Research On Construction Control Of Continuous Beam Bridge Based On BP Neural Network

In recent years in our country bridge, whether it is urban overpass bridge, railway bridge, highway bridge, or across rivers, valleys, cliffs of bridge, prestressed concrete girder bridge has the advantages of large span, and dominant. It is due to the large capacity, high terrain adaptability, design, construction technology is mature, reasonable cost, is widely used. But the construction parameters and design parameters of the different, resulting in the bridge’s alignment and stress and design. Therefore, in order to ensure the smooth construction of the whole bridge to meet the design requirements of linear and stress, are needed.In this paper, the Hong Shu Lin bridge (45+3 X 70+45m) is a practical engineering background. The main conclusions of this paper are as follows:(1) Finite element simulation analysis of the mangrove bridge using finite element analysis software Midas. Combined with the existing experience and the results of simulation analysis, it can be used to determine the displacement and stress of the structure simulation analysis of mangrove bridge.(2) The process of linear and stress control of Hong Shu Lin bridge is described in detail. Calculation of the elevation of the mangrove bridge erection, and the ultimate bridge elevation control of the bridge is within ±15mm, and the stress error is within reasonable limits.(3) Using BP neural network calculation method to predict the 8#,9# beam elevation model. By analyzing and optimizing, the hidden layer neurons number is 8, momentum and an adaptive gradient descent method for optimal BP neural network model. Finally,8#,9# beam elevation model prediction error was 9.7mm, with satisfactory results.The results of this study provides references for using BP neural network calculation method of continuous beam bridge construction control and similar bridge construction control.