A Neural Network Based Approach for Reliability Analysis of UAV Flapping Wing

During the flight experiment of the flapping-wing UAV, the hinged position of the crank and connecting rod of the mechanism is worn very seriously, and there is significant nonuniform wear phenomenon, which quickly leads to the unbalance of the mechanism's movement until the mechanism parts are broken and the movement is terminated. In order to solve the reliability problem caused by the wear of the rotating pair in the complex mechanism, this paper constructs the functional functions of the flapping-wing drive mechanism, investigates the random distribution of the factors affecting the reliability of the mechanism, and establishes the training samples of the proxy model. Based on the BP neural network method, the agent model under 50 different cycles is obtained, and the wear depth distribution of the mechanism in each period is obtained successively, according to which the reliability evolution of the mechanism is analyzed. The research content of this paper is of great significance for further reliability analysis of the driving mechanism of the flapping-wing UAV and providing suggestions for improved design.