Digital twin for component health- and stress-aware rotorcraft flight control

This paper pursues a probabilistic digital twin methodology for designing component health- and stress-aware system control, and demonstrates the proposed methodology for the problem of rotorcraft maneuver control. The probabilistic digital twin uses sensor data to infer up-to-date knowledge regarding the component’s current health state and the associated uncertainty, and predicts future degradation of the system and the stress experienced under specific operational trajectories. The stochastic optimization problem considers the stress predicted for the degrading system to decide suitable operational controls over time such that the system safely and reliably completes the desired task or mission. The operational optimization process using the probabilistic digital twin is demonstrated by conducting asset-specific simulation experiments for a light-weight rotorcraft and optimizing the pilot control actions.