Error Analysis On Construction Control Of Long Span Cable-Stayed Bridges With Steel Box Girder

The main bridge of Sutong Yangtze River Bridge under construction is a cable-stayed bridge with steel box girder of double pylons and cable planes with main span of 1088m, which will be the longest span cable-stayed bridge in the world after completion. In this thesis, taking the construction control of Sutong Yangtze River Bridge as the project background, the sources and reasons of the errors of the construction of Sutong Yangtze River Bridge are studied.Because the study of the influence of parameters on whole construction course and completed bridge state of long-span cable-stayed bridges with steel box girder is lack, the characteristics and the reasons of the errors are investigated in the thesis. Then based on sensitive analysis of parameters, the effects on the structure in construction course and completed bridge state of dead-weight of the bridge girder,rigidity of stay cable and so on are studied. At last, the key parameters are determined.With regard to parameters that may remarkably affect the computed results of the bridge structure, the method of parameter identification based on neural network technology is put forward. The possibility of the application of the neural network method in the construction control of cable-stayed bridge is explored. By analyzing the Sutong Yangtze River Bridge using the nonlinear analysis system-NLABS, samples are obtained. Based on BP nerval network, a parameter identification model of construction control is established. A parameter identification program is designed by using MATLAB programming language. At last, the application is verified feasible in the construction control of cable-stayed bridge by comparing its result with the real value of parameters which are identified.