基于小波分析的进化谱及在故障诊断中的应用

Evolutionary spectrum estimation based on the wavelet analysis and its application in fault diagnosis

在Priestley提出的基于小波分析进化谱理论基础上 ,对改进进化谱估计的性能进行了研究 .首先 ,重点分析了小波分解的尺度与信号时变趋势和分辨率之间的关系 ,并指出了根据信号时变特征和分辨率要求确定合理的小波分解尺度 .其次 ,讨论了权函数对进化谱估计能量泄漏现象的影响 ,指出了选用Hanning窗等作为权函数可以有效地抑制谱泄漏现象 .仿真及在故障诊断中的应用与理论分析的结论一致 .

The improvement of the accuracy of the wavelet-based evolutionary spectrum estimation is discussed, which is based on the original work of Priestley. Firstly, the relation between wavelet decomposition level, the time-dependent feature of signal and resolution requirement is analyzed for determining the best depth of wavelet decomposition in order to get an evolutionary spectrum with clear time-dependent trend and high resolution. Secondly, the role of weight function in the estimation of evolutionary spectrum is presented, which shows that Manning window is more effective than rectangle window in reducing the energy leakage in spectrum analysis. Finally, a series of simulation and application results are given and these results show that the method proposed in this paper is feasible, which can offer a more accurate evolutionary spectrum.