Modeling load-settlement curves of behaviour bored piles using artificial neural networks

Accurate prediction of pile behavior under axial loads is necessary for safe and cost effective design. This paper presents the development of a new model, based on artificial neural networks (ANNs), to predict the load-settlement relationship of behavior of bored piles subjected to axial compression loads. ANNs have been recently applied to many geotechnical engineering problems and have shown to provide high degree of success. In this paper, two ANN models are developed; one for bored piles installed in sand and mixed soils, and the other for cohesive soils. The data used for ANN model development are collected from the literature and comprise a series of in-situ bored pile load tests as well as cone penetration test (CPT) results. Predictions from the ANN models are compared with those obtained from the experimental tests, and statistical analysis is used to assess the performance of ANN models. The results indicate that ANN models are able to accurately predict the load-settlement relationships of behavior of bored piles with high accuracy.