基于QSTFT的同步压缩变换及在轴承故障诊断中的应用

Synchrosqueezing Transform Based on QSTFT and Its Application in Mechanical Fault Diagnosis

高阶同步压缩变换以同步压缩变换为基础,能够有效地估计纯调制信号的瞬时频率,提升时频面的能量集中水平。但由于受到窗口固定的限制,其难以对频率快速变化的多分量信号获得清晰的时频表示。为此,引入缝式短时傅里叶变换,其能够根据信号的特点,在特定区域自适应地选择最佳的窗口,克服了窗口时宽不变的缺陷。在此基础上,结合高阶同步压缩变换,提出一种基于缝式短时傅里叶变换的自适应高阶同步压缩变换算法,进一步提升对复杂信号时频特征的提取能力。通过对仿真信号和轴承外圈故障信号进行分析,验证了该算法的有效性。

Despite the fact that high-order synchrosqueezing transform(HSST)based on synchrosqueezing transform can effectively estimate the instantaneous frequency of the pure modulated signal and improve the energy concentration level in the time-frequency plane,it is difficult to obtain a clear time-frequency representation for multi-component signals with rapidly changing frequencies due to the limitation of fixed window.A quilted short-time Fourier transform,therefore,is introduced,which can adaptively select the best window in specific area according to the signal features and overcome the defect of constant window width.On account of this and by combination with HSST,an adaptive high-order synchrosqueezing transform based on a quilted short-time Fourier transform method is proposed to further improve the ability to extract time-frequency features of complex signals.The validity of the algorithm is verified by analyzing the simulation signal and the fault signal of bearing outer ring.