摘要
工业管道因失效导致泄漏往往会造成严重的后果,分析其失效原因并合理预估风险等级对管道安全服役显得尤为重要。故障树是一种分析系统失效原因和评估安全风险的常规方法,但由于管道系统的复杂性,难以获得故障树分析所需各类事件发生的精准概率,故无法实现定量评估。针对此问题,该文将三角直觉模糊数引入故障树,并结合打分法,提出一种评估工业管道系统风险等级的方法。该方法采用故障树分析管道失效原因并确定风险打分项,针对无法获得客观数据的打分项,运用三角直觉模糊数收集整合专家评价意见并转换为具体评分,最终依据打分法规则计算管道失效分数和风险等级。运用该方法对某天然气净化厂多条工业管道进行分析,验证了方法的可行性和适用性。
The material leakage caused by the failure of industrial pipelines usually causes serious consequences,so it is very important to analyze the failure reasons and reasonably estimate the risk level for the safety of pipelines.Fault tree analysis is a common method to analyze the cause of system failure and evaluate safety risk.However,due to the complex characteristics of the pipeline system,it is difficult to obtain the precise probability of the occurrence of various events used in fault tree analysis,and quantitative evaluation cannot be realized.To solve this problem,this paper introduces the triangular intuitionistic fuzzy number into the fault tree,and combined with the scoring method,proposes a method to evaluate the risk level of industrial piping system.In this method,failure trees are used to analyze the failure causes of pipelines and determine the risk score items.For the scoring items that cannot obtain objective data,triangular intuitionistic fuzzy numbers are used to collect and integrate expert evaluation opinions and convert them into specific scores.Finally,pipeline failure scores and risk levels are calculated according to the scoring regulations.This method is used to analyze several industrial pipelines in a natural gas purification plant,and the feasibility and applicability of the method are verified.
作者
都心爽
武玮
张金阳
淡勇
DU Xinshuang;WU Wei;ZHANG Jinyang;DAN Yong(School of Chemical Engineering,Northwest University,Xi′an 710069,China)
出处
《西北大学学报（自然科学版）》
CAS
CSCD
北大核心
2021年第4期632-642,共11页
Journal of Northwest University（Natural Science Edition）
基金
国家自然科学基金资助项目(21576224)
陕西省自然科学基础研究计划(2020JM-436)。
关键词
工业管道
风险等级
故障树
直觉模糊集
industrial pipelines
risk level
fault tree
triangular intuitionistic fuzzy number
