Studies On Fault Diagnosis Of Aircraft Fuel System Based On Particle Swarm RBF Network

As a typically complex system, aircraft may inevitably encounter various failures in the course of using because of human error, material defects, manufacturing errors, environmental fluctuations and other factors as fatigue, wear and aging. Therefore, timely and accurately fault diagnosis of aircraft systems and given troubleshooting strategies to improve maintenance efficiency are very important to ensure the safety of normal operations.For the complex characteristics of the aircraft failure type, this paper takes aircraft fuel system of Boeing 737 as the research object, analyzed the process of fault diagnosis, using particle swarm optimization on RBF neural network fault diagnosis of aircraft fuel systems in order to achieve rapid and accurate diagnosis. This method can reduce the time of aircraft troubleshooting and improve the economic efficiency of airlines. Specifically: taking the civil aircraft maintenance manual, maintenance program and the reliability of the report as fault diagnosis knowledge sources to build a knowledge composite model of civil aviation fault diagnosis, then analysis the aircraft fuel system fault to get neural network training sample.Design RBF network based on particle swarm optimization model for fault diagnosis, by this step, this paper uses the subtractive clustering algorithm to determine the number of hidden units of RBF networks, uses particle swarm algorithm to optimize RBF network, then calculate the center's width value with the least square method, in the last part, implement model simulation on Matlab to verify the fault diagnosis method based on particle swarm RBF network. The results show that with this method can effectively improve the efficiency of fault diagnosis and reduce the flight schedule due to flight delays, it has practical value.