小波相关特征尺度熵和隐半马尔可夫模型在设备退化状态识别中的应用

Application of Wavelet Correlation Feature Scale Entropy and Hidden Semi-Markov Models to Equipment Degradation State Recognition

为正确识别机械设备当前所处的退化状态,预防设备进一步退化和故障的发生,提出一种基于小波相关特征尺度熵和隐半马尔可夫模型(Hidden semi-Markov models,HSMM)的设备退化状态识别新方法。对采集到的设备振动信号进行小波相关滤波处理,得到信噪比较高的尺度域小波系数,在此基础上结合信息熵理论提出了沿尺度分布的小波相关特征尺度熵概念。构造信号的小波相关特征尺度熵/矢量,并以此矢量作为HSMM的输入进行训练,建立基于HSMM的机械设备运行状态分类器,从而实现设备退化状态的识别。以滚动轴承为例,对正常和几种故障程度不同的滚动体运行状态进行了识别,同时还与基于小波相关特征尺度熵-HMM的状态识别法进行了比较,试验结果表明该方法能有效识别设备的退化状态。

In order to correctly recognize the current degradation state of equipment for preventing equipment from farther degradating and going wrong, a new method of equipment degradation state recognition based on wavelet correlation feature scale entropy（WCFSE） and HSMM is proposed. The gathered vibration signal of equipment is processed by way of the wavelet transform correlation filter（WTCF）. In order to get the high signal-to-noise scales wavelet coefficients, the conception of WCFSE is presented based on the integration of information entropy theory and WTCF, and the WCFSE eigenvectors of signal are constructed. Those WCFSE eigenvectors are inputted to the HSMM for training, and a running state classifier of medical equipment based on HSMM is constructed to recognize the equipment degradation state. Roller bearings are taken as examples, and the running states of a normal rolling element and several rolling elements with different degree of fault are recognized by using the proposed method. This method is compared with the state recognition method based on WCFSE and HMM, experimental results show that this method can effectively recognize the degradation state of equipment.