p95 | p99 |
|---|---|---|---|---|---|
| fuel_remaining_kg | 2.5 | 35.5 | 95.0 | 153.9 | 175.7 |
| max_cargo_g | 5.5 | 5.6 | 5.6 | 22.8 | 25.4 |
| touchdown_vz_mps | 0.2 | 0.4 | 0.8 | 1.2 | 1.3 |
| touchdown_vh_mps | 0.1 | 0.2 | 0.5 | 3.3 | 5.0 |
| flight_time_s | 556.7 | 560.8 | 572.4 | 584.8 | 587.2 |
| landing_error_m | 4.8 | 8.5 | 258.8 | 882.7 | 1,207.3 |
Outcomes
| outcome | runs | share |
|---|---|---|
| missed_target | 134 | 67.0 % |
| crash_hull | 26 | 13.0 % |
| landed | 25 | 12.5 % |
| tipped_over | 10 | 5.0 % |
| not_at_rest | 2 | 1.0 % |
| landed_off_site | 2 | 1.0 % |
| terrain_impact | 1 | 0.5 % |
Sensitivity
Spearman rank correlation of each dispersed parameter with miss distance and with failure. |ρ| above 0.14 (2/√n) is marked significant.
| parameter | ρ miss | ρ failure | |
|---|---|---|---|
world.wind.speed | +0.42 | +0.34 | significant |
vehicle.cargo.mass | +0.20 | +0.05 | significant |
world.temperature_offset | -0.17 | -0.19 | significant |
world.wind.gust_max | -0.11 | -0.16 | significant |
vehicle.aero.cd_scale | +0.14 | +0.09 | |
world.wind.from_deg | -0.13 | +0.01 | |
vehicle.engine.thrust_vac | +0.09 | +0.06 | |
vehicle.engine.misalignment_deg[1] | -0.06 | -0.08 | |
vehicle.engine.throttle_tau | -0.07 | -0.01 | |
vehicle.mass.dry_cg_z | +0.07 | +0.03 | |
vehicle.engine.misalignment_deg[0] | -0.06 | -0.06 | |
vehicle.engine.isp_vac | -0.05 | -0.05 | |
vehicle.mass.dry | +0.03 | -0.05 | |
vehicle.rcs.thrust | -0.04 | +0.01 | |
vehicle.grid_fins.cn_alpha | -0.03 | -0.00 |
Method
Each run samples the true vehicle and environment from the dispersion file; the flight software flies with the nominal vehicle, the forecast environment and the pre-flight ascent plan. Success probability uses the Wilson score interval. Ellipses assume a bivariate normal landing distribution. These are simulation results: model-form error is not included beyond the dispersed parameters (see docs/models).