Research And Application Of Fire Probability In High-rise Civil Buildings

The accelerating process of urbanization in China has led to plentiful consequences including larger urban population,scarcer land resources and tenser urban land.The high floor area ratio of high-rise buildings effectively alleviates the problem of land shortage,but meanwhile the building structure and population capacity of high-rise buildings also bring unavoidable potential safety menace.Therefore,predicting the probability of fire in high-rise buildings and revealing the factors that affect the probability of fire in high-rise buildings are of great significance for the establishment of modern smart cities and the improvement of disaster prevention and mitigation capabilities.This paper takes high-rise buildings as the research object,predicts the fire probability of high-rise buildings based on Bayesian network and maximum likelihood estimation,so as to quantitatively evaluate the fire risk of high-rise buildings and reduce the occurrence of fire accidents.The main research work of this paper is as follows:(1)A statistical analysis of fire accidents in high-rise buildings at home and abroad was carried out,and a focused analysis was carried out on the hazard factors,building types,fire causes and severity of the accident,which laid the foundation for the subsequent research on the fire risk factors of high-rise buildings.(2)Based on the basic concept of Bayesian network,the modeling process of Bayesian network and the main functions and characteristics of each process are analyzed.According to the fire triangle theory and the characteristics of accidents,a high-rise building fire risk factor map is established,which is composed of 3 first-level risk factors,6 second-level risk factors,and 15 third-level risk factors,which provides the basis for the establishment of the Bayesian network model structure of high-rise building fire.In addition,the estimation process of the prior probability and conditional probability of Bayesian network is introduced,which provides a basis for the parameter learning of Bayesian network model of high-rise building fire.(3)Based on the maximum likelihood estimation theory,a set of building fire probability risk assessment model parameter estimation framework is established,and the mathematical model of fire probability and building fire distribution are summarized,and then the parameters of the British probability risk assessment model are estimated with the help of Poisson distribution.(4)The case analysis is divided into two parts.The first part is based on the analysis of high-rise office building fires in a city in the past 10 years,using Ge NIE software to build a Bayesian network model of high-rise building fires,and estimating the fire probability of high-rise buildings in this city based on Bayesian network causal reasoning,and studied the influence of various risk factors on fire probability and verified the rationality of the model based on Bayesian network diagnostic reasoning and sensitivity analysis of various risk factors.The second part is based on the fire and floor area of high-rise office buildings in the past five years in the city,using Matlab software to program the maximum likelihood estimation process between the fire probability and floor area of the city,and study the relationship between the fire probability and floor area of high-rise buildings in the city.