中文摘要：

时变工况下的行星齿轮箱振动信号具有明显的时变调制特点,常规的频谱分析方法难以识别齿轮故障特征频率。提出了基于Vold-Kalman滤波和能量分离的时频分析方法,识别行星齿轮箱的时变特征频率,诊断齿轮故障。与传统的时频分析方法相比,基于Vold-Kalman滤波和能量分离的时频表示具有良好的时频分辨率,而且没有交叉项干扰,能够有效提取非平稳信号中的时变频率成分。分析了行星齿轮箱时变工况下的实验信号,准确地诊断了齿轮故障,验证了该方法的有效性。

英文摘要：

The vibration signals of planetary gearboxes under nonstationary running conditions have significant time-varying modulation feature.Conventional spectral analysis methods are unable to identify the gear fault characteristic frequencies from such nonstationary signals.A method based on Vold-Kalman filter and higher order energy separation is proposed to analyze the vibration signals of planetary gearboxes under nonstationary conditions in time-frequency domain,thus to identify the time-varying characteristic frequencies and diagnose the gear faults.The time-frequency representation derived from Vold-Kalman filter and higher order energy separation provides nice time-frequency resolution and is free from cross term interference,and thus it can effectively pinpoint the time-varying constituent frequency components of nonstationary signals.The proposed method is validated with analysis of lab experimental signals of a planetary gearbox under time-varying running conditions.