In order to solve the problems of modeling difficulties,low simulation accuracy and insufficient fault samples required for fault diagnosis in the traditional single-domain modeling method of axial piston pump,a fault diagnosis method based on multi-do-main coupling modeling was carried out for the axial piston pump.The multi-domain coupling model of the axial piston pump was con-structed by Simscape,and three common faults of plunger leakage,spindle bearing wear and combined fault were simulated,then the pressure and flow data under various working conditions and different fault degrees were obtained by fault injection technology and MATLAB fast restart function.The fault data were extracted from time domain and frequency domain,and the fault features were selected by one-way ANOVA.Finally,the obtained features were used to train five fault diagnosis algorithms,including K-nearest neighbor,na-ive Bayes,decision tree,neural network and support vector machine,and the algorithm with the highest accuracy of fault diagnosis was obtained,with an average diagnostic accuracy of 98.5％.The method improves the accuracy of multi-domain coupling modeling of axial piston pumps,and realizes effective fault diagnosis of axial piston pumps.
维普
万方数据