Application Of Gray Neural Network Model BP Algorithm In High-speed Rail Settlement Prediction Of Improvement

Deformation monitoring is particularly important for construction and management invarieties of engineering structures such as dams, bridges, tunnels, buildings and highways,railways, etc. Construction technology of deformation monitoring for high-speed railway is alsobeing promoted in China, It is a project that is large sophisticated and demanding strictlyconstruction of all aspects, so it is necessary of controlling strictly settlement of underlineengineering, effectively processing data and reasonably evaluating in order to ensure the safety ofthe construction and operation. Subsidence observation of bridge and culvert structures is crucialto the construction of high-speed railway. Bridge structure must be of sufficient strength andstiffness which makes high-speed railway capable of withstanding larger power, in order tomaintain high speed, comfort and safety. Therefore, the settlement for piers as the basic of bridgeis critically important.At present, there are many methods for deformation, such as Grey system theory, RegressionAnalysis Prediction Method, artificial neural network and so on, which are widely applied inexcavation, buildings, dams, but it has shortcoming in theory reference for settlement predictionmodel of high-speed railway. Based on the River pier in Nanchang Station of Shanghai-KunmingPassenger Dedicated Line, this thesis studied subsidence prediction model that applies to bridgepier by relevant theory.The paper firstly overviewed related theory in underline engineering, simple describingdeformation monitoring methods, and then several key technology was detailedaddressed——GM(1,1) model, grey linear regression model and the BP neural network model,for which was analyzed in basic principle and modeling process, submitting shortcoming andOptimization in artificial neural network. Based on the theories, Gray BP Neural NetworkCombined Model was introduced. Finally, with the optimization model applied in observationdata of the River pier of Nanchang in Nanchang Station of Shanghai-Kunming PassengerDedicated Line, corporation between the optimization model and the single model was taken,calculating by using MATLAB.The studies was showed that the optimized BP neural network that has better efficiency thantraditional BP neural network model improved forecast accuracy, which was better suited forsettlement prediction of high speed railway bridge pier, and it has Feasibility of certain and agood practicality.