| In recent years,the pressure of ground transportation has increased with the increase of urban construction,the existing ground resources can not meet the needs of the city completely.At present,all kinds of engineering construction also develop towards deeper and higher direction,and subway engineering is one of the deeper representative.In recent years,urban underground space engineering develops rapidly in our country.Subway has been used in more and more cities.Air pavilion is one of the auxiliary structures of subway,and sometimes it can be used as shield machine receiving well in the last stage of subway engineering,which plays a very important role.However,the construction of the wind pavilion is generally a deep foundation pit,so people pay more and more attention to the problem of deformation and settlement caused by deep foundation pits.In order to ensure the safety of the foundation pit,onsite monitoring for the foundation pit is indispensable.However,safety problems will inevitably occur during construction.Because of the uncertainty of manual monitoring,there are still hidden dangers in prevention only from daily site monitoring data.Therefore,how to make timely risk prediction for foundation pit during or even before construction is one of the problems that need to be solved at present.This paper combines the deep foundation excavation project of People's Square Station of Changchun Metro Line 5,the law of deformation caused by excavation is studied.Firstly,the field monitoring data are extracted and classified to study the influence of excavation on its own deformation and surrounding environment.Secondly,the numerical model of foundation pit excavation is established,and the simulation data and monitoring data are extracted for comparative analysis to verify the accuracy of the model.Finally,the data extracted from the numerical model is used as a sample for training to obtain the risk of foundation pit excavation and predict the maximum surface settlement around the foundation pit.The main work done in this paper is as follows:(1)Introduce the site condition of foundation pit and extract the monitoring data.The reasons of foundation pit deformation are analyzed from four aspects: surface settlement,horizontal displacement of retaining structure,axial force of internal support and vertical displacement of retaining pile top.(2)Considering the actual situation of site construction,determine the soil constitutive model,boundary conditions,simulation parameters and actual construction conditions,simplify the model appropriately to improve the calculation speed.MIDAS GTS software is used to establish a THREE-DIMENSIONAL numerical model of deep foundation pit excavation,and the simulation data under corresponding working conditions are extracted and compared with on-site monitoring data to verify the rationality and accuracy of the model.(3)Based on RBF neural network to predict the surrounding ground settlement caused by deep foundation pit excavation,the K-means clustering algorithm is used to improve the local optimal defect of RBF neural network.The data extracted from the above numerical model are used as input samples for training,the risk prediction model of foundation pit excavation is established,the surrounding ground settlement is predicted,the prediction value is compared with the monitoring value to verify the precision of the prediction model,and the prediction value is compared with the predicted value obtained by ordinary RBF neural network to verify the precision improvement effect of the prediction model. |
PDF Full Text Request