Surface Settlement Prediction Model And Its Application In Subway Pit Excavation

With the rapid development of China's social economy and science and technology,the construction of various large-scale engineering project facilities has greatly increased.In addition,the increase in population and the acceleration of urbanization have made the construction of transport facilities particularly important.However,the surface collapse caused by this phenomenon is not uncommon,threatening the personal safety of construction personnel and surrounding residents.In order to discover and prevent this phenomenon in time,it is necessary to conduct safety inspection on the surrounding surface during construction,and a settlement monitoring mechanism must be established to monitor and predict the surface settlement trend so as to be able to timely take corresponding measures.Measures to prevent the occurrence of engineering disasters and ensure the safety of structures and personnel.This paper takes the ground settlement data of the Linquan West Road station of that subway line No.3(under construction)as an example,and selects two typical representative monitoring sites DBC14-2 and DBC5-2 for the study.The maximum amount of sinking is-17.01mm and-16.80mm,respectively.The gray model,Kalman filter model and BP neural network model are used to implement and check the Matlab program.The first monitoring point is predicted using different gray models before and after noise reduction,and the optimal grey model is obtained through analysis and comparison.The Kalman filter model and the BP neural network model were used to predict the monitoring points,and the conclusions were analyzed and compared.The same method as above was used for the prediction calculation of the second monitoring point to verify the pros and cons of the three prediction models under the engineering geological conditions.The comparison prediction and verification results show that the gray model with noise reduction is superior to the gray model without noise reduction,and the four-dimensional gray model with noise reduction has the highest prediction accuracy;the Kalman filter and BP neural network prediction results It shows that the results predicted by the 4-D gray model of noise reduction are better than the Kalman filter model.The results obtained by using the BP neural network model prediction model are better than those by the 4-D gray model of noise reduction;the study of this topic predicts the settlement of the ground surface.The construction of similar projects and the prevention of disasters are of great significance.