An Unknown Input Observer-Based Fault Diagnosis Method of eVTOL Under Closed-Loop Control

With the development of the low-altitude economic and industrial ecosystem, the safety of electric vertical take-off and landing (eVTOL) has attracted attention in recent years. To increase the fault diagnosis accuracy of rotors without direct speed sensors under the close-loop control, an unknown input observer-based fault diagnosis method is proposed in this paper. Firstly, an unknown-input observer is constructed from a linear state system with selected flighting state variables to track the control signals and rotor speeds at the balanced control point. Then, the residual between the ideal rotor speeds at the balanced control point and equivalent rotor speeds computed from sensorial control signals is calculated. Finally, the normalized residual is used for fault detection, isolation and level identification. The experimental results demonstrate the effectiveness of the proposed method under normal conditions, single-fault conditions and multiple-fault conditions.