Research strategy of fault diagnosis considering WOA-VMD algorithm

Abstract To address the problem of low accuracy fault information extracted from motor-bearing vibration signals, a feature extraction method based on Whale Optimization Algorithm (WOA)-Variable Modal Decomposition (VMD) energy entropy is proposed, and modified WOA-Support Vector Machine (SVM) used for diagnosis. Firstly, optimal number modes K penalty parameters α variable modal decomposition are found by WOA. WOA to find Then components obtained VMD processing signal, further as vector after screening. Finally, shortcomings population iteration mechanism, random variation strategy introduced improve WOA-SVM diagnosis bearing signal. The results show that WOA-VMD can decompose current signal overcome problems confusion pseudo-component EMD method, which an effective deal with rolling faults. Meanwhile, this has been verified experiments accurately extract identification rate data 99.2%.