Abstract

Electric vertical takeoff and landing (eVTOL) aircraft transforms future transportation systems by alleviating transportation congestion on the ground. This eVTOL technique possesses unique features, including reduced noise, low pollutant emissions, efficient operating costs, and flexible maneuverability. Meanwhile, battery consumption poses critical challenges to flight task duration. Thus, optimal takeoff trajectory design is essential due to immense power demands during eVTOL takeoffs. Conventional design optimization, however, iteratively evaluates high fidelity simulation models, making the design process computationally intensive. In this work, we implement a machine learning-enabled inverse mapping optimization concept, i .e., directly predicting optimal design based on design requirements (including design constraints and flight conditions). This inverse mapping configuration requires a high-fidelity simulation-based design optimization for each training sample. Thus, more effectively generating training data becomes crucial to the success of inverse mapping. We adopt an adaptive sampling strategy for intelligent data acquisition to efficiently train machine learning surrogates. Specifically, the adaptive sampling strategy is based on proper orthogonal decomposition (POD) and infills samples for POD basis improvement. We demonstrated the proposed approach on optimal takeoff trajectory design of the Airbus A3 Vahana and compared its performance against Latin hypercube sampling (LHS). Results exhibited that the adaptive sampling-based deep neural network surrogates consistently outperformed LHS-based surrogate modeling, given the same computational budget and neural architecture. Specifically, the adaptive sampling-based surrogates achieved over 98% accuracy for predicting optimal trajectories of electrical power and wing angle, as well as total takeoff time, using around 380 training samples, whereas LHS-based surrogates cannot make this accuracy even with 780 samples.

Department(s)

Mechanical and Aerospace Engineering

Publication Status

Full Access

International Standard Book Number (ISBN)

978-162410723-8

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2025 American Insstitute of Aeronautics and Astronautics, All rights reserved.

Publication Date

01 Jan 2025

Share

 
COinS