Prediction Model And Application Analysis Of Subway Structure Deformation

The subway has the advantages of saving space,as well as having large capacity,avoiding interference with road traffic on the ground,saving a lot of time for people and solving the problem of urban traffic congestion,which has become a major travel mode for people.The construction of subways is increasingly developing,but the construction and operation of the subway are mostly underground,which means that a large number of tunnels need to be built,the construction environment should be totally enclosed,there are much more difficulties in construction and operation of the subway than ground traffic.But the safety is the first factor to be considered whether in the process of construction or operation.The deformation tendency as well as hidden danger of the subway structure can be predicted and detected by modeling and analyzing the deformation monitoring data,which is of great significance to guarantee the safety in construction and operation of the subway.The main research contents of this paper are as follows:(1)The factors affecting structural deformation of subway tunnels during operation were analyzed,which contains the soil under the tunnel,as well as the construction above and around the subway tunnel,the differences between the tunnel sections and stations,the vibration of the subway vehicles during the operation period,the surface settlement and natural disasters;The control standards for the deformation monitoring of subway structures and the design of the monitoring network were studied,in order to ensure that the monitoring network meets the specifications and project requirements;The principle,method and specific process of automatic deformation monitoring of subway structure were described,which were compared with traditional manual monitoring,then the advantages of automatic monitoring were analyzed.(2)Several deformation prediction models for the subway were detailedly introduced and analyzed,the modeling process of several prediction models for the subway deformation,such as time series model,as well as BP neural network model,fusion model of time series and BP neural network,optimization of BP neural network model by genetic algorithm,and fusion model of genetic algorithm optimized time series and BP neural network were mainly studied.The innovation of the genetic algorithm for the optimization of the model and its application in deformation prediction of subway structures were proposed.The concrete project demonstrates that,the RMSE of time series model,BP neural network model were respectively ?0.667 mm,?0.359 mm,the RMSE of fusion model of time series and BP neural network,BP neural network optimized by genetic algorithm and fusion model of genetic algorithm optimized time series and BP neural network were respectively ?0.271 mm,?0.266 mm,?0.221 mm,which were respectively 59%,60%,67% higher than that of the time series model(3)The safety monitoring system with software and hardware integration for subway based on measuring robots was implemented,the monitoring time was shortened and monitoring efficiency was improved by applying data acquisition device based on 4G(RTU),the data acquisition and transmission efficiency were improved by developing a data acquisition platform of multivariate measurement sensors,the data storage and reading functions were realized,and a structural deformation intelligent monitoring and management system was established by using the technology of SOA and cloud computing.The structural deformation of the subway can be monitored in real time,the data can be transmitted back to the center,and can be analyzed as well as predicted automatically,the possible danger can be forecasted by this system,which ensure the safety of the subway structure.