基于事故树与贝叶斯网络的管道泄漏事故溯源方法

发布时间：2018-04-20 04:38

本文选题：油气管道 + 事故树　；参考：《油气储运》2017年09期

【摘要】：随着城市快速发展,通讯、热力、油气、供排水等管道相继建设并投入使用,由于规划、管理或部分历史遗留原因使得各管网在地下纵横交错,形成一张巨大的"蜘蛛网",且无统一有效的管理机构,一旦发生油气管道泄漏,如何快速辨识及事故溯源成为当前安全及应急管理的难点。针对该问题,提出了基于事故树(Fault Tree,FT)与动态贝叶斯网络(Dynamic Bayes Network,DBN)的油气管道事故快速辨识及深度溯源分析方法,充分利用FT建模的简便有效性、DBN反向推理能力及后验概率表达准确性的优势,通过分析城市油气管道泄漏事故可能原因建立相应的事故树,将其转化为DBN模型后融合历史统计数据和多源观测信息进行事故场景快速辨识及溯源分析。分析结果与后期事故调查报告结论一致,表明该方法具有较好的准确性和实用性,能够为事故应急救援、事故调查提供指导。
[Abstract]:With the rapid development of the city, communication, heat, oil and gas, water supply and drainage pipes have been built and put into use one after another. Due to the planning, management or some historical reasons, the pipe network is crisscross in the underground, forming a huge "spider net", and there is no unified effective management organization. Once the oil and gas pipeline leaks, how to quickly identify and matter Therefore, traceability is the difficult point of current security and emergency management. In view of this problem, a fast identification and depth tracing analysis method for oil and gas pipeline accidents based on Fault Tree (FT) and dynamic Bayesian network (Dynamic Bayes Network, DBN) is proposed. The simplicity and effectiveness of FT modeling, the DBN reverse reasoning ability and the posterior probability table are fully utilized. By analyzing the possible cause of the leakage accident in urban oil and gas pipeline, the corresponding accident tree is established and transformed into the DBN model to fuse historical statistics and multi source observation information to identify and trace the accident scene quickly. The analysis results are in agreement with the conclusion of the later accident investigation report, indicating that the method is better. The accuracy and practicability can provide guidance for accident emergency rescue and accident investigation.

【作者单位】：中国石油西部管道公司;中国石油工程建设有限公司华北分公司;
【基金】：国家自然科学基金资助项目“注油过程储油罐内静电动态聚集模式机理研究”,51204196
【分类号】：TE88

【相似文献】