基于多尺度包络谱图的直驱式风电机组轴承故障特征提取  

Bearing Fault Feature Extraction for a Direct Drive Wind Turbine Using Multi-scale Enveloping Spectrogram

作　　者：滕伟[1] 姜锐[1] 张阳阳[1] 柳亦兵[1] 马志勇[1] 

机构地区：[1]华北电力大学能源动力与机械工程学院,北京102206

出　　处：《动力工程学报》2017年第5期373-378,共6页Journal of Chinese Society of Power Engineering

基　　金：国家自然科学基金资助项目(51305135);北京高等学校青年英才计划资助项目(YETP0701);中央高校基本科研业务费资助项目(2015ZD15)

摘　　要：针对直驱式风电机组低速、重载的运行特点,提出将多尺度包络谱图应用于实际机组的故障诊断,以提取轴承的故障特征.对原始振动信号进行复小波变换,其变换结果的包络谱即为轴承振动信号的多尺度包络谱图.结果表明:因其所具有的同步多尺度分解和包络解调能力,多尺度包络谱图能够提取出隐藏在噪声中的低频微弱轴承故障特征,与传统包络解调方法相比具有较好的智能性和准确性,适用于实际风电机组的状态监测.Considering the features of low rotational speed and heavy load of direct drive wind turbines, the multi-scale enveloping spectrogram （MuSEnS） is proposed for the diagnosis of actual wind turbines and for the feature extraction of faulty bearings, which is the envelop spectrum of transform results by processing the original vibration signal using complex wavelet transform. Diagnosis results show that the MuSEnS can effectively detect the weak faulty feature of bearing hidden in background noise, because it has the a- bility of simultaneous multi-scale decomposition and envelope demodulation. Comparing with conventional demodulation analysis, the MuSEnS is more intelligent and accurate, which can be applied to the fault di- agnosis of actual wind turbines.

关 键 词：直驱式风电机组 轴承 多尺度包络谱图 特征提取 

分 类 号：TM315[电气工程—电机]