水电工程施工事故发展的网络行为范式研究

The hydropower construction accidents frequently occur. By establishing the behavior paradigm of accident development, it explores the developmental mechanism of hydropower engineering accident. Based on the characteristics of hydropower construction accidents, this research project systematically considers the insecure factors, such as human, objects and environments, etc. And by adopting ontology, epistemology and methodology methods, it carries out the behavior paradigm research on accident development. By establishing high-risk job evaluation model of multi-attribute segmentation, specifying the risk constitutes and applying Kappa coefficient analysis, it measures the risk factor classification. With the HFACS tools, it establishes the structural framework of the accident development, and then it summarizes the coupling relations between the deduction elements, and reveals accidents formation mechanism. It also abstracts catastrophic path of the accident development, builds the transfer matrix of the emergency conditions, develops rheological strength model of accident derivation and spread, rebuilds the Bayesian network for accident evolution, and confirms the damage intensity of the accident. Aiming at spatiotemporal dynamics of the risk elements of for the hydropower project construction accident, it uses HIDs to portray the sequence relationships of accident's catastrophic path node, hence to build the behavior paradigm of accident development's dual random network model, and develops the BDD efficient algorithm to determine catastrophic event's key nodes and paths. This research project provides theoretical support for hydropower construction accident risk prediction, prevention and control, in the meantime, it also provides new ideas for similar risk factor problems.

水电工程施工事故多发易发，通过构建事故发展的行为范式，探索水电工程事故发生发展机理。 本课题针对水电工程施工事故发展的特点，系统考虑人、物和环境等不安全因素，按本体论、认识论和方法论的递进层次，开展其事故发展的行为范式研究：建立高危作业多属性分割评价模型，明确风险构成，并应用Kappa系数分析法测算风险要素分类；借助HFACS工具建立事故发展的结构框架，统计推演要素间的耦合关系，揭示事故形成机理；抽象出事故发展的灾变路径，构建事故状态转移矩阵，开发事故衍生和扩散的流变强度计算模型，重构事故演化的贝叶斯网络，确定事故致损强度；针对水电工程施工事故风险构成要素时空动态性，运用HIDs刻画事故发生灾变路径节点的序列关系，构建事故发展行为范式的二重权随机网络模型，开发高效BDD算法确定灾变关键节点与路径。 本课题为水电工程施工事故的风险预测与防控提供理论支撑，同时为类似风险因素问题研究提供新思路。

水电工程施工事故多发易发，通过构建事故发展的行为范式，探索水电工程事故发生发展机理。针对水电工程施工事故发展的特点，系统考虑人、物和环境等不安全因素，明确了水电工程施工风险要素的构成与分类，针对几类具体作业类型，相应提出了具体的风险因素体系。借助HFACS方法建立事故发展的结构框架，借助系列统计学方法（Chi-square test、Lambda系数和Tau-y）或现有成熟模型（ISM），阐明了事故风险构成要素耦合关系；根据风险要素的耦合关系，抽象出事故发展的灾变路径，应用隐马尔科夫模型，构建事故状态转移矩阵，揭示事故形成机理及发展趋势。在此基础上，针对水电工程施工事故风险构成要素时空动态性，基于故障树分析法，结合二叉决策图 (BDD, Binary Decision Diagram)的结构与算法，给出了安全事故流变强度与基本事件重要度的计算方法；最后，系统分析水电工程施工应急管理特点，从应急救援方案、应急能力评估、协同应急以及应急执行力培育等方面对应急管理进行深入研究，取得了系列成果。本课题为水电工程施工事故的风险预测与防控提供理论支撑，同时为类似风险因素问题研究提供新思路。