基于局部最大值2阶同步压缩变换的变工况轴承故障诊断

Fault Diagnosis of Bearings with VariableWorking Conditions Based on Local Maximum Second-order Synchro-squeezing Transform

由于工况变化,机械设备在运行过程中相应的状态监控信号也呈非平稳性。时频分析技术作为处理非平稳信号的有效工具在机械故障诊断应用中起着重要作用。针对2阶同步压缩变换(Second-order Synchrosqueezing Transform,SST2)时频分析技术存在的时频分布能量发散问题,利用SST2局部最大时频系数位置构建局部最大值估计算子,将短时傅里叶变换结果(Short-time FourierTransform,STFT)的时频能量进行重排。结果表明,局部最大值估算算子能够有效增强SST2时频分布的能量集中度,该方法不仅能更准确的捕捉非平稳信号中的时变信息,同时能获得能量更集中的时频分布,减少噪声对时频分布图谱的干扰。通过一组数值仿真信号与一组变速轴承故障实验数据分析验证了该方法的有效性,并结合阶次分析方法实现变工况滚动轴承故障诊断。

Condition monitoring signals of machines often demonstrate non-stationary characteristics due to speed and/or load changes during operation.Time-frequency analysis technique is an effective tool for the analysis of non-stationary signals which plays an important role in machinery fault diagnosis.In this paper,aiming at the energy dissipation problem in time-frequency analysis of the second-order synchro-squeezing transform(SST2),the local maximum time-frequency coefficient position is applied to construct a local maximum estimation operator so as to improve the energy concentration of the signal.The results show that the local maximum estimation operator can effectively enhance the energy concentration of the time-frequency distribution.This method can capture the time-varying information in the non-stationary signal more accurately,and reduce the noise interference to the analysis result.Validity of the proposed technique is verified through the analyses of a set of simulated non-stationary signals and a set of bearing experimental data acquired under varying speed condition.And the rolling bearing fault diagnosis under variable working conditions is realized combined with order analysis method.