PROM 5.0 validation: the same model on a new engine
PROM v5.0 runs the PROM v4.3 model on a new native compute engine. The engine solves the equations, but the equations and coefficients have not changed. The validated behaviour in the earlier campaign has not changed either.
The full dataset was rerun on the new engine. It covers 75 propellers tested against wind-tunnel data and 98 motors tested against dynamometer data. The v5.0 results reproduce the v4.3 campaign to every published digit.
The old and new engines were also compared on a representative grid. The grid included single-rotor and multi-rotor configurations, several throttle settings, different battery states and both battery thermal models. Every reported result had to agree within one part in a million. All 24 configurations passed with both thermal models. The largest difference in any reported result was 4.3 parts in a hundred million.
A separate check regenerated the report from scratch with one engine build. All 61 published statistics matched the v4.3 campaign at the four significant figures used in the report. The native engine also handled every case without sending unsupported configurations to the previous engine.
What the measurements cover
The propeller results compare geometry-only predictions with the UIUC wind-tunnel measurements. The model takes only measured blade geometry as input and applies no fitted correction, so wind-tunnel measurements can never leak into a prediction.
The motor results also expose clear limits. Shaft speed is the weak channel, especially at wide-open throttle, where PROM tends to predict more speed than the motor can achieve. Of the 95 motors tested on the bench, 54 fall outside a ±5% speed band. Across the bench corpus, mean absolute speed error is 6.10%, and 49% of rows are within ±5%.
Below 30% throttle, motor current has roughly 21% mean absolute error, with only about a quarter of rows inside ±10%. The model performs best from 60% to 99% throttle. Motor efficiency error is roughly 5 percentage points, and the predicted efficiency stops rising near 84%.
The propeller model has its own boundary. It systematically under-predicts thrust at heavy loading and low advance ratio. The effect is strongest for short-pitch, large-diameter propellers near static thrust. Every wind-tunnel measurement in this report used airflow along the rotor axis. The report does not establish the same accuracy for angled inflow, stacked rotors or steep descent.
What to take away
PROM continues to return results outside the measured envelope without warning. Custom components are also outside that envelope, and the report's error bars do not apply to them. Outside-envelope results can still help compare similar designs because many errors move similar configurations in the same direction. They are much weaker as absolute predictions. Measure any absolute value outside the envelope that a design must rely on.
These results are engineering estimates, not certified data. Read the methods, tables and full limitations in the PROM 5.0 validation report.