基于次序二叉树SVM的矿井提升机制动系统故障诊断  

Fault diagnosis on braking systems of mine hoists based on ordering binary tree SVM

作　　者：王莹[1] 高雅利[2] 马建伟[1] 

机构地区：[1]河南科技大学电子信息工程学院,河南洛阳471003 [2]洛阳理工学院电气工程与自动化系

出　　处：《矿山机械》2011年第9期46-49,共4页Mining & Processing Equipment

摘　　要：介绍几种传统的支持向量机(SVM),提出了一种基于次序二叉树支持向量机的多类分类算法。该算法采用样本分布半径和分布距离估算各类别样本在高维特征空间中的分布情况,更精确地确定其在特征空间中的分类区域。利用该算法对提升机制动系统的故障诊断进行仿真分析,结果表明,该方法具有诊断速度快且故障识别率高的特点。The paper introduces several traditional support vector machines（SVM）,and proposes a multi-class classification algorithm based on ordering binary tree SVM.In the algorithm,the sample distribution radius and sample distribution distance are used to estimate the distribution situation of various samples among the high-dimension characteristic space,so as to determine their classification zones in the characteristic space more precisely.After the simulation analysis of fault diagnosis on the hoist braking system by the algorithm,it is showed that the algorithm is characterized by high diagnosis velocity and high fault identification ratio.

关 键 词：支持向量机 故障诊断 二叉树 矿井提升机 

分 类 号：TD534.5[矿业工程—矿山机电] TP18[自动化与计算机技术—控制理论与控制工程]