Flight Simulation Model Of A Multi-Fidelity Digital Twin Of An Evtol Drone

Abstract

In the last decade, novel advancements in electric motors have revived the possibility of developing new VTOL airplanes. While their potential is clear, significant engineering challenges remain. The goal of this study was to test different advanced modeling and simulation techniques, including both physics-based and data-driven approaches. This work is part of the MODEL-SI project, where a previous contribution reviewed applicable methods for the development of an eVTOL Digital Twin (DT). A modular approach was chosen to build a comprehensive Flight Simulation Model (FSM). Conventional physics-based methodologies were primarily used, while some modules were implemented using a state-ofthe- art, i.e. Co-Kriging, and a novel Machine Learning (ML) method: the Bayesian Neural Network with Transfer Learning (BNN-TL). An initial assessment between the two methods demonstrated the superior capabilities of BNN-TL in handling complex problems. Ultimately, the final FSM is capable of performing static and dynamic analyses, such as aircraft trim, transition simulation, and aeroelastic analyses.