基于贝叶斯网络的地下空间火灾风险评估方法研究

Risk Assessment Method for Fire in Underground Space Based on Bayesian Network

为提高地下空间火灾风险定量评估的准确性,将贝叶斯网络(BN)方法引入火灾风险定量评估过程中。根据地下空间典型火灾场景建立事件树模型,由事件树模型转化得到BN模型。证明2种模型的一致性。并在考虑不确定性因素影响的条件下,修正BN。最后,利用经修正的BN模型,计算火灾发展到不同阶段的概率,以财产损失估计值表示事故的严重程度,计算火灾风险值。结果表明,不确定性因素的存在对风险评估结果影响显著;基于BN的定量评估方法比事件树分析ETA方法能更好地处理不确定性问题,更适合地下空间火灾的风险评估。

In order to improve the accuracy of quantitative assessment of underground space fire risk, the, BN approach was introduced into fire risk assessment process. Firstly, an event tree model was built in accordance with typical fire scene and transformed into a BN model. Moreover, the consistency of two models was proved. Then, the network model was modified in order to consider the effects of uncertain factors. Finally, this modified model was used to calculate the probabilities of fire in different development stage. The loss of property was used to estimate the severity of fire, afterward the fire risk was calculated. The results from example analysis show that effect of uncertain factors is remarkable for risk assessment, and that compared with ETA, the novel method handles uncertainty better and is more suitable for assessing underground space fire risk.