摘要
提出一种结合多层结构和稀疏最小二乘支持向量机(Sparse Least Squares Support Vector Machine,SLSSVM)的机械故障诊断方法。该方法构建了多层支持向量机(Support Vector Machine,SVM)结构,首先在输入层利用支持向量机对信号进行训练,学习信号的浅层特征,利用“降维公式”生成样本新的表示,并作为隐藏层的输入,隐藏层支持向量机对新样本训练并提取信号的深层特征,逐层学习,最终在输出层输出诊断结果。针对因多层结构带来算法的复杂度以及运行时间增加的问题,采用最小二乘支持向量机(Least Squares Support Vector Machine,LSSVM)技术,并将稀疏化理论与最小二乘支持向量机结合,通过构造特征空间近似最大线性无关向量组对样本进行稀疏表示并依此获得分类判别函数,有效解决了最小二乘支持向量机稀疏性缺乏的问题。最后,通过滚动轴承故障诊断实验验证了该方法的有效性。
A mechanical fault diagnosis method combining multi-layer structure and sparse least squares support vector machine is presented.This method constructs a multi-layer support vector machine structure.Firstly,the fault signal is trained and the shallow features are learned by support vector machine at the input layer.And then a new representation of samples is generated through the“dimension reduction formula”and used as the input for the hidden layers.The hidden layer support vector machine trains new samples and extracts the deep features of the signal layer by layer.Finally the diagnostic results are output at the last layer.Considering the increase of the running time and algorithm complexity caused by the multi-layer,this paper combines the sparsification theory with least squares support vector machine technology.The approximate linear independent vector set as sparse representation of the samples in feature space is searched and thus the discriminative function can be constructed.Lack of sparseness is effectively solved and the validity of this method is verified by the fault diagnosis experiments of the rolling bearing.
作者
张瑞
李可
宿磊
李文瑞
ZHANG Rui;LI Ke;SU Lei;LI Wen-rui(Jiangsu Key Laboratory of Advanced Food Manufacturing Equipment and Technology,Jiangnan University,Wuxi 214122,China;School of Mechanical Engineering,Jiangnan University,Wuxi 214122,China;School of Mechanical Engineering,Donghua University,Shanghai 201620,China)
出处
《振动工程学报》
EI
CSCD
北大核心
2019年第6期1104-1113,共10页
Journal of Vibration Engineering
基金
国家自然科学基金资助项目(51775243)
江苏省重点研发计划(BE2017002)
江南大学自主科研计划重点项目(JUSRP51732B)
关键词
故障诊断
滚动轴承
多层结构
支持向量机
稀疏化
fault diagnosis
rolling bearing
multi-layer structure
SVM
sparsification
