双树复小波滤波器构造及其在机械故障诊断中的应用

摘要

In order to reduce the energy leakage and to improve the time-frequency analysis ability of the dual tree complex wavelet transform(DT-CWT) for non-stationary signal of machinery fault,the transfer function of the filter is improved to guarantee the orthogonality and to enhance the symmetry.The filter algorithm is also improved to make the reconstruction accuracy better.Finally,the initial shock response vector of the filter of scaling function is optimized.The calculation method is presented to construct the dual tree complex wavelet and filter based on the Hibert transform and Q-shift.The basic requirements for the filter design and calculation is proposed.The comparative experiments were conducted for the proposed method and the ordinary complex wavelet transform and discrete wavelet transform.The experimental results show that the reduction effect of noise and energy leakage for the proposed wavelet is better than othersThe present method was used to extract the weak failure information in machinery vibration signals to diagnose the mechanical fault.The experiments indicate that this method can obtain high signal-to-noise ratio (SNR) than the common methods and extract the fault feature submerged in a heavy noise effectively.%为了减少能量泄露,提高双树复小波对机械故障非平稳信号的时频分析能力,从滤波器的构造入手,首先改进滤波器传递函数,保证滤波器正交性,增强对称性;优化分层滤波器算法,加强滤波器紧支性,提高重构精度;最后对尺度函数滤波器初始冲击响应向量进行时频域优化.根据复小波实部和虚部互为hilbert变换的关系和Q-移位的思想,给出了滤波器设计计算的基本要求和原则.将该法构造的小波滤波器引入能量泄露和信号降噪仿真实验,并在实践中诊断滚动轴承的机械故障.实验表明:该法构造的双树复小波对能量泄露有明显的抑制作用,信号降噪可以获得比常规双树复小波、离散小渡更高的信噪比,能更好地提取隐含在强噪声中故障信息.说明该双树复小波滤波器算法的有效性和可行生.