The time-dependent increase in pile capacity after its driving may be significant. A reliable prediction of this increase (setup) may lead to a significant saving in pile design. A neural network model to predict pile setup was developed. A database derived from field tests reported in the literature showing setup of driven piles was compiled, with six variables selected as input parameters: soil type, pile type, pile diameter, pile length, time after pile installation, and initial effective stress at tip. Ultimate pile capacity at beginning of restrike (QBOR) is the sum of pile capacity (QEOD) at end of drive and increased in capacity (ΔQBOR) caused by setup, which in this study is predicted by a backpropagation neural network. The results demonstrate that the neural network model provides a better prediction than predictions by the available empirical methods. A neural network model can serve as a reliable tool for the prediction of pile setup, and further training with additional data will lead to additional improvement in the quality of prediction.
A Neural Network Model for Prediction of Pile Setup
Transportation Research Record
Transportation Research Record: Journal of the Transportation Research Board ; 2004 , 1 ; 12-19
2007-01-01
Article (Journal)
Electronic Resource
English
A Neural Network Model for Prediction of Pile Setup
Online Contents | 2007
|Prediction of Pile Setup in Clay
Transportation Research Record | 1999
|Prediction of Pile Setup in Clay
Online Contents | 1999
|Transportation Research Record | 2010
|Regression versus artificial neural networks: Predicting pile setup from empirical data
Online Contents | 2014
|