城市燃气供应系统Natech事件风险评估研究

Study on risk assessment of Natech events in urban gas supply systems

为提高城市燃气供应系统抗灾能力和韧性水平,基于不同自然灾害诱发的城市燃气供应系统突发事件案例样本统计数据,探究城市燃气供应系统可能面临的典型自然灾害类别。首先,构建不同自然灾害诱发的城市燃气供应系统突发事件链,并转换为事故树模型和贝叶斯模型;然后,采用基本事件发生近似概率公式,对比分析不同模型的计算结果;最后,运用GeNIe 2.0软件逆向推理目标节点概率为100%时其他各节点的后验概率,总结关键节点,确定城市燃气供应系统可能面临的典型自然灾害。研究表明,城市燃气供应系统可能面临的典型自然灾害为暴雨、洪水、滑坡,且因考虑节点条件概率,贝叶斯模型计算结果更符合实际。

To improve the anti-disaster ability of the urban gas supply systems,the damage caused by natural disasters to urban gas systems was studied from the aspect of accident cases.First of all,according to the literature and network research,we collected and sorted out the Natech incident cases of urban gas supply systems at home and abroad in recent years,counted and sorted out the emergency cases of urban gas supply systems caused by different natural disasters,classified the natural disasters in the cases according to the Classification and Coding of Emergencies(GB/T 35561—2017),and determined typical natural disaster types that urban gas supply system might face.Secondly,based on the failure process of urban gas systems caused by different natural disasters,the emergency chain of city gas supply systems caused by different natural disasters was constructed,and it was transformed into a fault tree model and a Bayesian model.Besides,the approximate probability formula of the basic event was put forward,the probability of the occurrence of the event was calculated according to the relevant accident data,and the calculation results of different models were compared.Finally,using GeNIe 2.0 software,the model is reversely reasoned,the key nodes of the Bayesian model were summarized and compared with the key nodes of the fault tree model,the typical natural disasters that urban gas supply systems may face were determined,and the scientific and reasonable preventive measures for typical natural disasters that the city gas system may face were put forward.The research shows that the typical natural disasters that urban gas supply systems may face include rainstorms,floods,and landslides.Moreover,considering the conditional probability of nodes,the results calculated by the Bayesian model are more in line with reality than those calculated by the fault tree model.