基于多尺度排列熵与双核极限学习机的滚动轴承故障诊断方法

Fault diagnosis of rolling bearings based on multi-scale permutation entropy and dual kernel extreme learning machine

针对极限学习机隐含层节点数需人为设定,分类的准确性与稳定性较差,核极限学习机(K-ELM)对核函数选取要求较高,单一核函数难以对非线性样本充分学习、泛化性仍有不足等缺点,提出一种基于多尺度排列熵(MPE)和非线性加权组合的双核极限学习机(DK-ELM)的滚动轴承故障诊断方法并证明了其可行性与优越性。首先,计算不同故障状态轴承信号的多尺度排列熵,获取一系列无量纲特征;然后,利用双核函数计算其高维特征向量集并输入DK-ELM中建立轴承信号状态分类模型,对不同状态的轴承信号进行分类。实验结果证明,核函数的引入可以有效提高ELM分类性能,DK-ELM的分类模型比支持向量机(SVM)、ELM以及各单核极限学习机具有更高的分类精度,而且对训练样本数量较少的情况有更好的分类效果。

The number of hidden layer nodes of the extreme learning machine needs to be set by artificial, and the accuracy and stability are not good, The kernel extreme learning machine ( K-ELM ) has high requirements for the selection of kernel functions. The single kernel function cannot fully learn the non-linear feature, and the generalization is still insufficient. A fault diagnosis method of rolling bearing based on multi-scale permutation entropy ( MPE) and nonlinear weighted-dual kernel extreme learning machine ( DK-ELM ) is proposed and its feasibility and superiority are proved. First, the MPE of bearing signal in different fault states is extracted. Next, the high dimension feature vector is expressed by dual kernel and used as the input of the DK-ELM to establish the classification model to determine the state of signal. The result of experimental shows that the addition of kernels can improve the performance of ELM, classification model of DK-ELM has higher classification accuracy than sopport vector machine ( SVM ), ELM and single-kernel ELM , and has better classification effect on the insufficient number of samples.