基于MRMD的滚动轴承损伤程度识别方法研究

Research on the Recognition Method on Rolling Bearing Fault Severity Based on MRMD

为解决滚动轴承损伤程度难以识别的问题,提高故障诊断的准确率,将特征选择方法应用到滚动轴承故障诊断中。在建立多域特征集的基础上提出一种基于MRMD(Max Relevance Max Distance)评价准则的特征选择方法完成对轴承损伤程度的评估。首先从原始信号中提取能够表征轴承运行状态变化的时频域统计特征并建立多域特征集;然后利用MRMD特征选择方法去除特征集中的无关特征和冗余特征,筛选出敏感特征;最后将筛选出的故障特征样本输入到概率神经网络(PNN)中得到损伤程度的评估结果,利用该特征选择方法可以实现轴承裂纹损伤程度的识别。以分类器正确率为依据,验证了基于MRMD特征选择方法的有效性和优越性。

In order to solve the problem that the fault severity in rolling bearings is difficult to identify and improve the accuracy of fault diagnosis, the feature selection method is applied to the fault diagnosis of rolling bearings. Based on the establishment of multi-domain feature sets, a feature selection based on maximum relevance maximum distance evaluation criteria is proposed. The method to complete the severity assessment of bearing damages. Firstly, the time-frequency statistical characteristics that can characterize the change of the running state of the bearing are extracted from the original signal and multi-domain feature set is established;then, the irrelevant features and redundant features in the feature set are removed by using the MRMD feature selection method, and the sensitive features are selected;The fault feature samples are input into the probability neural network to obtain the assessment result of the degree of damage. With this feature selection method, the bearing fault severity can be identified. Based on the correct rate of classifiers, the effectiveness and superiority of the MRMD feature selection method were verified.