二叉树型多分类器融合的轴承故障诊断方法

Method of binary tree structure based multiple classifier fusion in bearing fault diagnosis

多分类器融合能有效集成多种分类算法的优势,实现优势互补,提高智能诊断模型的稳健性和诊断精度。但在利用多数投票法构建多分类器融合决策系统时,要求成员分类器数目多于要识别的设备状态数,否则会出现无法融合的情况。针对此问题,提出了一种基于二叉树的多分类器融合算法,利用二叉树将多类分类问题转化为多个二值分类问题,从而各个节点上的成员分类器个数只要大于2即可,有效避免了成员分类器数目不足的问题。实验结果表明,相比单一分类器的诊断方法,该方法能有效地实现滚动轴承故障智能诊断,并具有对各神经网络初始值不敏感、识别率高且稳定等优势。

The fusion of multiple classifiers harnesses the advantages of various classification algorithms, and thus improves the robustness and accuracy of intelligent diagnosis models. When majority voting scheme is employed to construct a multi-classifier decision fusion system, the number of the required member classifiers is usually bound to be larger than that of the patterns to be recognized. Otherwise, it is difficult to achieve decision fusion in certain cases. Aiming at this issue, a multi-classifier fusion algorithm is presented using the form of binary tree, which transforms the multi-classifica-tion problem into a series of binary classification problems. In each binary classification, the number of member classifier is to be larger than 2, thus avoiding the requirement on large number of member classifiers. Experimental results demon-strate the proposed paradigm can effectively improve the recognition accuracy and stability of rolling bearing fault diagnosis in comparison with the diagnosis method based on a single classifier.