Generation of Parametric Climb Trajectories considering Operational Inputs for Aircraft Engine Thrust Extraction

Abstract

For aircraft engines costs related to maintenance, repair \& overhaul make up a great proportion of the overall direct operating costs. As the aviation sector is about to face substantial technological shifts towards hybrid-electric and all-electric propulsion, tools are required to model engine operating costs and their strong interdependence to operational factors. This study presents a method for adapting the parametric climb trajectory generation of the aircraft performance model OpenAP for considering operational inputs of flight distance and ambient conditions. Flight data of Airbus A320 operated in North America are analysed for the characteristic climb parameters. The data is used to train a XGBoost machine learning model in order to link the operational inputs to the trajectory parameters. The results show, that the model is able to represent global trends in the data while staying within the limits of the original model. However, the model shows some singularities, which could be addressed by parameter tuning and expanding the data base. Eventually, the generated trajectories differ from the default trajectory of the original model, such that the average thrust per segment varies in the range of ±20% to ±10%.