Rotorcraft pitot-static systems calibration process to reduce error in all flight regimes and all rotorcraft configurations

Abstract

Calibration of pitot-static systems is a priority activity to be conducted at the earliest stage of the rotorcraft certification process in order to provide accurate airspeed and altitude flight test data and to develop production pitot-static air data systems with acceptable errors. Legacy air data system did not allow any correction for position error causing great effort for the system design. State of the art Air Data Units allow correction of the airspeed and altitude errors but only for one flight regime and one configuration. Those corrections might present residual errors that are not acceptable for all the other flight regimes and rotorcraft configurations requiring re-design and re-testing. The paper explains the pitot-static systems calibration method used to discern static and total pressure errors in level flight by field test with the use of a Weather Station. Then it describes the results obtained using the GPS quasi-static Head and Tail wind method. This allowed to obtain calibration data for all flight conditions and for all helicopter configurations using a limited number of flights and without changing the helicopter configuration. The paper explains the equivalence of the level flight results between the NTPS GPS method and the GPS quasi-static Head and Tail wind method and suggests a correction for the Orbis GPS quasi-static technique. A mathematical method is shown to infer calibration curves for pitot systems which guarantee an optimal performance in all the flight conditions and configurations, leading to the creation of lookup tables for the state of art Air Data Units. Finally, the paper shows a new methodology developed based on a multi-input regressor using a fully connected Neural Network which is capable of providing a correct value of airspeed and pressure altitude in real-time for every aircraft configuration and every flight regime. In this way, the aircraft manufacturer will be able to optimize development and production costs, while having a fully calibrated airspeed in flight. The study provides detail on how to collect flight test data points, pre-processing techniques and analysis. Moreover, the paper describes the findings in optimizing the algorithm structure and in implementing such algorithm in an embedded system.