writer：Kong, De-peng, Shou-xiang Lu, Quan-sheng Kang, Siu-ming Lo, and Qi-miao Xie.
keywords：Building fire, Risk assessment, Fuzzy set, Event tree, Uncertainty
source：期刊
specific source：Fire Technology no. 50 (4):977-991.
Issue time：2011年
Many uncertainties are introduced inevitably during estimating fire risk for life safety in buildings. Probabilistic methods, the most widely used method to deal with uncertainties, require a large number of historical data. Due to the uniqueness of building fires, such data are rarely available. In order to address this limitation, a fuzzy risk assessment method for life safety under building fires is presented. Event tree is constructed to analyze potential fire scenarios arisen from the failure of fire protection systems. According to the definition of fire risk for life safety, occurrence likelihoods and expected casualty numbers of fire scenarios are estimated with consideration of some uncertainties. Due to limited statistical data and poor knowledge about fire dynamic and evacuation, fuzzy numbers are employed to describe these uncertainties. Based on calculated occurrence likelihood and expected casualty number for a fire scenario, the risk of life safety for a fire scenario is obtained as a fuzzy number. A practical case study for a hypothetical one storey commercial building is carried out with the assessment method presented in this paper and compared with the conventional probabilistic assessment results.