基于故障树和贝叶斯的高层住宅火灾概率计算

Calculating high-rise residential fire probability based on fault tree and Bayesian Network

结合北京市某高层住宅小区火灾构建故障树,合理建立各事件之间的逻辑关系,并将故障树转化为贝叶斯网络模型,再对其中的起火部分进行优化,通过贝叶斯网络的推理计算出顶上事件发生的概率和中间事件、基本事件的后验概率。计算得到的火灾发生概率1.46×10^-4起/户,与国内住宅火灾平均发生概率较为接近。

Considering a high-rise residential community in Beijing, the fault tree was built, and the logical relationship between the events were arranged reasonably. The Fault tree was transformed into a Bayesian network model, and then the fire outbreak part was optimized. The probability of the top event as well as posterior probability of intermediate events and basic events was calculated by the Bayesian network. The fire probability calculated is1.46×10^-4,which is closed to the average probability of residential fire in China.