Study On Risk Assessment And Settlement Early Warning Of Land Subsidence In Shield Tunnel Construction Of Metro

In the process of subway tunnel construction,shield tunneling construction method is widely adopted due to its advantages such as high degree of mechanization,wide application range of soil layer and fast construction progress.When shield construction cuts through the soil,it will inevitably disturb the stratum to varying degrees,and then cause ground subsidence,which will not only affect the construction safety,but also damage the surrounding buildings and underground pipelines.Based on the project of Chengdu Metro Line 8 Airport Expressway Station to Yixin Avenue Station,this paper identifies the risk factors of land subsidence during the construction of EPB shield,establishes the risk evaluation index system and quantitative risk evaluation model,establishes the prediction model of land subsidence combined with large data and intelligent algorithm,and designs the early warning system of land subsidence on the basis of the prediction model.Through the project example verification,it is found that the research has certain practical value for the risk control of land subsidence in the shield construction process.The main research contents and achievements are as follows:(1)Through literature review,summarize common risk identification methods,combined with the characteristics of subway shield construction,select WBS-RBS risk identification method to identify land subsidence risk factors comprehensively,and finally determine five first-level risk factors and 36 second-level risk factors.At the same time,in order to eliminate the information overlap among various risk factors,the principal component analysis method is used to deal with the risk factors.Finally,36 secondary risk factors are fitted to 17 independent representative risk factors,and on this basis,a set of universal land subsidence risk evaluation index system is established.(2)Combining the idea of "variable weight" with the extension theory,a "variable weight extension land subsidence risk assessment model" is constructed.The risk of land subsidence between airport expressway station and Yixin Avenue station is pre-evaluated by the model.The evaluation results show that: The risk grade correlation coefficient of variable weight model is more discrete than that of constant weight model,and the risk grade is easier to determine.At the same time,the main factors affecting land subsidence are geological factors,tunneling parameters,groundwater factors,excavation and dumping factors,slurry factors and monitoring auxiliary factors,which should be paid more attention to during construction.(3)In order to warn the land subsidence,it is necessary to predict the land subsidence firstly.Based on the pre-evaluation of subsidence risk,an improved PSO-RBF neural network prediction model is constructed.The influencing factors of land subsidence and the amount of land subsidence are regarded as input and output variables of the model respectively.The model is trained and predicted by monitoring data collected from construction site.The error between the predicted results and the actual settlement is within 8%,which meets the requirements of deformation monitoring.The design of land subsidence early warning system is carried out by using the graphical interface design module(GUI)of MATLAB.The settlement prediction model is embedded into the designed early warning system in the design process.The purpose of early warning of land subsidence is achieved by setting the threshold of settlement early warning.The paper combines the “variable weight” and extension theory to pre-evaluate the ground subsidence risk of subway shield construction,combine the improved PSO algorithm and RBF neural network to predict the ground subsidence,and design the settlement warning based on the prediction model.These studies provide a new idea for risk management of ground subsidence in metro shield construction.