基于WPD-tSNE-SVM方法的电站机组主轴故障诊断分析  

Fault Diagnosis Analysis of Power Station Spindle Based on WPD-tSNE-SVM Model

作　　者：曹康栖 李灿[2] CAO Kangxi;LI Can(Binhai County Power Supply Branch,State Grid Jiangsu Electric Power Co.,Ltd.,Yancheng 224599,China;Department of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou 450001,China)

机构地区：[1]国网江苏省电力有限公司滨海县供电分公司,江苏盐城224599 [2]郑州大学机械与动力工程系,河南郑州450001

出　　处：《机械制造与自动化》2023年第6期226-228,共3页Machine Building & Automation

基　　金：河南省高等学校重点科研项目(21B535003)。

摘　　要：为提高电站机组主轴故障诊断效率,设计一种WPD-tSNE-SVM组合模型,采用小波包混合特征与支持向量机(SVM)对电站机组轴承开展故障诊断。研究结果表明:采用t分布式邻域嵌入方法降维数据呈现规律分布特征,说明小波包混合特征提取方法能够满足有效性。非线性SVM多故障分类器能够满足小波包混合特征的精确故障分析,各分类器都可以实现小波包混合特征集的高效分类,以径向基核函数设置的非线性SVM诊断方式达到了更高的准确率,从而为之后的维护保养过程提供参考价值,促进维护效率的进一步提升,有效保障电站机组主轴处于稳定运行状态。根据该方法诊断主轴轴承运行故障,为后续维护保养提供指导意义,获得更高的维护效率,确保电站机组主轴运行稳定性。In order to improve the fault diagnosis efficiency of power station unit spindle,a WPD-tSNE-SVM combined model was designed,and wavelet packet mixed feature and support vector machine were used to carry out fault diagnosis of power station unit bearing.The results show that the wavelet packet mixed feature extraction method can satisfy the effectiveness of the regular distribution of dimensionality reduction data by using tSNE method.The nonlinear SVM multi-fault classifier is in line with the precise fault analysis of the wavelet packet mixed features,each classifier can effectively classify the wavelet packet mixed feature set,and the radial basis kernel function is applied to set the nonlinear SVM diagnosis method to achieve higher accuracy,thus providing reference value for the subsequent maintenance process,promoting the further improvement of maintenance efficiency,and effectively guaranteeing the stable operation of the main shaft of the power station unit.With the method,the operation fault of spindle bearing is diagnosed,which provides guidance for subsequent maintenance,achieves higher maintenance efficiency,and ensures the operation stability of power station unit spindle.

关 键 词：电站机组 主轴 故障诊断 小波包分解 t分布式随机邻域嵌入 支持向量机 

分 类 号：TH17[机械工程—机械制造及自动化]