基于故障树和贝叶斯网络的建筑施工火灾风险评价

针对建筑施工现场频繁发生的火灾事故，将这一事故作为顶事件并分析导致顶事件发生的基本事件，构建故障树，表示出各基本事件与顶事件之间的逻辑关系。再将故障树转化成贝叶斯网络模型，通过贝叶斯网络的推理能力计算出顶事件发生的概率和各基本事件的后验概率。计算出各个基本事件的结构重要度、概率重要度、关键重要度，分别对三类重要度进行排序，定量地评估建筑施工火灾风险发生的可能性，找出最薄弱的环节，有针对性地指导现场施工。并引用案例数据计算施工过程中火灾事故的风险值，证明了将故障树与贝叶斯网络模型结合分析建筑施工安全风险这一方法的有效性和简便性，找出最能影响顶事件发生的风险因素，以及最容易降低概率的风险因素来更好地指导施工，将火灾风险降低到最小。

Fire Risk Assessment of Building Construction Based on FaultTree and Bayesian Network

For fire accidents occurring frequently at the construction site，we identify a fire event as a top event and then analyzethe basic events that cause the top event. A fault tree is established to show the logical relationship between the basic event and thetop event. Then, the Fault Tree will be transformed into Bayesian network，which will be used to calculate the occurrence probabilityof top event and the posterior probability of each basic event by using Bayesian network reasoning ability. This study used Bayesiannetwork to calculate the structural importance，probability importance and critical importance of each event to quantitatively assessthe fire risk on construction site and find the weakness of the accident. The result is significant to the onsite construction safety.Finally，a case study is used to verify the proposed model to prove its effectiveness and simplicity. Results show that the risk factorsthat can significantly affect the occurrence of the top event，and the measures are most likely to minimize the risk of fire.