基于时变滤波与CWT的齿轮箱无转速计阶次跟踪

Tacho-less Order Tracking for Gearbox based on Time-varying Filter and CWT

针对转频波动工况下复杂齿轮箱系统振动信号中各啮合阶次成分相互干扰,导致对啮合频率估计困难的情况,提出一种基于时变滤波与连续小波变换(CWT)结合的无转速计阶次跟踪齿轮箱故障诊断技术。基于短时傅里叶变换(STFT)设置时变滤波器;通过时变滤波及连续小波变换(CWT),获得平滑的瞬时频率估计(IFE);再基于Vold-Kalman滤波(V KF)获得参考轴相位。该方法获得的阶次谱与计算阶次跟踪(COT)相当。最后,结合快速谱峭度算法可准确判断故障齿轮位置。通过仿真和实验表明,无转速计方法能为强噪声转速明显波动工况下的机械故障诊断提供一种新的有效手段。

In the fault diagnosis of complex gearbox system under the rotation frequency fluctuation, the orders of meshing frequency interfere with one another,so it is difficult to estimate gear meshing frequen cy. For this problem,a tacho-less order tracking technique combining with time-varying filter and continuous wavelet transform(CWT)is proposed. Time varying filter is set based on short-time Fourier transform(ST FT). The smooth instantaneous frequency estimation(IFE)is obtained based on time-varying filter and CWT. The phase of the reference shaft is calculated with Vold-Kalman filter(VKF). The order spectrum obtained by the proposed method is basically consistent of the computed order tracking(COT). Lastly,the position of fault gear can be accurately judged by Fast-kurtogram algorithm. The results of simulation and experiment show that the proposed method can provide a new effective method for mechanical fault diagnosis under the condition of strong noise and apparent speed variation.