小波包能量熵和改进的LSSVM在风力机轴承故障诊断中的应用

Application of Wavelet Packet Energy Entropy and Improved LSSVM in Fault Diagnosis of Wind Turbine Bearings

针对极端复杂工况下风力机轴承运行状态监测中的故障诊断问题,提出一种基于小波包能量熵故障特征提取并结合鲸鱼算法(WOA)优化最小二乘支持向量机(LSSVM)进行故障分类识别的风力机轴承故障诊断方法。通过小波包分解提取各频带成分的能量熵值构建故障特征集,同时针对LSSVM参数的选取依赖人工选择的盲目性问题,采用鲸鱼优化算法寻找LSSVM中最优的2个关键参数正则化参数和核函数参数,以此提高故障诊断模型的分类精度。通过不同工况下的试验数据集测试,实现了对不同故障状态特征参数的准确分类。结果表明,所提方法诊断结果优于遗传算法(GA)和粒子群算法(PSO)分别优化的LSSVM.远优于传统的LSSVM算法。

Aiming at the problem of fault diagnosis in wind turbine bearings running condition monitoring under extremely complex working conditions,a fault diagnosis method of wind turbine bearings was proposed based on wavelet packet energy entropy and the least square support vector machine(LSSVM) optimized by whale optimization algorithm(WOA).The energy entropy of the components of each frequency band was extracted by wavelet packet decomposition to construct the fault feature set.Aiming at the blindness problem that the selection of parameters in LSSVM depends on manual selection,WOA was used to find the optimal key parameters in LSSVM,regularization parameters and kernel function parameters,so as to improve the classification accuracy of the fault diagnosis model.The accurate classification of the characteristic parameters of different fault states was realized by testing the test data sets under different working conditions.The results show that the proposed method is superior to the LSSVM optimized by genetic algorithm(GA)and particle swarm optimization algorithm(PSO),and far superior to the traditional LSSVM algorithm.