Study And Application Of Evolutionary Neural Network On Forecast Of Single Pile Ultimate Bearing Capacity

Pile foundation has been widely studied because of its practical values. The ultimate capacity of single pile(UCSP) is seriously concerned by Rock and Soil technicians. They advanced many kinds of math models in the light of different hypothesis . Because the parameters employed in the models are excessively simplified , the math models have much limitation on practical uses. This paper, firstly, introduces the transferring tnachanism of load which the pile is undergone and summarize the factors which influence IJCSP and secondly introduces the basic theory of Artificial Neural Network (ANN) and Genetic Algorithm (GA) and subsequently discusses how to use GA to evolve the weights and thresholds of ANN and then we could constitute the optimal ANN structure桬volutionary Neural Network (ENN).On the combining of ENN and the influencing factors, We set up ENN models to forecast the IJCSP of different kinds of piles. At last, the models are verified throught constuction project experiments.