基于VMD和改进多分类马田系统的滚动轴承故障智能诊断

为了有效提取滚动轴承的故障信号,选择合适的智能分类器识别故障状态,提出基于变分模态分解及多重马氏距离法的多分类马田系统的故障智能诊断系统。通过变分模态分解将振动信号分解为多个本征模函数并提取相关特征;并采用了多重马氏距离法的马田系统,以特征子集代替特征参与分类器的构建,以解决特征参数众多的问题;通过正交表和信噪比,筛选出各状态的敏感模态分量,并提出多分类马田系统,用于多类故障智能识别;将其应用于滚动轴承故障数据中,验证算法的有效性,并与其他算法对比分析。结果表明,基于变分模态分解及改进的多分类马田系统算法能简化诊断系统、训练耗时少,识别准确率高,是一种更为有效的故障智能诊断方法。

Intelligent fault diagnosis of rolling bearings based on the VMD and improved-multi-classification Mahalanobis Taguchi system

In order to effectively extract the fault signal of rolling bearings and to select a suitable intelligent classifier for the identification of fault types, the variational mode decomposition and multi-classification Mahalanobis Taguchi system with multiple Mahalanobis distance were proposed to construct an intelligent fault diagnosis system.Vibration signals were decomposed into several intrinsic mode functions by the variational mode decomposition.Features of each mode function were extracted.To deal with the issue of excessive number of features in diagnosis system, the Mahalanobis Taguchi system was adopted based on multiple Mahalanobis distances.In stead of individual features, feature subsets were used in participating in the construction of the classifier.Orthogonal arrays and signal-to-noise ratios were used to select key intrinsic mode functions and the multi-classification Mahalanobis Taguchi system was utilized to diagnose faults intelligently.To verify the validity of the method, rolling bearing fault data were tested and the results were compared with other intelligent algorithms.The results indicate that the proposed algorithm is of advantages of algorithm higher accuracy, simplified diagnostic complexity, and decreased training time.It is an efficient intelligent fault diagnosis method.