基于复杂网络理论的符号有向图(SDG)化工故障诊断  

Research on chemical process signed directed graph(SDG) fault diagnosis based on complex network

作　　者：王政[1] 孙锦程[1] 王迎春[1] 姜英[1] 贾小平[2] 王芳[2] 

机构地区：[1]青岛科技大学化工学院,山东青岛266042 [2]青岛科技大学环境与安全工程学院,山东青岛266042

出　　处：《化工进展》2016年第5期1344-1352,共9页Chemical Industry and Engineering Progress

基　　金：国家自然科学基金(21136003;41101570);山东省自然科学基金(ZR2009BL021)项目

摘　　要：化工过程系统的大型化和复杂性,仅通过常规方式来描述故障机理越来越受到限制。本文以流程图建模法构建的符号有向图(signed directed graph,SDG)故障模型为基础,将化工过程系统抽象为网络拓扑结构,通过对网络模型的统计特征描述,判断网络的复杂性、小世界性和无标度性,进而以复杂网络中心性理论定量计算网络中各个节点的重要性,分析比较各指标来确定网络中的核心节点,并通过Capocci算法对网络进行社团结构的定量划分,最后以网络中的核心节点确定化工过程中易引起安全事故的关键变量,并用社团划分的结果绘制出化工故障诊断模型的关键路径,确定重点监测部位。案例应用结果表明:该方法可行,为化工过程系统中故障节点和监测提供了新的解决思路,丰富了化工过程故障诊断和预防控制的相关理论。Chemical process systems are large scale and complex. It was limited to describe the fault only through conventional model. This paper based on the SDG（signed directed graph）fault model which constructed with flow modeling method and made the chemical process system abstract the network topology. We described the network model statistical characteristics and judged the network characteristic of complexity, small world and scale-free. The centrality theory was used to quantitatively calculate the importance of each node in the network. Then, those indexes compared to determine the core node in the network, as well as the community with the Capocci algorithm for network structure quantitatively. Finally, we used the core node of the network to determine the chemical process key variables that easy to cause safety accidents. The result of the community partition maps out the protected path for the chemical fault model. The key monitoring area was identified. Results showed that the proposed method can find the fault nodes of chemical process system and the place that need to be monitored and controlled,which can then be used as a support in fault diagnosis and preventive controls.

关 键 词：化工流程 SDG 复杂网络 故障诊断 

分 类 号：X92[环境科学与工程—安全科学]