Study on Fault Diagnosis of Planetary Gearbox in Unmanned Aerial Vehicle Using Multi sensor Data

Purpose: The harsh operating conditions of unmanned aerial vehicles cause crack propagation in the gear teeth in the actuator, which eventually leads to breakdown. For the prevention of service delays or failure because of crack propagation, a diagnostic process based on the built-in and add-on sensors’ signals is proposed for the planetary gearbox of unmanned aerial vehicles.BRMethods: A planetary gearbox test rig is constructed, and the motor current, the position and vibration signals are acquired for the normal and crack-induced states. Features representing the health state are extracted and selected by using a suitable performance measure. Subsequently, the K-Nearest Neighbor algorithm is applied to the selected features, from which the classifications of the normal and fault states are performed.BRResults: Among the acquired signals, the vibration signal obtained from add-on sensor showed better performance (100%, J₃ = 46.6) than the motor current (88.7%, J₃ = 3.89) and angular position (100%, J₃ = 5.88) acquired from the built-in sensor. In the features of the vibration signal, those in the frequency band with strong energy concentration showed outstanding separability.BRConclusion: The proposed method successfully classifies the fault from the normal in the planetary gearbox in the basis of the frequency-domain features of the vibration signal.