论文部分内容阅读
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。
With the rapid development of the city, pipelines such as communications, heat, oil and gas, water supply and drainage have been built and put into use one after another. As a result of planning, management or part of historical legacy, the pipe networks are staggered in the underground to form a huge “spider web” And there is no uniform and effective management organization. Once the oil and gas pipelines leak, how to quickly identify and trace the accident has become the current safety and emergency management difficulties. Aiming at this problem, a rapid DBN-based identification and depth traceability analysis method for oil and gas pipeline accidents is proposed. The advantages of simple and effective FT modeling, reverse inference of DBN and accuracy of posterior probability representation are put forward. By analyzing the possible causes of urban oil-gas pipeline leakage accidents, a corresponding fault tree is established, which is transformed into a DBN model and integrated with historical statistics and multi-source observation information to quickly identify and trace the accident scene. The analysis result is consistent with the conclusion of the investigation report of the late accident, which shows that the method has good accuracy and practicability and can provide guidance for accident rescue and accident investigation.