基于复杂网络的化工过程状态监测方法  

Condition monitoring method of chemical process based on complex networks

作　　者：崔超宇 郭丽杰[1,2] 康建新[1,2] CUI Chaoyu;GUO Lijie;KANG Jianxin(Hebei Key Laboratory of Applied Chemistry,Yanshan University,Qinhuangdao,Hebei 066004,China;School of Environmental and Chemical Engineering,Yanshan University,Qinhuangdao,Hebei 066004,China)

机构地区：[1]燕山大学河北省应用化学重点实验室,河北秦皇岛066004 [2]燕山大学环境与化学工程学院,河北秦皇岛066004

出　　处：《燕山大学学报》2021年第1期33-41,共9页Journal of Yanshan University

基　　金：国家自然科学基金资助项目(21606193)。

摘　　要：为了有效地从化工过程监测数据中提取故障特征,及时准确地辨识出故障,提出了基于复杂网络的化工过程状态监测方法。首先依据复杂网络平均节点度和平均聚集系数的稳定性确定建立复杂网络所需的最佳样本数量,以保证所建复杂网络能刻画系统中各个节点之间相互作用的真实状态;然后利用复杂网络的拓扑结构,构建了复杂网络的节点异常系数和其阈值计算方法,用以提取实时工况特征来实现定量化状态监测;最后以TE过程为应用实例,结果表明,本文所提出的方法能够以图形化的方式快速、有效地检测出化工过程故障,可为操作和管理人员提供可靠的决策依据,预防事故的发生。In order to effectively extract fault characteristics from the data in the chemical process and identify faults in a timely manner,a condition monitoring method of chemical process monitoring based on complex networks is proposed.In this method,based on the stability of average node degree and average aggregation coefficient of complex networks,the optimal numbers of sample for complex networks is first determined so that the complex networks established can describe the real state of interaction between nodes in the complex system.Then,by using the topology of the complex networks,the node anomaly coefficient of the complex networks and its threshold are constructed in extracting the real-time operating characteristics and carrying out quantitative condition monitoring.Finally,the TE(Tennessee-Eastman)process is used in a case study.The results show that the proposed method can quickly and effectively detect chemical process faults in a graphical way,which focuses on providing a reliable decision-making basis for operators and managers to prevent accidents.

关 键 词：复杂网络 化工过程 状态监测 节点异常系数 

分 类 号：X937[环境科学与工程—安全科学]