| With the plant-scale construction of urban utility tunnel in China,the safety problems can not be ignored.Among them,the cable compartment has great fire risk.Once the cable compartment fires,it will have a great impact on people’s property safety.The fire risk assessment of the cable compartment is of great significance to prevent the occurrence of fire accidents in the utility tunnel.Therefore,through the dynamic analysis of the fire probability and fire consequences of the cable compartment,this paper provides theoretical and data support for the fire design and management of the utility tunnel.The specific research conclusions are as follows:1.This paper describes the whole process of the fire in the cable compartment of the utility tunnel from the causal factors to the consequences of the fire.Through the hazard identification path in cable fire,the causes of cable fire accidents are analyzed,and the cable compartment fire accident tree model is constructed.Then,the safety barrier limiting the spread of cable fire and the possible fire accident scenarios are analyzed by using the event tree model.Finally,the bow-tie model of cable compartment fire is constructed combined with the fault tree,and then the whole process of cable compartment fire development from fire risk factors,accident occurrence to accident consequences is revealed intuitively.2.A dynamic risk assessment method for cable compartment fire is proposed.Firstly,the mapping algorithm is used to realize the transformation from bow-the tie model to the Bayesian model,and the Bayesian model is optimized according to the actual situation of cable fire occurrence and development.Secondly,causal reasoning,diagnostic reasoning,and auxiliary decision-making reasoning are carried out based on the optimized Bayesian network.Then,numerical simulation is carried out to study the effects of different ventilation modes and fire opening and closing on the temperature and visibility of cable compartment fire,so as to realize the fire consequence analysis.Finally,the fire risk level of cable compartment is evaluated by risk matrix method.3.Probability prediction analysis based on Bayesian network model.Firstly,the forward reasoning and diagnostic reasoning analysis are carried out based on BN model,and the probability of cable compartment fire is predicted to be 4.967×10-3,the probability of cable compartment fire developing to each stage is 3.533×10-3、1.378×10-3 and 1.232×10-3;Through Bayesian network aided decision analysis,it is found that the largest cause chain of cable compartment fire is insulation damage((X6)→poor connection(D1)→leakage(C2)→cable core overheating(B1)→cable spontaneous combustion(A1)→cable fire(T).Then,through the probability adaptive analysis of the BN model,the probability of occurrence and consequence of cable compartment fire increase with time from 2007 to 2016,and its upward trend indicates the deterioration of working conditions of cable compartment.Finally,based on the dynamic Bayesian model and reasoning,the cable fire probability is predicted from4.967×10-3 increased to 1.469×10-2 at T29.3.Taking an underground utility tunnel in Kaizhou,Chongqing as an example,the case study of cable fire risk assessment is carried out.Referring to the utility tunnel system,the geometric model of the cable compartment is constructed to simulate and analyze the consequences of cable fire.Firstly,based on the development stage I of cable fire,six different ventilation modes and simulation conditions of fire door opening and closing are designed.Then the temperature and visibility slice distribution in the cable is obtained by simulation.Finally,based on the dynamic Bayesian network reasoning results and numerical simulation results,the overall risk level of the cable compartment of the utility tunnel can be predicted and evaluated by using the risk matrix,and the T9 time can be obtained.In the best ventilation mode,that is,when the fire door is closed under the comprehensive mechanical ventilation system,the fire risk level of the cable compartment increases to level II when t=100s and to level III when t=200s. |