Research On FMS Fault Diagnosis Based On Support Vector Machine

With the continuous development of the society and the continuous improvement of the consumption levels,people's demand for high-quality and diversified products is also increasing day by day.Nowadays,some traditional single production are not in line with the development concept of today's manufacturing industry.As one of the important development branches in the future intelligent manufacturing field,flexible manufacturing system(FMS)can rely on limited Shared resources to achieve the production of diversified components and parts,representing the advanced manufacturing technology in the 21 st century.However,in the actual production process,due to the complexity of the flexible manufacturing system structure,the ambiguity and diversity of fault information,it is difficult to reflect the advanced manufacturing technology.Therefore,it is imperative to study the fault diagnosis of FMS.Based on the analysis of FMS fault characteristics and system constitution,this paper proposes a fault diagnosis method based on support vector machine by combining tatistical learning theory and support vector machine(SVM)theory.The main research contents of this paper are as follows:(1)According to the system structure of FMS,a framework of FMS fault diagnosis system is proposed.By analyzing the fault diagnosis of FMS,the fault diagnosis of logistics system and processing system is determined as the focus of this paper.(2)According to the characteristics of logistics system,a fault diagnosis strategy based on genetic algorithm(GA)to optimize SVM parameters is proposed.The SVM diagnostic model and GA-SVM diagnostic model are established by combining the relevant theories of support vector machines,and the classification performance of GA-SVM diagnosis model is better than that of SVM diagnosis model through simulation experiments,which greatly improves the accuracy of logistics system fault diagnosis.(3)According to the characteristics of machining system,a fault diagnosis strategy based on improved particle swarm optimization(PSO)to optimize SVM parameters is proposed.In order to avoid the traditional particle swarm optimization trapped in local optimal solution,an adaptive particle swarm optimization with compression factor and asynchronous learningfactor is proposed.The improved particle swarm optimization is used to establish PSO-SVM diagnostic model and the simulation experiment results show that PSO-SVM diagnosis model has excellent classification performance,which verifies the feasibility and effectiveness of the improved particle swarm optimization for SVM parameter optimization.(4)Combined with the diagnosis model established in this paper,the FMS fault diagnosis system is designed and developed by using Matlab and C# mixed programming technology.By showing the system's fault diagnosis function,the reliability of the support vector machine algorithm for FMS fault diagnosis and the practicability of the FMS fault diagnosis system are verified.