结合VMD符号熵与SVDD的滚动轴承性能退化评估

Performance Degradation Assessment Model of Rolling Bearings Combining VMD Symbol Entropy and SVDD

针对滚动轴承的早期故障难以检测的问题,提出了一种基于变模态分解(Variable mode decomposition,VMD)符号熵和支持向量数据描述(Support vector data description,SVDD)的滚动轴承性能退化评估模型。首先对振动信号进行VMD分解并提取各个分量符号熵,并采用双样本Z值对各个分量符号熵进行评价,选取双样本Z值最大的特征作为特征向量。特征提取完毕后,采用SVDD模型进行性能退化评估,使用全寿命数据进行模型的验证。实验结果表明,相比于排列熵特征提取方法以及模糊C均值聚类(Fuzzy c-means clustering,FCM)方法,该模型可以更好显示出滚动轴承性能退化规律。

In order to solve the problem of early fault detection of rolling bearings,a performance degradation assessment model of rolling bearings combining variable mode decomposition(VMD)symbol entropy and support vector data description(SVDD)was proposed in study.Firstly,the vibration signal is decomposed by VMD and the symbol entropy of each component is extracted.Then,the Z value of double samples is used to evaluate the symbol entropy of each component,and the feature with the largest Z value of double samples is selected as the feature vector.After feature extraction,SVDD model is finally used to evaluate performance degradation,and life cycle data is used to verify the model.The experimental results show that this model can better show the performance degradation law of rolling bearings compared with permutation entropy(PE)feature extraction method and fuzzy c-means clustering(FCM)method.