基于云理论的水库大坝应急临机决策研究

The emergency provisional decision making of the dam has salient features: high degree of uncertainty, dynamic, and timeliness and so onin the disaster environment. Cloud theory can be used to fully integrate uncertainty, randomness, vagueness of the mass information. The program is aimed at characteristic parameters of disaster-causing factors and the uncertainty of dam effect size the factors effect, and uses the digital characteristics of cloud model to overall characterize the dynamic monitoring information. Besides, with the help of forward cloud and backward cloud generators, it can realize the conversion between qualitative and quantitative multi-source heterogeneous emergency knowledge, such as text knowledge, dynamic monitoring data, expert tacit knowledge, and build a fusion cloud model which matches the dam emergency knowledge. Moreover, exploring the decision-making bias and their mechanism under disaster environment, it do further introduction of psychological pressure, working strength, risk preference and other impact factors. Through the cloud transform, it creates a dynamic emergency decision-making cloud model, and by the changes of cloud gravity center, it can rapidly recognize the disaster characteristics, such as deformation, seepage, stress-strain and achieve a systematic research of the dam emergency decision-making model under disaster environment. As a result, it can reduce or eliminate the systematic errors caused by different mathematical methods during multi-stage decision process and be verified by the case of typical dam disaster events at the same time. This research has important theoretical and practical significance for improving and enriching emergency decision-making theory and methods, and guiding disaster prevention and mitigation work of water conservancy under disaster environment.

灾变环境下水库大坝应急临机决策具有高度复杂性、动态性、时效性等显著特征。云理论能将海量信息的不确定性、随机性、模糊性等集成统一，并加以有效利用。本项目拟针对致灾因子特性参数及大坝效应量的不确定性，采用云模型的数字特征来整体表征动态监测信息；在此基础上，借助正向云和逆向云生成器，实现动态监测数据、专业文本知识、专家隐性知识等多源异构信息的定性与定量间映射转换，构建大坝应急知识匹配的融合云模型；探索灾害情景下决策偏差机制，进一步引入心理压力、工作强度和风险偏好等消缺因子，通过云变换，建立应急临机决策云模型，利用云重心的变化快速辨识变形、渗流、应力应变等灾变特征，实现灾变环境下水库大坝应急临机决策系统性研究，降低决策过程中多重数学方法所造成的系统误差，并以典型水库大坝灾害事件案例加以验证。该项研究对完善和丰富应急决策的理论和方法，指导突发事件下水利工程的防灾减灾工作，具有重要的理论和现实意义。

灾变环境下水库大坝应急临机决策具有高度复杂性、动态性、时效性等显著特征。云理论能将海量信息的不确定性、随机性、模糊性等集成统一，并加以有效利用。本项目拟针对致灾因子特性参数及大坝效应量的不确定性，采用云模型的数字特征来整体表征动态监测信息；在此基础上，实现动态监测数据、专业文本知识、专家隐性知识等多源异构信息的定性与定量间映射转换，构建大坝应急知识匹配的融合云模型；研究了水库大坝应急临机决策影响因素，并建立了临机决策的测度指标体系，针对灾害情景下组织机构、救援能力、群众避险能力，并引入决策者的工作强度、心理压力以及风险偏好等消缺因子，利用云重心的变化来快速辨识灾变环境下的水库大坝应急临机决策效能评估模型，科学识别大坝灾害警情，提高应急决策效率和准确率。最后利用青海省温泉水库险情应急处置作为案例进行了验证，同时对相关功能进行了云管系统实现。该项研究对完善和丰富应急决策的理论和方法，指导突发事件下水利工程的防灾减灾工作，具有重要的理论和现实意义。