A simplified digital twin framework of a trainer aircraft

Reinhardt, Dominik; Bouferrouk, Abdesselam

Abstract

This paper aims to establish a simplified, user-friendly framework for a Digital Twin (DT) of a PC21 jet trainer aircraft. The implementation of the DT is carried out in Matlab with the aid of embedded Neural Network tools. The aircraft is trained and tested with flight data provided by QinetiQ/ETPS, as well as simulation data obtained via UDP connection from X-Plane, via Simulink. The paper investigates the potential of DTs for enhancing the design, testing, and certification process of a trainer aircraft, with the goal of minimising human error, costs, effort, and environmental impact. The testing of the DT framework with manoeuvres from real flight data shows the feasibility of using a Neural Network to simulate the behaviour of a trainer aircraft. Despite the limitations faced during the execution of the project, the predictions of the network, for a similar manoeuvre as the one it was trained with, show promising results. This proves that the DT framework, with further development and careful adjusting, can be utilised as a simple tool for simulating the flight dynamics of an aircraft.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: Aerospace Europe Conference 2023
• Start Date: Jul 9, 2023
• End Date: Jul 14, 2023
• Acceptance Date: Jul 8, 2023
• Online Publication Date: Oct 2, 2023
• Publication Date: Oct 2, 2023
• Deposit Date: Jul 18, 2023
• DOI: https://doi.org/10.13009/EUCASS2023-965
• Public URL: https://uwe-repository.worktribe.com/output/10956180