Research On Fire Risk Assessment Of Subway Station Based On Bayesian Network

With the vigorous advancement of my country’s urbanization construction,the traffic demand and traffic pressure are increasing day by day,and the subway is favored by citizens in the choice of transportation due to its advantages of convenience,comfort,environmental protection and small footprint.However,the subway station is located in the underground space and the environment is airtight.If a fire occurs in the subway station,it will be difficult to escape and rescue work,with high risk factor and a large scope of influence.In the operation of subway stations,there are still problems such as insufficient fire risk prevention and insufficient fire emergency capability.Therefore,how to use a scientific and efficient method to evaluate the fire risk level of a subway station and discover the key risk factors of a subway station fire in advance will be of great significance for the subway station to grasp the hidden fire situation and improve the level of risk prevention and control in the station.First,based on the causes of fire accidents in subway stations,and referring to the industry standards of subway stations,fire risks are divided into four basic elements,namely personnel,management,environment and equipment,and the specific incentives are discussed based on these four aspects,and combined with literature research,to preliminarily determine the causative factors of fire risk in subway stations,and refer to expert suggestions to screen and adjust the influencing factors to complete the construction of the subway station fire risk assessment system.Secondly,the fire risk assessment model of subway station based on Bayesian network is established.According to the causal logical relationship,the Bayesian network structure is obtained by transforming the evaluation index system,and the state of the network nodes is determined in combination with the risk characteristics,thereby establishing the Bayesian network structure for the fire risk assessment of subway stations.After completing the qualitative modeling,the marginal probability of the risk occurrence possibility is calculated through triangular fuzzy language and questionnaire,and the accident risk loss is weighted by the improved AHP methodimproved entropy weight method,and then the conditional probability based on the accident loss weight is proposed.After the model parameters are determined,on this basis,the classification standard of risk level is given.Finally,taking Wuhan Mafangshan subway station as an example,GeNIe software is used to establish a Bayesian network structure to evaluate the fire risk of the subway station,predict the fire risk probability of the station through inference calculation,and find out the weak links and causes of the fire risk of the subway station through diagnosis and reasoning.The evaluation results are compared and analyzed,and then risk prevention and control measures are proposed.The research results show that the fire risk level of Mafangshan Station is "important",the main cause of fire accidents in Mafangshan Station is fire safety management,the risk probability is 78%,and the secondary causes are fire equipment,monitoring equipment,Station emergency equipment and daily fire inspections.By comparing with the fuzzy comprehensive evaluation method,the Bayesian network model can obtain a clear and intuitive risk probability,find out the loopholes in the fire risk management of the station,and can reliably evaluate the fire risk of the subway station,which has a broad application prospect in measuring the fire risk level of the subway station.