«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

期数:

2012年04期

页码:

175-178

栏目:

应用开发研究

出版日期:

1900-01-01

文章信息/Info

Title:

SVM Fault Diagnosis Research Based on Attribute Reduction and Parameters Optimization

文章编号:

1673-629X（2012）04-0175-04

作者:

李莎 陶红 高尚

江苏科技大学计算机科学与工程学院

Author(s):

LI Sha; TAO Hong; GAO Shang

School of Computer Science and Engineering,Jiangsu University of Science and Technology

关键词:

支持向量机; 关联分析; 最大最小距离; 粒子群; 故障诊断

Keywords:

support vector machine;  correlation analysis;  max-min distance;  particle swarm optimization;  fault diagnosis

分类号:

TP39

文献标志码:

A

摘要:

应用数据挖掘的方法从实时数据库中提取相应的故障诊断知识是一种有效途径,也是很有现实意义和研究价值的问题。为提高汽轮机组故障诊断的效率,并考虑其计算成本和复杂性,把关联分析作为数据的前处理器,通过计算属性间的相关系数,结合最大最小聚类方法,删除冗余属性。然后采用支持向量机进行故障诊断,构造SVM多分类器,采用粒子群优化算法对参数寻优并训练样本。并与BP神经网络和线性判别分析做比较,实验表明此故障诊断方法诊断速度快、准确率高,可以很好地应用于设备故障诊断

Abstract:

Applying data mining methods to extract the appropriate fault diagnosis knowledge from the real-time database is an effective way,also is an practical significance and research value problem.In order to raise the efficiency of fault diagnosis of steam turbine units and consider its costs and complexity,use the correlation analysis as data pre-processor.Calculate the correlation coefficients between attributes,and combine with max-min distance,then keep only one of the attributes which most highly correlates.Then construct support vector machine classifier,applying particle swarm optimization to find optimal parameter.Experimental results show that SVM outperforms linear discriminant analysis（LDA） and back-propagation neural networks（BPN） in classification performance and can be well applied in fault diagnosis

备注/Memo

备注/Memo:

人工智能四川重点实验室开放课题（2009RY001）李莎（1986-），女，江苏泰州人，硕士研究生，研究方向为数据挖掘、模式识别与智能系统；高尚，副教授，硕士生导师，主要从事系统理论及智能计算方面的研究

常用功能

更新日期/Last Update: 1900-01-01