基于自适应随机共振方法的轴承故障特征提取  

Fault Feature Extraction of Wind Turbine Bearing Based on Adaptive Stochastic Resonance Method

作　　者：俞勤新 彭艳来 杨晓峰 杨宏宇 张佐辉 葛岳 王志新[3] YU Qinxin;PENG Yanlai;YANG Xiaofeng;YANG Hongyu;ZHANG Zuohui;GE Yue;WANG Zhixin(Longyuan Power Group(Shanghai)New Energy Co.,Ltd.,Shanghai 202155,China;Shanghai Proinvent Information Tech.Ltd.,Shanghai 201111,China;Shanghai Jiaotong University,Shanghai 200030,China)

机构地区：[1]龙源电力集团(上海)新能源有限公司,上海202155 [2]上海博英信息科技有限公司,上海201111 [3]上海交通大学,上海200030

出　　处：《机械制造与自动化》2023年第5期92-95,共4页Machine Building & Automation

基　　金：国家重点研发计划项目(2018YFB1503000)。

摘　　要：针对现有故障诊断方法对风力发电机组轴承微弱故障特征识别效果较差的问题,提出一种峭度值优化的自适应变尺度随机共振方法。基于经验模态分解理论对风电机组轴承振动信号进行分解,利用峭度准则分别计算出各个固有模态分量的峭度值,并设置合适的峭度阈值对各分量信息进行提取。利用蚁群算法对随机共振系统进行智能参数优化,以信噪比为目标函数,获取最佳的输出信号,进而识别风电机组中轴承故障类别。理论分析与实验结果表明:此方法能够准确地识别风电机组轴承信号中的微弱故障类别,在风力发电机组轴承故障诊断与性能维护中具有良好的应用前景。Aiming at the ineffectiveness in identifying weak fault features of wind turbine bearings by existing fault diagnosis methods,an adaptive variable scale stochastic resonance method with kurtosis optimization is proposed.Based on Empirical Mode Decomposition theory,the vibration signals of wind turbine bearings are decomposed,and the kurtosis values of each intrinsic modal component are calculated by kurtosis criterion,and appropriate kurtosis thresholds are set to extract the information of each component.The ant colony algorithm is applied to optimize the intelligent parameters of stochastic resonance system,and the signal-to-noise ratio is taken as the objective function to obtain the best output signal,and identify the bearing fault categories in wind turbines.Theoretical and experimental results show that the proposed method can accurately identify weak fault types in wind turbine bearing signals,and has a good application prospect in wind turbine bearing fault diagnosis and performance maintenance.

关 键 词：风力发电机组 滚动轴承 微弱故障 故障诊断 

分 类 号：TM614[电气工程—电力系统及自动化]