Study On Accidents Statistical Analysis And Risk Assessment Methods Of Urban Subway Tunnel

With the large-scale development of urban subway construction,more and more attention has been paid to the risk management of urban subway tunnel construction.As a high-risk construction project,tunnel engineering has the characteristics of large scale,high risk,many risk factors and complicated objective conditions.In this paper,shallow-buried urban subway tunnel was taken as the main research object.According to the construction characteristics of urban subway tunnel,there are some main problems in risk mamagement of tunneling engineering such as lack of accident data,insufficient research on risk dynamic analysis,risk assessment over-reliance on empirical judgment and lack of theoretical study of analytical methods.This paper made a series of studies on the construction risk management of urban subway tunnel by means of theoretical analysis,literature retrieval,data research,Bayesian network,numerical simulation,artificial neural network and engineering practice.The following four aspects of the study are specifically carried out:(1)The statistics and data analysis of urban subway construction accidents in 2003?2016 were carried out.Statistics are made from the types of accidents,the time of occurrence,the direction of risk sources,the occurrence of regions,the occurrence of location,the occurrence of emergency and accident levels.The regularities was analyzed including the accident year and the number of deaths,the type of accident and the number of deaths,the type and the level of accidents,the type and the year of the accident and the type of accident and the source of the risk.According to the source of risk points,risk indicators were identified,and the risk transitivity was analyzed,which provided data support for later research on risk assessment..(2)A dynamic risk assessment method for shallow-buried tunnel construction based on Bayesian network was put forward and a dynamic risk rating system was constructed.The dynamic risk level of tunnel construction was divided into three parts:objective factor level,subjective factor level and monitoring factor level.According to the specific situation of tunnel construction,the specific parameters of the three factors were classified according to certain rules,and the dynamic risk rating system structure was constructed,so the Bayesian network structure was obtained.According to the selection of various factors,the nodes of the Bayesian network are determined,and the node range is determined according to the factor classification standard.The probability distribution of the node range was determined by historical statistics and expert experience.The posterior probability and sensitivity were analyzed by Netica software.Finally,the Bayesian network was used to calculate the comprehensive dynamic risk level based onthree factors,and the engineering application was carried out.(3)Aiming at the objective situation caused by the leakage of underground pipelines in the accident statistics,the evaluation of the collapse risk of urban subway tunnel construction is carried out under the condition that the existing pipeline leakage is very common.A method of forecasting the risk of collapse is put forward.The method integrates numerical simulation,artificial neural network and Monte Carlo principle to predict the collapse risk of different pipeline positions.The original data of tunnel excavation under different engineering conditions are obtained by numerical simulation,which was used as training data of the neural network,and the risk of collapse is calculated according to the Monte Carlo principle.(4)In order to solve the scale problem in analytic network process,the scale of analytic network process is improved.An exponential scale is introduced to assess the risks,and a method is proposed to determine the importance parameter for the exponential scale.The improved analytic network process is used to establish a general method for risk analysis in tunnel engineering,which is applied to the risk analysis of the Beijing subway line 6 Wuluju-Cishousi running tunnel.The results from the proposed method more accurately reflect the actual situation than those of the analytic network process using the 1-9 scale.