基于SDG技术的炼油催化裂化装置智能自解释报警研究  

Intelligent Self-Explanation Alarming Study on FCCU Unit Based on SDG Technology

作　　者：单晓雯[1] 张卫华[2] 吴重光[2] 王春利[1] 

机构地区：[1]中国石油化工股份有限公司安全工程研究院,青岛266071 [2]北京化工大学信息科学与技术学院,北京100029

出　　处：《石油化工自动化》2009年第1期31-36,共6页Automation in Petro-chemical Industry

摘　　要：基于符号定向图(SDG)的故障诊断方法是一种典型的半定量故障诊断方法。该方法是根据系统不同节点之间的影响关系,建立系统模型,一旦系统出现报警,该方法能够及时找出报警原因(根结点)及危险从原因节点传播到报警节点的"剧情",这样有利于提前采取措施预防事故的发生。根据炼油厂催化裂化装置常见流程建立了FCCU的反应再生系统的SDG模型,并在FCCU动态仿真软件平台上进行了自解释报警仿真实验。实验表明SDG的故障诊断方法识别故障的完备性较高,能够揭示故障演变的路径和历程,推理速度快,可以应用于多种动态连续过程系统的在线危险识别、故障分析和诊断的研究。The fault diagnosis method on signed directed graphs （SDG） is a typical semi-quantitative method. In the continuous industrial system, the SDG model is built based on the relationship of different nodes. When the system gave alarm it can find the cause and the propagation path of alarms, and help take action early to prevent accident. The reaction-regeneration system SDG model of normal process FCCU is built and the simulation experiments of self-explanation alarming has been carried out. The experimental results indicate that the fault diagnosis methods on SDG can diagnose faults and give the exact course of the fault evolution. It has very high inference speed and can be used in the research of hazard identification on line and fault analysis in series continuous dynamic processes system.

关 键 词：催化裂化装置 符号定向图 定性推理 自解释报警 

分 类 号：TE967[石油与天然气工程—石油机械设备]