摘要
为了准确评价天然气管道事故后果严重程度,提高管道安全管理水平,以属性区间识别理论为基础,建立了天然气管道事故后果综合评价的属性区间识别模型;同时,为解决属性区间识别模型中的权重确定问题,在收集天然气管道历史事故数据的基础上,建立投影寻踪-遗传算法模型,从历史事故数据中获得客观权重;最后依据置信度准则对天然气管道事故后果进行属性识别,实现天然气管道事故后果严重程度的等级划分。应用所建立的模型对近年来发生的13起天然气管道事故进行实例验证,结果表明,属性区间识别模型操作简单,结果合理,具有一定的参考价值,为天然气管道事故后果评价提供了一种新的方法,进一步可为天然气管道安全管理和应急救灾措施的制定提供技术支持。
In order to evaluate the severity of the natural gas pipeline accident scientifically and reasonably,as well as improve the level of the safety management of natural gas pipeline,the paper has established an attribute interval recognition model for the comprehensive evaluation of natural gas pipeline accident consequence based on the attribute interval recognition theory. The model includes 3 evaluation indexes: the number of death,the number of injured and the direct economic loss. The natural gas pipeline accident is usually divided into 4 grades,whereas the severity of the accident consequence tends to increase gradually from grade 1 to grade 4. In order to solve the problem of subjectivity in the process of determining the indexes weight of attribute interval recognition model,it is necessary to establish a projection pursuit genetic algorithm model based on a large number of historical related data of natural gas pipeline accidents,whose objective weight used to be found from the historical accident data. And,finally,according to the confidence criterion,the attribute recognition of the accident consequence of natural gas pipeline can be carried out to realize the consequence of the natural gas pipeline classification of accident severity. In order to verify the feasibility of the attribute interval recognition model,we have analyzed 13 typical natural gas pipeline accidents in recent years by classifying the model type and the grades of the consequence of the said accidents. The analysis results show that the evaluation results of the aforementioned accidents are in line with the human subjective cognition for the natural gas pipeline accidents. And,so,the attribute interval identification model is easy to operate with the evaluation results gained being reasonable. The said model is used to be of their own application value in the safety management,which can provide for the accident consequence assessment with a new method of natural gas pipeline,so that the method can provide a certain scientific and technical support for the safety management of the natural gas pipeline and formulation of the emergency relief measures.
作者
李玉星
杨昊
王武昌
刘鹏
黄维和
LI Yu-xing;YANG Hao;WANG Wu-chang;LIU Peng;HUANG Wei-he(Shandong Key Laboratory for Oil-Gas Storage and Transportation Safety,Qingdao 266580,Shandong,China;CNPC LNG and Heavy Gas Storage and Transportation Technology Laboratory,Qingdao 266580,Shandong,China;PetroChina Co.Ltd.,Beijing 100007,China)
出处
《安全与环境学报》
CAS
CSCD
北大核心
2021年第3期943-949,共7页
Journal of Safety and Environment
基金
国家重点研发计划课题(2016YFC0802104)。
关键词
安全工程
天然气管道
事故评价
属性区间识别
投影寻踪模型
遗传算法
safety engineering
natural gas pipeline
accident evaluation
attribute interval recognition
projection pursuit model
genetic algorithm
