基于改进模糊聚类的齿轮箱故障诊断算法  

Gearbox Fault Diagnosis Algorithm Based on Improved Fuzzy Clustering

作　　者：王印松[1] 刘艳艳 李牡丹 WANG Yin-song;LIU Yan-yan;LI Mu-dan(Department of Automation,North China Electric Power University,Baoding 071003,China)

机构地区：[1]华北电力大学自动化系,河北保定071003

出　　处：《控制工程》2023年第3期412-418,共7页Control Engineering of China

摘　　要：针对风电机组齿轮箱故障数据维度高、相关性强、信息特征交叉等特点,采用保局投影和模糊c-均值聚类相结合的方法进行齿轮箱故障诊断。通过保局投影进行故障特征提取,解决数据处理难、信息交叉等问题;加入数据权和抑制因子改进模糊c-均值聚类算法以实现特征分类,克服传统聚类算法分类“非此即彼”的缺点;通过诊断模型计算隶属度诊断数据属于某类的可能性,以避免强制分类带来的误判。最后,通过QPZZ-II旋转机械振动实验平台的齿轮箱故障数据对所提算法进行实验,结果表明该方法可以有效进行风机齿轮箱故障诊断,并对一定变工况下的诊断具有鲁棒性。Aiming at the characteristics of high dimensionality,strong correlation,and cross-information of different fault data of wind turbine gearboxes,this study uses a combination of locality preserving projections and improved fuzzy c-means clustering to diagnose gearbox faults.The locality preserving projections algorithm is used for fault feature extraction,which solves the problems of difficult data processing and information crossover.By adding data weights and suppression factors to improve the fuzzy c-means clustering algorithm for feature classification,it overcomes the shortcomings of traditional clustering algorithm classification of“either/or”;the diagnostic model calculates the possibility that the membership degree of the diagnostic data belongs to a certain category,and avoids the misjudgment caused by forced classification.Finally,the algorithm is tested with the gearbox fault data of the QPZZ-II rotating machinery vibration test platform.The results show that the proposed method can effectively diagnose and classify gearbox faults in wind turbine and is robust to diagnosis under certain variable conditions.

关 键 词：齿轮箱 故障诊断 保局投影 聚类 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]