Dynamic Risk Assessment And Risk-based Support Decision Analysis On Tunnel Squeezing

With the development of tunneling engineering towards the trends of deeper and longer,engineering disasters have become increasingly prominent.Especially in the western mountainous area of China,which is renowned for the significant complexity of the geological environment.Squeezing in rock tunnels often occur in soft(weak)rock masses at great depth,like phyllite,mudstone,fault fracture zone and other bad geological combinations.It is also common in tunnels subjected to high horizontal in situ stresses in tectonically active regions.Squeezing may cause tunnel collapses,budget overruns and construction delays.Given the variability of geologic conditions and construction uncertainties faced by tunnel project,it's of great importance to systematically assess and predict the risk of squeezing,as a basis for effectively managing the risks,thus assuring the safety of the construction process.NATM naturally highlight the concept of information construction and updating the construction strategy.This paper aims to propose a novel method to quantitatively and dynamically assess the risk of squeezing,based on which a decision supporting component was developed by the decision tree analysis to determine the optimal supporting strategy.The main content of the paper is summarized as follows:(1)On the basis of the geo-mechanical analysis of squeezing,several damage factors have been identified and five main parameters of which have been employed in the probabilistic analysis of the prediction of squeezing.Specially,Ordinal Logistic regression was applied to an extensive well-documented database to achieve the probabilistic classification of squeezing intensity.(2)The consequences and losses of the disaster mainly include economic losses and time delays,and the severity of consequence is associated with the intensity of excessive deformation.First 4 levels of construction strategy have been proposed corresponding to the intensity of the excessive deformation,then a PERT-based Monte Carlo simulation methodology is used herein to obtain the cumulative distribution of economic losses and time overruns.Subsequently,decision tree analysis is involved to choose the optimal supporting strategy,where optimal refers to minimum risk.By then this paper achieved the quantitative risk analysis and risk-based decision support.(3)Geological model featured by 4 geologic parameters is described by a continuous-time,discrete states Markov process.With the excavation process,more geological information could be obtained through the Geological follow-up survey,in which some quantitative indexes have been used to enhance the reliability of the observation.Based on the geological information revealed in some specific locations,the state probability of each single geological parameter and further the state probability of ground condition along the tunnel alignment are inferred through Bayesian updating and transition probability calculation.Through the combination of those two methods could achieve the dynamic risk assessment and decision making.The above method is applied to the Zhuokeji tunnel project,for dynamic assessment and support decision of the squeezing risk of the un-excavated section.The results could be used to guide the construction,to provide support decision-making recommendations,as a way to achieve information construction,so the results have practical significance for safe construction,disaster prevention and mitigation of tunnel engineering under high ground stress.