基于云边交互机制的铁路致灾机理及防灾预警理论

The high-speed railway has become the key infrastructure for the implementation of China's “The Belt and Road” strategy. The safety and reliability are the cornerstones to ensure the sustainable and stable development of the railway system. However, quite a lot of accidents have happened owing to the complicated environment including the complex terrain and variable weather conditions. Besides, the complete mechanism theory of disaster prevention monitoring and forewarning was missing in existing studies, and hence it is urgent to improve the accuracy of disaster detection and ensure the real-time disaster forewarning. First, this project studied the spatial and temporal distribution and the evolution mechanism of railway disaster-inducing factors from the mechanism perspective, and also analyzed its influence on the railway safety conditions to support the study of the disaster prevention monitoring and forewarning methods. Second, the optimization model for the network resources deployment was established to improve the sensing accuracy of disaster-inducing factors and optimize the allocation of computing resources. On this basis, the methods of capacity retention for disaster monitoring and forewarning were formed. Finally, the dynamic allocation model for edge-cloud computing resources was proposed based on multi-agent and game theory considering the dynamic requirements and variable monitoring objects, to build the enhancement mechanism of disaster monitoring and forewarning via the interaction of edge and cloud. The methods of disaster prevention monitoring and forewarning were transformed from the traditional “perceiving on the edge and decision-making on the cloud” to the “synergistic perceiving and decision-making via the edge-cloud collaboration”. The results of this project will further develop the theoretical system of railway disaster prevention monitoring and forewarning and improve the monitoring efficiency of disaster prevention to promote the safe development of China's railway industry.

高速铁路是我国“一带一路”等战略实施的关键性基础设施，安全性和可靠性是铁路系统持久稳定发展的基石。然而，地形地势复杂、气象特性多变的运行环境导致铁路安全事故频发，但现有防灾监测预警方法缺乏完备的机理理论支撑，致灾要素感知精确性和灾害预警时效性亟待提升。首先，本项目从机理探究角度，研究铁路致灾要素时空分布和演化机理，及其与铁路安全态势的映射关系，形成防灾监测预警方法的支撑理论；然后，建立防灾监测网络资源部署优化模型，实现致灾因素精准感知和计算资源部署优化，形成防灾监测预警能力保持理论与方法；最后，针对监测对象和需求的动态性，研究基于多智能体博弈的边云计算资源动态配置优化模型，形成基于云边交互机制的防灾监测预警能力强化机制，实现防灾监测从传统“边缘感知-云端决策”向“边云协同感知决策”转变。本项目成果将进一步完善铁路防灾监测预警理论体系，提升防灾监测预警效能，助力我国铁路行业安全发展。

目前对决定铁路防灾监测预警能力的机理理论研究尚不成熟，亟需从致灾要素时空分布演化机理及其与铁路运行态势相互作用关系研究出发，以防灾感知、计算资源优化部署和调度为方法，进行铁路防灾监测预警能力生成、保持和强化理论研究。因此，本项目主要分为以下四点内容：.首先，依据致灾要素与铁路灾害的影响进行等级划分，并建立铁路网络致灾要素风险评价模型，形成致灾要素风险时空分布特征。结合时空分布特征以及各地的地理环境特征，分析致灾要素所引发的次生灾害分布特征。.其次，将路网与列车运行时刻表耦合，建立铁路网络运行模型，建立铁路系统局部风险演化机理。并将该模型与致灾要素演化机理相结合，对铁路系统面临的不同等级、不同类型的灾害风险全局运行态势进行评估。.然后，进行监测系统架构的搭建，构建系统需求体系，在需求引领下对现有系统不足进行分析，保障架构设计的合理性和实用性；其次，对系统功能进行分类合并，构建系统功能架构；之后，研究面向不同业务需求，构建系统逻辑架构；最后，在系统需求引领下，完成从系统功能到系统物理实体的映射。并且在监测系统架构的支撑下，完成整体监测系统的搭建，完成节点位置和资源部署。.最后，以监测系统的演化机理为支撑，提出基于边缘云协同方案的铁路监控系统有效延迟与能耗最小化模型，通过在复杂铁路监控场景下方案进行比较，验证所提出的模型对系统整体性能的优化。.综上所述，本项目剖析各致灾要素对铁路局部风险和全局运行态势的映射关系，形成复杂系统动态演化模型；并以此为理论基础，指导建立边云协同防灾监测网络资源部署和动态配置优化模型，保持和强化铁路防灾监测预警能力，支撑新一代基于云边交互机制的防灾监测预警理论体系构建。

内容获取失败，请点击重试