Fault Simulation and Diagnosis of the Aero-Engine Fuel Regulator

Effective fault diagnosis is very important to improve the performance and safety of aero-engine fuel regulator. In this paper, a novel method based on PSO-BP neural network is presented to diagnose the faults of aero-engine fuel regulator. Firstly, a nonlinear model of the fuel regulator is established in AMESim software. Subsequently, on the basis of nonlinear model, six typical faults are simulated corresponding to their fault mechanisms, and the fault characteristic table is obtained after detailed analysis. Further, in fault diagnosis, the particle swarm optimization (PSO) algorithm is combined with BP neural network, where the weights and biases of neural network are optimized by PSO algorithm. Finally, the simulation results demonstrate that the presented method can realize the accurate diagnosis of faults, and it achieves faster convergence speed and higher diagnosis precision than classical BP neural network, providing a new way for the fault diagnosis of aero-engine fuel regulator.