Research On The Prediction Of Driven Pile's Ultimate Bearing Capacity Using Neural Network

Driven pile is widely used in pile foundation and how to get it's ultimate bearing capacity is one of the problem that should be determined at first. At present, there are maily two methods to obtain the ultimate bearing capacity: the direct and indirect method. Nevertheless, the limitations of these methods make it impossible to determine the actual ultimate bearing capacity of every pile with accurate, economic and convenient way in practice projects. How to get the ultimate bearing capacity accurately and conveniently is considerable concern and intractable for engineers. Fortunately, thanks to its great learning ability and non-linear massive simultaneous managing capacity, the neural network provides a model for determining the ultimate bearing capacity of single pile. This paper, based on the knowledge of the ultimate bearing capacity of driven pile and the mechanism of the neural network, proposes the neural network model on the prediction of ultimate bearing capacity of driven pile.Firstly, this paper expands the principles of driving pile, summarizes the change of resistance in different soils and the mechanism load transmission, and conducts a detailed, all-around analysis of all factors with regard to the ultimate bearing capacity. Then, it introduces the basic concept and fundamental principle of neural network, the relevant characteristics of BP net work and the principles of the genetic algorithm. Finally, on the basis of previous work, the paper sets up the neural network model on the prediction of ultimate bearing capacity of driven pile.The paper mainly focuses on the feasibility and superiority of the neural network model in solving the problem. Based on the analysis of all factors concerning the ultimate bearing capacity, it establishes the 3-level BP network with 13 factors involving pile, hammer and soil perform as neuron of input level and vertical ultimate bearing capacity as neuron of output level, and works out a relevant program, making use of the neural network toolbox of MATLAB7.0. Contrapose the disadvantages of BP network, the paper employs the self-adaption method to adjust the learning rate, provisional estimate to determine the number of neuron joints of hiden layer and using genetic algorithm to optimize the initial weight of BP network. In the end, the paper applies the neural network model to the 56 driven pile in Nanjing and Shanghai, compares the result by using neural network with results of static test to prove a favorable prediction effect.