Realization Of Random Field In Geotechnical Engineering And Face Stability Analysis For A Shield Tunnel

The construction of urban underground tunnel project has attracted the attention of more scholars because of its complicated engineering and high construction risk.The main risk of shield construction lies in the size of support stress of the pressure chamber,the support stress is too large,easy to push out the soil,too small may cause the soil of tunnel excavation surface to collapse.This kind of accident has appeared many times,so it is very important to study the critical support stress of the face for a shield tunnel.In the study of the face stability analysis for a shield tunnel,there are two factors that affect the value of the critical support stress of the simulated shield excavation: first,the non-homogeneous characteristics of the earth body are ignored,that is to say,the human idealization of the soil body in each layer of soil parameters are the same.But in fact,according to geological survey and research,the soil body own parameters in the space with a high degree of uncertainty and variability after many years of external influence and experience driven by the physical effect;In the face stability analysis for a shield tunnel and analysis of the critical support stress probability,the traditional Monte-Carlo simulation(MCS)calculation reliability requires more samples,and the statistical probability distribution requires more random field operations,in order to achieve a certain statistical accuracy,heavy workload.Therefore,based on the theory of random field,an active learning method combined with the Kriging model and Monte-Carlo method—AK-MCS(Active learning reliability method combining Kriging and Monte-Carlo Simulation)is introduced to analyze face stability analysis for a shield tunnel.Based on random field simulation and deterministic condition simulation(based on random field simulation)The soil body is homogeneous earth body)to determine a certain function function,in the calculation of the probability of failure of the critical support stress of simulation calculation under deterministic conditions,the use of functional function prediction value to calculate the random field estimation of Monte-Carlo simulation sample,probability statistical analysis.The research shows that AK-MCS method can effectively utilize fewer sample points,calculate the failure probability of deterministic condition simulation,and can estimate the critical support stress of the large sample random simulation with fewer samples under the premise of meeting the precision requirement,and thus facilitate the probability statistical analysis,which is of great significance to the choice of the support stress in actual engineering.