Fault diagnosis of induction motor based on ICEEMDAN and POA-SVM
It is difficult to extract the stator current fault features of induction motor,and the selection of Support Vector Machine(SVM)penalty coefficient c and kernel function parameter g has great influence on the diagnosis results.An induction motor fault diagnosis method based on Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(ICEEMDAN)and Support Vector Machine(POA-SVM)optimized by Pelican Optimization Algorithm(POA)was proposed.Firstly,ICEEMDAN was used to decompose the stator current filtered by notch filter to obtain a series of Intrinsic Mode Function(IMF).Then,the first 7 order IMF components of each state signal were selected and the energy entropy was calculated as the fault feature vector.Finally,the fault feature vector was input into the POA-SVM model to obtain the diagnosis result.Through the simulation software Ansoft/Maxwell,the motor model was established to obtain the current data,the diagnosis accuracy reaches 100％,and the fault diagnosis of induction motor was realized.In order to further verify the superiority of the diagnosis method,a motor fault simulation test bed was built to collect current signals.The results show that the diagnosis accuracy of the proposed method can reach more than 97.5％under three load conditions:no-load,half-load and full-load.Compared with other fault diagnosis meth-ods,the proposed method has better recognition ability for induction motor electrical faults.
国家科技期刊平台
NSTL
万方数据
