Research On Key Technologies Of Monitoring And Control For Cable-stayed And Buckle-suspended Construction Of Long-span CFST Arch Bridge

Long-span concrete-filled steel tube arch bridges are usually constructed by cable hoisting,cable-stayed buckling and suspension method.The key of monitoring is cable force and arch truss alignment.The purpose of this paper is to find a fast and accurate method for calculating the cable forces.The feasibility of the least square support vector machine(LS-SVM)method to predict the alignment of arch truss during hoisting is verified.The feasibility and accuracy of the two methods are verified in engineering.Furthermore,it scientifically guides the construction monitoring of long-span CFST arch bridge with arch truss cable-stayed and buckled suspension.To ensure the quality and safety of bridge construction.Taking Mengdong River Bridge of Yongji Expressway with a span of 268.0 m as an example,A three-dimensional simulation model is established.The hoisting process of arch truss is simulated.The cable tension and arch truss alignment during cable-stayed and buckle-hanging construction are analyzed.According to the monitoring data of the arch truss hoisting process of Mengdong River Bridge and the large and small well bridge with a span of 450 m,Write regression prediction model program.In this paper,the vertical displacement of the buckling control point during the construction of arch truss cable-stayed buckling is predicted and analyzed.The following is the specific content of this paper:A fast method for calculating reasonable cable forces is derived.Based on the forward iteration method,The tension of the cable is calculated by the simplified moment balance method as the initial iteration matrix.In the iteration process,a modified matrix is added.The improved forward iteration method is obtained.Then,the method is applied to solve the cable-stayed and cable-stayed forces of Mengdong River Bridge.Compared with the forward iteration method,the convergence efficiency of the improved forward iteration method is verified.The cable forces solved by this method are applied to engineering.The results show that the maximum vertical displacement of the control point can be obtained by the forward iteration method through 12 iterations.The seventh iteration of the improved forward iteration method can obtain the maximum vertical displacement of the control point with the accuracy of-19.2mm.At the same time,the measured values in the construction site are in good agreement with the theoretical calculation values,which verifies the rationality and accuracy of the method.It shows that the improved iteration method can improve the calculation efficiency and be successfully applied in engineering.During the erection of the arch truss of Mengdong River Bridge,To explore the applicability of least squares support vector machine(LS-SVM)to the prediction of arch truss alignment.The main parameters affecting the shape of arch truss are selected,the program is compiled,and the prediction model of least squares support vector machine is established and trained.By analyzing and comparing the predicted results with the measured values,the predicted accuracy of the model can not meet the requirements.The main reason is that the small number of samples is not conducive to the self-learning training of the model.The absolute value of sample data is large,the discreteness of data is large,and the prediction is difficult.The cable tension of Mengdong River Bridge adopts multi-stage cable adjustment scheme,which has great influence on the prediction results.In a word,the least squares support vector machine(LS-SVM)method needs to satisfy certain conditions to predict the vertical displacement of the arch truss in the process of hoisting.The exploratory results of the least squares support vector machine(LS-SVM)method for the linear prediction of Mengdong River Bridge are analyzed.Aiming at selecting more hoisting sections of arch truss,less absolute value of sample data and less discrete data,and the project of Dajingqiao Bridge with one-time adjustment of initial tension of cable,According to the bridge data,the least squares support vector machine prediction model is established and trained.To verify the feasibility of least squares support vector machine(LS-SVM)in predicting the vertical displacement of buckling control points during the construction of long-span CFST arch truss cable-stayed buckling and suspension.The results show that the MSE value of the training model is only 0.04,while the MSE value of the predicted model is 0.27.The fitting accuracy of the model is high in training,and the predicted results also meet the engineering accuracy requirements.Least Square Support Vector Machine(LS-SVM)has been successfully applied to predict the vertical displacement of buckling control points in the process of arch truss hoisting of Xiaojing Bridge.In a word,the improved forward iteration algorithm has high calculation efficiency and strong engineering applicability,which provides theoretical and methodological support for the calculation of cable-stayed and cable-stayed forces of long-span arch bridges.Least squares support vector machine(LS-SVM)can effectively predict the arch truss alignment in the process of arch truss hoisting of long-span arch bridges under certain conditions,and provide data support for the real-time adjustment of arch truss alignment and cable tension.Two methods are used to supervise and control the construction of long-span CFST arch bridge by cable-stayed,buckled and suspended method,which can effectively guarantee the quality and safety of the bridge construction process and completion.