Research On Fault Diagnosis Of Aviation Case Electrochemical Machining Based On BP Neural Network Optimized By Genetic Algorithm

Due to the fact that the casings of aviation engines are often ring-shaped thinwalled structural components,traditional cutting methods can cause machining deformation.Electrochemical machining,which has advantages such as no cutting force and good surface quality,can be used for the machining of casing parts.In practical applications,there are many casing parts that are processed using electrochemical machining.However,due to the many factors that can affect electrochemical machining,malfunctions can occur during the process,which can affect the efficiency and yield of the process.This paper mainly studies the mechanism and method of fault diagnosis in electrochemical machining of casing parts.Based on the BP neural network,a diagnostic model for electrochemical machining faults in casing parts is established,and genetic algorithms are introduced to improve diagnostic accuracy,achieving effective diagnosis of electrochemical machining faults in casing parts.Firstly,the types of faults in the machining process of the aircraft engine casing are analyzed,and the specific causes of each type of fault are analyzed to select the fault feature parameters required for fault diagnosis.Based on the learning method of the BP neural network,a basic model for diagnosing faults in the machining process of the aircraft engine casing is established.The experimental analysis of various training parameters on the diagnostic accuracy is carried out.The experimental results show that although the BP neural network has certain fault diagnosis capabilities,its diagnostic accuracy is not high.To further improve the diagnostic accuracy of the basic model,multiple optimization schemes are proposed for the shortcomings of the BP neural network diagnosis.The genetic algorithm is used for optimization,and the effects of the crossover and mutation probability on the optimization effect are analyzed.An adaptive crossover and mutation probability method is proposed to improve the optimization effect of the genetic algorithm.The adjustment rules of adaptive crossover and mutation probabilities are improved to further improve the optimization effect of the genetic algorithm.Simulation experiments are conducted to understand the impact of adaptive crossover and mutation probability on enhancing the optimization effect of the genetic algorithm.The experimental results show that after optimization by the adaptive genetic algorithm,the number of model iterations is reduced,the network convergence is accelerated,and the accuracy of the fault diagnosis model for the machining process of the aircraft engine casing can be improved to 98%.Finally,three sets of comparative experiments are designed to analyze the influence of fault characteristics,sample size and Confounding on the performance of the optimized ECM fault diagnosis model.The results show that the optimized electrochemical machining fault diagnosis model for the gearbox is less affected by interference,and has a high diagnostic rate in diagnostic tests of large and small samples.At the same time,it also shows better resolution in diagnostic experiments of similar fault types with similar features.The diagnostic results of three sets of experiments also indicate that the fault diagnosis model for gearbox electrochemical machining optimized by adaptive genetic algorithm has better generalization and diagnostic accuracy.Embed the trained diagnostic model into the gearbox electrochemical machining fault diagnosis system designed based on Matlab GUI,visualize the model training process,and greatly facilitate user training of diagnostic models and fault diagnosis of gearbox electrochemical machining.