时空受限条件下海事事故协同应急决策方法研究

Recently, major accidents occur occasionally worldwide, in order to reduce the severity of consequence, it is necessary to take quick and effective response to maritime accidents. From the literature review, the majority of the researches focus on the end state (i.e. search and rescue and oil spill response) of maritime accidents, while few research focus on initiating event and mid states, such as collision and grounding. Moreover, the research on emergency decision-making, from the perspective of restrictions on spatial-temporal accessibility and cooperative among multiple organizations, is also lack from the literature. Therefore, this project will intend to meet the requirements of navigation safety, focus on the emergency response to typical maritime accidents, and conduct systematic research on the key scientific problem on “cooperative emergency decision-making for maritime accidents under restricted spatial-temporal accessibility”. In order to achieve this objective, this project will manage to solve three sub-problems, which are spatial-temporal accessibility analysis of emergency resources under uncertain restrictions, cooperative decision-making method for maritime accidents under restricted conditions, and optimized resource allocation method in coupling of risk, time and cost. Afterwards, relationships among emergency resources, maritime accidents, and restrictions will be discovered, and the methods for cooperation decision-making and dynamic resources scheduling in multiple states for multiple tasks will be proposed. Finally, these two models will be applied to the typical accidents (e.g. collision, grounding and not under control) in typical waterway (Yangtze River and South China Sea). The simulation experiments will be carried out in the maritime emergency simulation platform, and these experiments will be used to verify that the proposed models can achieve quick and effective selection of response actions and optimized resources allocation according to the dynamic requirements of multiple tasks in the timeline of accident development. The outputs of this project can provide the theory and technology for emergency response to maritime accidents.

近年来重特大航运事故时有发生，对事故进行快速有效的处置以减少事故损失十分必要。当前应急研究对象侧重于最终事件（目标搜寻和溢油处置），对初始事件（如碰撞和搁浅）及其演化过程的研究较少。此外，从资源时空可达性限制和多部门协同的角度开展研究也较少。鉴于此，本项目面向船舶航行安全需求，以事故应急决策为对象，紧密围绕“时空受限条件下海事事故协同应急决策方法”这一关键科学问题，力争突破“不确定约束条件下应急资源时空可达性分析、限制条件下海事事故快速协同应急处置决策方法和风险-时间-费用耦合的应急资源动态调度决策方法”这三个难点，阐明资源、事故和约束的映射关系，提出协同应急处置决策方法和多阶段多任务资源调度方法，将其应用于典型水域（长江和南海）的典型事故（碰撞、搁浅和失控等），实现最优处置方案选择及有序合理的应急资源调度，并在海事应急仿真平台开展实验验证。研究成果将为事故应急处置提供理论基础和技术支撑。

航运承担大量贸易运输，苏伊士运河搁浅事故造成经济损失约570亿美元，究其原因为船员操纵失误，因此，如何提升船员的风险评估与应急能力是需要解决的关键问题。本项目以长江为对象，解决了动态瞬变下的船舶航行风险评估与辅助决策问题，形成了三个创新成果：开展了水位-季节-空间耦合下事故前兆航行条件辨识，建立了小样本数据下的事故前兆航行条件网络模型，量化了各影响因素对船舶航行事故的影响，实现了事故前兆的航行条件辨识，辨识准确率约92%；融合船舶操纵特性的航行风险及时预警方法，建立了考虑港航工程设计尺度约束和船舶操纵特性的航行态势认知模型，提出了船舶航行风险及时预警方法，相比轨迹数据挖掘可降低虚警率约25%；构建了方案-准则-因素的航行事故应急决策框架，提出了动态发展过程中情景风险效用值的计算方法（情景+后果+概率），引入TOPSIS解决了事故发展轨迹动态造成不完整信息，实现了紧急情情景下的辅助决策，可降低约35%的决策失误概率。研究成果如下：在国内外学术期刊或会议发表论文26篇，其中高被引论文4篇，SCI论文12篇；授权国家发明专利4项，软件著作权5项；共培养3名硕士研究生；项目负责人获得湖北省自然科学二等奖，并入选国家高层次人才特殊支持计划青年拔尖人才。本项目的主要研究成果应用于全任务海事应急处置演练系统的研发，该系统是国际上第一个面向海事监管人员的仿真演练系统，项目于2020年成功获批深圳市科创委成果转化类项目，获批经费270万元。此外，项目成果还成功应用于广东海事局电子沙盘系统的研发，为船舶通航安全保障提供了解决方案。