Prediction of Lateral Behavior of Single and Group Piles using Artificial Neural Networks

In this paper, Artificial Neural Networks (ANNs) are applied to the prediction of lateral behavior of single and group piles. The results of neural networks are compared with the measured data from the model tests. A series of model tests were performed with single and group piles in the Nak-Dong Ri...

Full description

Saved in:
Bibliographic Details
Published in:KSCE Journal of Civil Engineering Vol. 5; no. 2; pp. 185 - 198
Main Authors: Kim, Byung Tak, Kim, Young Su, Lee, Sang Ho
Format: Journal Article
Language:English
Published: Seoul 대한토목학회 01.06.2001
Springer Nature B.V
Subjects:
Algorithms
Artificial neural networks
Back propagation networks
Feedback
Interaction models
Learning
Machine learning
Model testing
Neural networks
Piles
Soil-pile interaction
Studies
ISSN:1226-7988, 1976-3808
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, Artificial Neural Networks (ANNs) are applied to the prediction of lateral behavior of single and group piles. The results of neural networks are compared with the measured data from the model tests. A series of model tests were performed with single and group piles in the Nak-Dong River Sand. For the verification of applicability of BPNN (Back Propagation Neural Network), the Modified BPNN and SNN (Sequential Neural Network), a total of 200 model test results for single and group piles were used. Also, in this study, the structure of neural network with one input layer-two hidden layerone output layer was used. The number of neuron for each hidden layer determined to be 30 from the results of model test and the learning rate determined to be 0.9 to optimize network learning. Compared with BPNN, the learning epoch of the modified BPNN for the same learning pattern is reduced to 88% at the maximum and the convergence to a global minimum is guaranteed. The investigation confirmed that a SNN with feedback is more effective than a conventional ANN without feedback to simulate the lateral behavior for single and group piles. It is concluded from this study that artificial neural ne tworks based on pile-soil interaction models can be developed by properly predicting and learning algorithms based on a comprehensive data set, and that useful inferences can be made from such models.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1226-7988
1976-3808
DOI:10.1007/BF02829074