基于SDG故障诊断的传感器分布优化设计  

Design of Sensor Network Based on Signed Directed Graph Fault Diagnosis

作　　者：王强[1] 吴重光[1] 张贝克[1] 魏环[1] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029

出　　处：《计算机仿真》2006年第7期308-312,共5页Computer Simulation

基　　金：国家高技术研究发展计划(863)项目(2003AA412310);北京市教委资助项目(XK100100435)

摘　　要：故障诊断是化工企业安全生产中一项重要的工作,其诊断方法的效率主要取决于监视过程变量传感器的配置。现有的双向图的设计方法配置过程繁琐,容易出错。针对这个问题,提出一种改进的基于SDG故障诊断的计算机程序算法来设计传感器的网络分布。通过对一个化工实例的仿真研究表明,这种改进的计算机程序算法具有效率高、信息利用量大、准确率高和简单易用的特点,为将基于SDG的故障诊断的方法应用于实际控制过程提供了一种新的途径。Fault diagnosis is an important task for the chemical process, and the efficiency of the diagnostic appreach depends critically on the location of. sensors monitoring the process variables. Some algorithms based on directed graph （DG） and signed directed graph （SDG） were discussed in previous work. These algorithms are based on a bipartite graph to locate the network of sensors. In this article, the sensor location work is extended to use the program based on SDG. Algorithms for sensor network design based on the computer are detailed and applied to a chemical engineering case study. It is shown that better design can be obtained by using the program based on SDG instead of the bipartite graph .

关 键 词：带符号的有向图 故障诊断 传感器 程序 

分 类 号：TP274.5[自动化与计算机技术—检测技术与自动化装置]