| The high-speed railway request not only the strength of the subgrade meeting the requirement, but also more strict settlement control of the foundation, which including the control of the settlement during the filling period and the final settlement after the filling period. The theory of the consolidation and settlement is not complete yet and the actual situation of large area loading is complex because of the changing type and length of the pile, making the theoretical calculation and numerical simulation more difficult, this article is based on the observational data of the composite foundation settlement test on Jinan West railway station to improve the existing methods of prediction and study the methods used in the settlement predict of the composite foundation with large area loading. The study is as followed:1. Compare the hyperbolic method, the neural network, the grey system used in the settlement prediction to find their advantage and disadvantage and improve these methods;2. Analyzing on the feature of the settlement simply, including the difference of settlement because of the changing width of the subway and the difference because of the changing depth of the foundation;3. Using the Elman neural network to predict the settlement of various types of the composite foundation and comparing the effect on the prediction with the height of the filling subway considered or not;4. Using the hyperbolic method and the wavelet-BP neural network to predict the settlement of various types of the composite foundation and comparing the effect and applicability of two methods;5. Using the BP neural network to predict the settlement of various types of the composite foundation and comparing the effect and applicability of two methods;6. Using the BP neural network to predict the settlement at different depth of the composite foundation at the station area, than generalizing the trained neural network and comparing the effect, at last proposing a scheme to improve.
|
PDF Full Text Request