DC3: Robust Autonomous Control of Onboard Generation AWE System

Objectives

This DC’s aim is to find a robust, full autonomy-capable control scheme for a onboard generation AWE system. The general control scheme is a cascade of PID controllers (e.g. attitude and speed controllers) and a control allocation, which computes actuator demands like control surface deflections by essentially inverting the nonlinear parts of the control plant such that the by-PID-controllers-demanded moments and forces are actuated. A key research part and contribution of the DC is to find how robustness can be ensured or even guaranteed. Specifically, the DC will identify what redundancies of the sensors, control computers, communication buses, and actuators are required, and how to optimise these through fusion algorithms/software in terms of minimum cost, minimum total number of sensors, and minimum complexity. The DC will investigate off-nominal situations and corner cases, such as actuator failure during a gust (e.g., based on the IEC standard) and during a critical moment of the flight trajectory. Critical situations are identified, and the entire control software is validated for flightworthiness through Monte Carlo batch simulations. The DC will identify and design control algorithms specifically for such off-nominal situations, e.g., a stall recovery controller, such that the entire Software-Hardware-Control combination of the AWE system is likely (or even guaranteed, if possible) to always maintain control and stay in a safe state, or, respectively, does not stay in a safe state with a very low probability (e.g., 10 –9 per hour as specified for commercial airlines). The DC implements control algorithms in C++ and integrates into the company’s systems engineering framework, which includes the actual flight control software as well as software-in-the-loop simulation models. The DC will help to set up flight tests to validate their research, do data analysis, and confirm their algorithms’ effectiveness or propose/implement improvements. From this research, the DC will contribute generalized assertions (e.g., what sensors and redundancies are necessary and what are not), fault models, control algorithms (also real-time capable), all or at least in part applicable to other AWE systems or related technologies (e.g., drones, UAM vehicles, aircraft, wind turbines) and specifically applicable for a fly-gen system.

Expected Results

Reference design for an onboard generation hard-wing AWE. Statistics on various faults with corresponding statistics on catastrophic failure. Robust fault-tolerant control strategies. Experimental verification of the advanced controller.

Supervisory team

Florian Bauer is main supervisor, Lorenzo Fagiano is academic supervisor. Maximilian Isensee is a second company supervisor, while Mark Kelly, Christoph Hackl and Alessandro Croce are co-supervisors.

Planned secondments

Politecnico di Milano M21-M27 to learn about control and stability of onboard generation concept, supervised by Lorenzo Fagiano and Alessandro Croce.