摘要
为评估建筑火灾动态风险,从防火工程的角度将火灾发展过程划分为"火情-火警-火险-火灾"4个阶段,分别研究了不同阶段的主要风险评估参数。采用Bayes网络方法构建了动态风险评估模型,确定了网络结构与参数。采用敏感度分析法研究了评估参数对火灾风险的影响程度。以2座典型建筑为例,分别计算得到每个阶段风险和综合风险。研究结果表明:建筑火灾风险是一个动态变化的过程,各阶段风险、评估参数均存在差异;评估节点和依赖关系构成了因果网;评估模型可以有效地将消防监测终端采集的消防大数据与人工智能分析技术相结合,有助于提升建筑消防安全管理的智能化水平。
The development of building fires was divided into four stages for risk assessment as fire initiation, fire alarm, fire behavior, and fire spreading based on fire engineering theory with analyses of the main risk assessment parameters of each stage. The dynamic risk assessment model was based on a Bayesian network. A sensitivity analysis was then used to evaluate the influences of key parameters on the fire risk. Two typical buildings were then used as examples to evaluate the risk at each fire stage and the overall risk. The results illustrate how the building fire risk is a dynamic process with different risk and impact parameters in each stage. The model nodes and dependencies constitute a causal network. The evaluation model can effectively combine large amounts of fire data collected by a building fire monitoring terminal using artificial intelligence analyses. This research can effectively improve building fire safety management.
作者
疏学明
颜峻
胡俊
吴津津
邓博誉
SHU Xueming;YAN Jun;HU Jun;WU Jinjin;DENG Boyu(Department of Engineering Physics,Institute of Public Safety Research,Tsinghua University,Beijing 100084,China;China Institute of Industrial Relations,Institute of Safety Engineering,Beijing 100048,China;Beijing Key Laboratory of City Integrated Emergency Response Science,Beijing 100084,China)
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2020年第4期321-327,共7页
Journal of Tsinghua University(Science and Technology)
基金
国家重点研发计划项目(2017YFC0806600)
国家自然科学基金资助项目(71774094,71790613)
中国劳动关系研究生教育教学改革项目(YJG1702)。
