基于FOA-MKSVM的滚动轴承故障分类方法

由于滚动轴承实际各状态数据一般具有不均衡的特点,所以分类时采用单一核函数存在一定的局限性。针对此问题以及支持向量机多参数选择的盲目性,建立一种基于果蝇优化算法的多核支持向量机模型。该模型可以通过核函数权值来调节全局核函数和局部核函数在该模型中的作用,兼具了良好的学习能力和泛化能力。同时,将多核支持向量机参数与果蝇算法中食物的味道浓度值建立一定关系,通过模仿果蝇觅食行为,对各参数进行优化选择。为了验证所提方法的有效性,先利用UCI标准数据集进行实验,再将其应用到滚动轴承故障分类中,并对单核核函数与多核核函数及参数优化算法进行比较。结果表明,提出的方法具有初始化参数少、参数设置简单、全局搜索能力强和分类准确率高的优点,可有效地应用到滚动轴承故障分类中。

Fault classification method of rolling bearing based on FOA MKSVM

Because of the uneven characteristics of rolling bearing actual state data, the single kernel function used in classification stage has certain limitation. Aiming at this problem and the blindness of multi parameter selection of support vector machine, a multi kernel support vector machine model is proposed based on fruit fly optimization algorithm (FOA). The actions of global kernel function and local kernel function in the model can be adjusted with kernel function weight, and the model has both good learning ability and generalization ability. Meanwhile, a certain relation between the parameters of the multi kernel support vector machine and the smell concentration of the food in FOA is built; by imitating fruit fly foraging behavior, the parameters can be optimized and selected. In order to verify the effectiveness of the proposed method, UCI standard data set was first used in the experiment, then the method was applied in rolling bearing fault classification. The single kernel function, multi kernel function and parameter optimization algorithms were compared, the results show that the proposed method has some advantages, such as less initialization parameters, simple parameter setting, strong global search ability and high classification accuracy rate, and can be effectively applied in rolling bearing fault classification.