储运过程管道堵塞故障小样本模式识别方法  

Pattern recognition method of blockage in pipeline fault small sample for storage and transportation process

作　　者：陈扶明[1] 税爱社[1] 李生林[1] 卿宇搏 孙耀磊[1] 

机构地区：[1]后勤工程学院后勤信息与军事物流工程系 [2]76174部队

出　　处：《计算机应用研究》2014年第7期2031-2034,共4页Application Research of Computers

基　　金：总后勤部军需物资油料部项目(油20080208)

摘　　要：针对储运过程典型的管道堵塞故障样本数量不足、诊断知识不完备的实际情况,提出储运过程管道堵塞故障小样本模式识别方法;以支持向量机(SVM)算法为基础,构建了储运过程故障模式识别框架,分析了基于SVM方法的故障模式分类;利用储运过程仿真模型采集正常及故障过程数据,建立SVM多分类模式识别器,通过识别器核函数选取和参数优化,进行了管道堵塞故障小样本模式识别的仿真应用研究。仿真实验结果表明,SVM方法能够准确、快速实现故障模式识别,为储运过程故障模式识别提供了一种实用方法。Aiming at blockage in pipeline fault insufficient sample number and diagnosis knowledge in storage and transportation process,this paper proposed a pattern recognition method of small sample. This paper built the fault pattern recognition framework and analyzed the fault pattern classification based on support vector machine（SVM） algorithm. It used the simulation model collect normal and failure process data,to establish the SVM classification pattern recognizer. Through the kernel function of recognizer selection and parameter optimization,it carried a simulation application research of small sample pattern recognition. The simulation results show that the SVM method can accurately and quickly realize fault pattern recognition,and provide a practical method for the storage and transportation process fault pattern recognition.

关 键 词：储运过程 模式识别 管道堵塞 支持向量机 参数优化 智能算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]