SDG故障诊断模型的检验和验证  

Verification and validation of SDG fault diagnosis model

作　　者：吴重光[1] 张卫华[1] 夏迎春[1] 纳永良[1] 王春利[2] 姜巍巍[2] 李传坤[3] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029 [2]中国石油化工股份有限公司青岛安全工程研究院,山东青岛266071 [3]青岛科技大学化工学院,山东青岛266042

出　　处：《北京化工大学学报（自然科学版）》2009年第6期105-110,共6页Journal of Beijing University of Chemical Technology(Natural Science Edition)

基　　金：国家"863"计划(2003AA412310)

摘　　要：将定量模型检验和验证的思想应用于符号有向图(SG或SDG)定性故障诊断模型,以使其达到较高的完备性和分辨率。基于SDG-HAZOP平台,提出了针对SDG定性模型检验和验证的非正式检验、静态检查、动态试验和诊断结果半定量风险分析的4个关键步骤,给出了静态检查中的操作点非正常原因可达性检验和不利后果可达性检验步骤的具体内容。以加热炉SDG故障诊断模型为例,验证了上述步骤的有效性。In order to achieve a better completeness and identification performance, the verification and validation （V＆ V） technique for qualitative models is applied to the signed graph （SG or SDG） quantitative fault diagnosis model. Based on the SDG-HAZOP platform, four key V＆V steps are proposed： informal verification, static check, dynamic test, and semi-quantitative risk analysis of diagnosis results. Two steps in the static check procedure, operational point teachability verification of abnormal reason and adverse consequence, are described in detail. Finally, application of these two steps is demonstrated using a heating furnace as an example.

关 键 词：符号有向图 检验 验证 模型 故障诊断 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]