基于小波相关特征尺度熵的预测特征信息提取方法研究

Research on Approach of Prognostics Feature Information Extraction Based on Wavelet Correlation Feature Scale Entropy

提出一种小波相关特征尺度熵WCFSE的预测特征信息提取方法。将小波相关滤波法与Shannon信息熵理论相结合,给出了沿尺度分布的WCFSE的定义及其计算方法。WCFSE定量表征不同尺度的能量分布,各尺度能量分布的均匀性反映设备运行状态的差别,选取最能反映故障特征的WCFSE作为特征参数来判断设备运行状态。正常和几种故障程度不同的滚动体运行状态的识别结果验证了该方法的有效性和实用性。

A prognostics feature information extraction approach was proposed based on wavelet correlation feature scale entropy WCFSE. The defining and computing way of WCFSE were presented based on integration of Shannon information entropy theory and wavelet transform correlation filter. WCFSE can quantitatively describe energy distribution of different scales which can response difference of equipment running states, the equipment running states were estimated by magnitude of the selected WcFsr which can mostly embody fault features. Several states of roller with normal state and different fault severity states were recognized by the proposed method. Experimental results show that the proposed method is effective and practicable, and a new approach is proposed for the fcature information extraction of equipment degradation state recognition and fault prognostics.