时频分析-数学进展-上海交通大学.PDFVIP

198 机械故障诊断技术中的信号处理方法：时频分析 2013 年6 月 文章编号：1006-1355(2013)03-0198-05 机械故障诊断技术中的信号处理方法：时频分析 王金福，李富才 ( 上海交通大学 机械系统与振动国家重点实验室，上海 200240 ) 摘 要：当机械设备的振动信号为非平稳信号和时变信号这类特殊信号时，时域分析和频域分析因其自身的局限， 无法取得很好的分析效果，需要使用时间和频率的联合函数来表示信号，即信号的时频表示。针对常用的时频分析振 动信号处理方法，总结多种算法的特征和优缺点。根据常见机械设备关键构件的振动特征，选择不同的信号处理和特 征提取算法进行分析，以便提高多种构件、多类故障的特征提取精度和可靠性，从而为有效地实现机械设备的故障提 供参考和指导。 关键词：振动与波；故障诊断；振动信号；特征提取；时频分析 + 中图分类号：TB53；TH 165 .3 文献标识码：A DOI 编码：10.3969/j.issn. 1006-1335.2013.03.045 ReviewReview ofof SignalSignal ProcessingProcessing MethodsMethods inin FaultFault DiagnosisDiagnosis forfor MachineryMachinery UsingUsing Time-frequencyTime-frequency AnalysisAnalysis WANG Jin -fu , LI Fu -cai ( State Key Laboratory of Mechanical System and Vibration, Shanghai Jiaotong University, Shanghai 200240, China ) AbstractAbstract :: When vibration signals of mechanical equipment are non-stationary or time-varying, neither time domain analysis nor frequency-domain analysis can provide satisfactory results. In contrast, the time-frequency analysis can signify the signals well. In this paper, the characteristics, advantage and disadvantage of several time-frequency analysis methods were summarized for the purpose of machinery fault diagnosis. According to the vibration characteristics of the key components of some popular mechanical equipment, some signal processing methods and feature extraction algorithms were selected and analyzed in order to improve the precision and reliability for multi-fault diagnosis of the components. KeyKey wordswords :: vibration and wave ; fault diagnosis ; vibration signal ; feature extraction ; time-frequency an