基于贝叶斯网络的地震次生燃气管道泄漏事件链构建

地震次生燃气管道泄漏事件是导致地震灾害影响扩大的主要原因之一，为了解建筑物内地震次生燃气管道泄漏事件的发展过程，参考国内外相关的文献和统计资料，确定建筑物内地震次生燃气管道泄漏事件贝叶斯网络的节点变量和取值范围。根据节点变量及其逻辑关联性构建贝叶斯网络结构图，通过对国内外研究数据的统计并结合专家经验估算确定各节点变量的条件概率。利用贝叶斯网络工具计算建筑物属性和环境变量在不同状态取值下建筑物遭受破坏、燃气泄漏、燃气扩散引发二次灾害等关键节点变量的后验概率。通过实例分析得出，建筑结构和地震烈度是建筑物遭受破坏的主要影响因素，风速、大气稳定度对燃气扩散引发二次灾害有显著影响。

Construction of event chain for secondary gas pipeline leakage induced by earthquake based on Bayesian network

The secondary gas pipeline leakage event induced by earthquake is one of the main reasons causing the expansion of influence by earthquake disaster. To investigate the development process of secondary gas pipeline leakage event induced by earthquake in building, the node variables and valuing range of Bayesian network for secondary gas pipeline leakage event induced by earthquake in building were determined by referring to the related literatures and statistics documents at home and abroad. The structure diagram of Bayesian network was constructed according to the node variables and their logical connection. The conditional probability of each node variable was determined by statistics on research data at home and abroad combined with experts experience estimation, then the posterior probability of key node variables, such as the destroyed building, gas leakage, gas diffusion causing secondary disaster etc., under different state valuing of building properties and environmental variables were calculated by using Bayesian network tool. The results of actual example analysis showed that the structure of building and the intensity of earthquake are the main influencing factors causing the damage of building, and the wind speed and atmospheric stability have a significant impact on the gas diffusion causing secondary disaster.