Time Frequency Diagnosis of Transformer Mechanical Fault Based on SSWT-GLCM and Improved WOA-SVM
In order to further improve the accuracy of transformer fault diagnosis,a time-frequency diagnosis method of transformer mechanical fault based on synchronous compressed wavelet transform and gray level co-occurrence matrix(SSWT-GLCM)and improved whale algorithm optimization and support vector machine(WOA-SVM)is proposed.Firstly,the time-frequency analysis of transformer vibration signal is carried out by using SSWT,and the two-dimensional time-frequency diagram with dense energy stack is obtained,which ef-fectively retains the characteristic information of transformer vibration signal.Then,the GLCM which jointly de-scribes the relationship between regional pixels extracts the main feature information of two-dimensional time-frequency map,providing effective feature parameters for subsequent fault diagnosis models.Finally,the key parameters of SVM are optimized by improved WOA,The time-frequency diagnosis model of typical mechani-cal faults of transformer based on improved WOA-SVM is established.The experimental results show that the improved WOA-SVM fault diagnosis model has high recognition accuracy and operation efficiency,and pro-vides technical support for transformer mechanical fault time-frequency diagnosis based on vibration signal.
NETL
NSTL
万方数据
