基于WT和SVD的水电机组故障特征提取方法  

Fault Feature Extraction Method for Hydropower Units based on WT and SVD

作　　者：丁晨 刘梦 王官佳 杜伟[1] 吴凤娇[1] 王斌[1] DING Chen;LIU Meng;WANG Guanjia;DU Wei;WU Fengjiao;WANG Bin(School of Water Conservancy and Construction Engineering,Northwest Agricultural and Forestry University,Yangling 712100,China;Beijing Kedong Electric Power Control System Co.,Ltd.,Beijing 100000,China)

机构地区：[1]西北农林科技大学水利与建筑工程学院,陕西杨凌712100 [2]北京科东电力控制系统有限责任公司,北京100000

出　　处：《水电与新能源》2024年第1期75-78,共4页Hydropower and New Energy

摘　　要：针对水电机组振动信号故障特征提取难,提出一种融合小波变换(Wavelet Transform,WT)和奇异值分解(Singular Value Decomposition,SVD)相结合的故障特征提取方法。首先,通过小波阈值降噪消除强噪声对模型特征提取的干扰,再利用小波变换将降噪信号分解成不同频率的模态子序列,应用SVD理论提起子序列的SVD值作为特征,最终将特征输入RF模型中实现水电机组故障的快速识别与诊断。通过在公开数据集和真实机组案例中应用,验证了对水电机组故障诊断的高效性。It is usually difficult to extract fault features from vibration signals of hydropower units.Thus,a fault feature extraction method combining the wavelet transform(WT)and the singular value decomposition(SVD)is proposed.Firstly,the interference of strong noise on model feature extraction is eliminated through wavelet threshold de-noising.Then,the de-noised signals are decomposed into modal subsequences in different frequencies using the wavelet transform.SVD theory is applied to extract the SVD values of the subsequences as the fault features.At last,the features are input into the RF model to realize the rapid fault identification and diagnosis of hydropower units.Application of the proposed method in public datasets and actual cases verifies its efficiency in fault diagnosis of hydropower units.

关 键 词：小波变换 奇异值分解 随机森林 特征提取 水电机组故障诊断 

分 类 号：TM312[电气工程—电机]