基于形态学特征及支持向量机的水电机组轴系故障识别  

Identification of Shafting Faults in Hydro-electric Units based on Morphological Features and Support Vector Machine

作　　者：王勇劲 王永建 苟方映 廖涛 林志鹏 胡波 WANG Yongjin;WANG Yongjian;GOU Fangying;LIAO Tao;LIN Zhipeng;HU Bo(Huadian Electric Power Research Institute Co.,Ltd.,Hangzhou 310030,China;Huadian Sichuan Power Generation Co.,Ltd.,Chengdu 610000,China)

机构地区：[1]华电电力科学研究院有限公司,浙江杭州310030 [2]华电四川发电有限公司,四川成都610000

出　　处：《水电与新能源》2023年第10期6-10,共5页Hydropower and New Energy

摘　　要：为解决水电机组主轴运行期间的潜在故障,提出了多重分形谱、粒子群寻优和支持向量机对主轴图像数据多维度故障识别的方法;实现机组轴系运行中的摆度波形形态学特性提取,获得多重分形谱数据;采用粒子群寻优算法从海量样本中寻优获得异常故障特征形成故障特征集,由支持向量机分类方法对故障特征集进行分类识别,有效识别图形数据堆积过程中故障样本分类,实现轴系故障识别,以便确定相应的解决措施。In order to effectively identify the potential faults during the operation of the main shaft of hydro-electric units,a multi-dimensional fault identification method based on the graphical data of the main shaft is proposed,in which the multi-fractal spectrum,particle swarm optimization and support vector machine are combined.The multi-fractal spectral data is obtained by extracting the morphological features of the swinging waveform of the main shaft during the operation of the unit.The particle swarm optimization algorithm is used to get abnormal fault features from massive samples to form the fault feature set.Then,the support vector machine classification method is used to classify and identify the fault feature set,which can effectively identify the fault sample classification in the graphical data accumulation process.Thus,the shafting faults of hydro-electric units can be identified and corresponding treatment measures can be determined.

关 键 词：形态学 支持向量机 故障识别 

分 类 号：TM312[电气工程—电机]