Hydraulic Pump Fault Diagnosis Based on VMD-CS-SVM Comprehensive Characteristic Index

Based on VMD algorithm, CS algorithm and SVM algorithm, the fault diagnosis model of VMD-CS-SVM hydraulic pump is established, and the fault diagnosis and classification of hydraulic pump are realized. Firstly, the fifth-order IMF component was obtained by decomposing the signals using VMD. The singular value, amplitude spectrum entropy and fuzzy entropy of each component were calculated respectively, which were mainly composed of comprehensive indexes. The comprehensive indexes were substituted into the support vector machine optimized by Sparrow algorithm for classification and recognition. The experimental results of hydraulic pump show that the recognition rate of this method can reach 100%, which improves the diagnostic accuracy and reliability of hydraulic pump.