Deformation Prediction And Control For High-speed Rail Bridge Based On Genetic Neural Network

Combined with project examples, applicatied of genetic algorithms and neural network, this papers investigated the way to control and prediction the large span prestressed concrete continuous girder bridge of high-speed rail, the main contents are as follows:(1)This paper comprehensively analyzed the control method for bridge construction, studied the control method of genetic algorithms combined with artificial neural network adaptive to applicability of control problems on bridge construction. The results are applied to construction control of many large span prestressed concrete continuous girder bridges along the Shanghai-Kunming high-speed rail(2)Through in-depth analysis of factors affecting the deformation of bridge and sensitivity analysis to the linear factors on the structural by the way of finite element model, this study concluded that weight, prestress as well as the gradient of temperature have an significantly affection to the linear of structure.(3)Combining genetic algorithm and BP neural network to construct the genetic neural networks, taking the weight, effective prestress, measuring temperature, segmental deflection of the main beam as parameters, this paper Proposed an adaptive genetic algorithm. Taking the minimization of sampling error as objective function, this paper realized the optimization of weights and thresholds to the neural network model to get a new BP neural network and implement procedures of the algorithm based on MATLAB implementation.(4)By application of the above algorithm, combined with project examples and contrast of measured data during construction, this paper analyzed the feasible of application to large span prestressed concrete continuous girder bridgesThis algorithm have the advantages of high precision to predict linearly changes of bridges during the construction and also be to provide reference to linear control of high precision for bridges of high-speed railway.