关键词: 白鲸优化算法, 变分模态分解, 支持向量机, 主轴承, 故障诊断

Abstract: This paper proposes a diagnostic framework based on the white whale optimization algorithm combined with variational mode decomposition ( VMD) and support vector machine ( SVM) to address the problem of weak early fault characteristics and susceptibility to strong noise in wind turbine main bearings. The algorithm aims to minimize the envelope entropy and adaptively search for the optimal number of VMD modes and penalty factors, Extract time-domain, entropy domain,and frequency-domain features from the selected eigenmodes and construct high discriminative feature vectors,The classification is ultimately completed by SVM. The experiment used long-term monitoring data from a 2 MW wind farm in Inner Mongolia ( 800 samples,including normal,inner ring fault,and outer ring fault conditions) to verify the method. The results showed that the average diagnostic accuracy of our method was 94. 2%, which was about 13% higher than the unoptimized VMD-SVM ( the same dataset 81. 3%) .

Key words: beluga whale optimization algorithm, variational mode decomposition, support vector machine, main bearing, fault diagnosis