摘要
将支持向量机应用到内燃机车柴油机燃油系统故障诊断中,将故障诊断问题转换为故障分类问题,成功地实现了小样本情况下的数据识别。以昆明机务段在修程期间记录的故障数据作为样本,采用1-a-1多类分类器对DF4B型内燃机车燃油系统部分故障信息进行了学习。计算结果显示采用支持向量机的识别方法,故障判断平均准确度可以达到99.2%。
The support vector machine is applied to fault diagnosis of fuel system for diesel locomotive. So the problem of fault diagnosis is transferred into problem of fault classification, and data identification via small samples is realized. Taken the fault data recorded during maintenance in Kun-ming depot locomotive as the sample, part of the fault data of the fuel system for DF4B diesel locomotive are studied, by use of 1-a-1 classification machine. The calculation results show that with the identification method of support vector machine, the average accuracy of the fault diagnosis could reach 99.2%.
出处
《机车电传动》
2008年第2期63-65,80,共4页
Electric Drive for Locomotives
基金
中国北车集团公司科研开发项目(2006NG039)
关键词
内燃机车
燃油系统
故障分类
支持向量机
小样本
diesel locomotive
fuel system
fault classification
support vector machine
small sample
