马氏田口度量的滚动轴承故障诊断的特征选择优化方法

Optimization method of characteristic variable reduction for Mahalanobis-Taguchi system

研究了基于马氏田口度量的滚动轴承故障诊断特征变量降维优化问题,提出一种新的混合马氏田口系统变量筛选方法。首先,计算马氏田口距离误分类率的加权和,建立变量筛选的混合整数线性规划模型;其次,运用Gompert二进制粒子群优化算法求解最优的滚动轴承故障诊断特征变量组合;最后,与标准二进制粒子群优化算法和二进制蚁群优化算法进行对比,结果表明,该方法能显著提高滚动轴承故障识别准确率和计算效率。

Studies the dimensionality reduction optimization of Mahalanobis-Taguchi system variables.A new method of variable selection for hybrid Mahalanobis-Taguchi system is proposed.Firstly,calculate the weighted sum of the Mahalanobis-Taguchi distance misclassification rate,and establish a mixed integer linear programming model for variable selection.Then,the Gompert binary particle swarm optimization algorithm is applied to solve the optimal variable combination.Finally,takes the fault diagnosis of rolling bearing as an example and compares it with the standard binary particle swarm optimization algorithm and binary ant colony optimization algorithm.