基于WSGMD和AMOMED的风力发电机齿轮箱轴承早期故障诊断  

Early Fault Diagnosis for Wind Generator Gearbox Bearing based on WSGMD and AMOMED

作　　者：王瑜 严国斌 王广玲 张弈鹏 Wang Yu;Yan Guobin;Wang Guangling;Zhang Yipeng(China Three Gorges Renewables(Group)Co.,Ltd.,Beijing 100053)

机构地区：[1]中国三峡新能源(集团)股份有限公司,北京100053

出　　处：《现代科学仪器》2023年第1期210-217,共8页Modern Scientific Instruments

基　　金：中国三峡新能源(集团)股份有限公司科研项目资助(合同编号三峡新能源合字【2021】254)。

摘　　要：围绕风力发电机齿轮箱轴承早期故障诊断困难这一难题,提出一种基于加权辛几何模态分解和自适应多点最优最小熵解卷积的诊断新方法。首先,采用辛几何模态分解将原始振动信号分解为多组辛几何分量,计算各分量的Gini系数并据此对各分量进行加权重构,得到信噪比提升的重构信号。为了进一步强化放大重构信号中的故障特征,通过鲸鱼优化算法对多点最优最小熵解卷积的滤波长度及冲击周期进行自动搜索,获取最佳参数后对重构信号做进一步解卷积处理,最终通过解卷积信号的包络解调分析来识别轴承服役状态。风电现场机组振动监测信号验证表明,该方法可准确辨识出轴承早期故障,具有广阔的工程应用前景。In order to solve the problem of early fault diagnosis of wind generator gearbox bearing,a novel diagnosis method based weighted symplectic geometry mode decomposition(WSGMD)and adaptive multipoint optimal minimum entropy deconvolution(AMOMED)is proposed.Firstly,the original vibration signal is decomposed into several symplectic geometry components by symplectic geometry mode decomposition(SGMD),the Gini index of every component is calculated and used to reconstruct the component,and then the reconstructed signal with higher signal noise ratio is obtained.In order to enhance and amplitude the fault feature of the reconstructed signal,the filter length and the shock period of the multipoint optimal minimum entropy deconvolution(MOMED)is automatically searched by the whale optimization algorithm(WOA),the reconstructed signal is further deconvolution processed after the optimal parameters are obtained,and then the bearing service condition can be judged by envelope demodulation analysis of the deconvolution signal.The analysis result of the unit vibration monitoring signal in the wind power field verifies that,the proposed method could accurately identify the bearing early fault,and has broad engineering application prospect.

关 键 词：风力发电机 轴承早期故障 辛几何模态分解 多点最优最小熵解褶积 鲸鱼优化算法 

分 类 号：TN9[电子电信—信息与通信工程]