Research On Construction Control And Prediction Of Long-span Cable-stayed Bridge

With the continuous improvement of the spanning capacity of long-span cable-stayed bridges,the bridge type is becoming more and more complex,which puts forward higher requirements for the quality of construction,followed by a series of problems about accuracy control in the construction.Therefore,the establishment of a reasonable monitoring plan is a prerequisite to ensure the smooth construction of the bridge.The construction monitoring work should not only ensure that the bridge construction is always in a safe state,but also carry out scientific and reasonable analysis to simplify the construction process and improve the construction efficiency.By consulting a large number of literature and related books,this paper understands the current development status and development trend of contemporary cable-stayed bridges,and the important theoretical methods,links and principles of construction control of cable-stayed bridges.This article takes the construction of the main bridge of the Jiayu Yangtze River Highway Bridge,an important controlling project in Hubei Province as the background,the main research contents and related results are as follows:Adopts an adaptive unstressed configuration control method to control and analyze the construction process of the bridge,and uses the BDCMS to whole process of construction was calculated and analyzed,which provided guidance for the construction monitoring of the main bridge in the later period.A detailed construction monitoring plan and advanced monitoring equipment were developed to ensure the accuracy control of the entire cable-stayed bridge during construction.Explore the possibility of the application of artificial intelligence,a cutting-edge technology,in the construction control of super-long-span cable-stayed bridges.A deep confidence network prediction model based on deep learning theory was established using MATLAB computer programming language design software.The linear elevation of the main girder of the bridge during the construction stage was predicted.For the problems of large differences in the unit and magnitude of the collected data,Normalized processing,supervised fine-tuning of the entire model using the reverse broadcast algorithm.The feasibility of this method is verified by comparison with the actual measured values on site,and the existing problems are pointed out.