Simulation and fault diagnosis for aviation engine starting system based on SIMULINK

According to the characteristics of aviation engine starting process, this paper mainly analyzes state parameters of the starting process, and establishes its mathematical model. On the basis of using the SIMULINK to simulate aviation engine starting system, some fault conditions are tested to obtain fault data, and the principal component analysis (PCA) is used for data dimension reduction and to obtain the characteristic data by MATLAB. Then the genetic algorithm is used to optimize the back propagation (BP) neural network training features, and finally with test data to verify the accuracy of the results. Diagnostic results indicates that the complexity of the neural network structure and the training time are both reduced, and the accuracy of diagnosis is improved by the genetic algorithm combined with BP algorithm (GA-BP).