Fault Diagnosis And Failure Analysis Of Reciprocating Compressor Valve

As the reciprocating compressor of general machinery,it has an extremely important position in the field of heavy industry.The fault of the valve in the reciprocating compressor is more than 60% of the total number of faults.Therefore,it is great practical significance and practical value to diagnose the fault of the reciprocating compressor valve.This paper takes the 4M80 multistage reciprocating compressor of Anyuan Chemical Co.,Ltd.as the research object.Through the field practice and investigation,the connection between the valve fault and the pressure signal change is analyzed according to the data and practice experience of the equipment.In this paper,the advantages of SVM in the field of mechanical fault diagnosis are summarized by studying the research learning and the relevant background of the support vector machine algorithm.Finally,the SVM was used by establishing the valve fault diagnosis model.The intake and exhaust pressure of the three compression stages of the 4M80-22/26-224-BX hydrogen compressor is taken as the input of the diagnostic model to determine the specific location of the fault.Finally,in the MATLAB software,the LIBSVM toolbox is used to realize the application of support vector machine algorithm in the valve fault diagnosis,and genetic algorithm(GA),cross validation(CV)and particle swarm optimization(PSO)are introduced to optimize the parameter of support vector machine.At the same time,the RBF kernel function and the optimal parameter C and G are selected to establish the fault diagnostic mode,and the diagnostic model is verified by the test set data.The simulation results show that the SVM model based on genetic algorithm optimizes the relevant parameters and has excellent recognition performance.