Genetic programming for predicting axial capacity of driven piles

Abstract

The behavior of pile foundations under axial loading is complex and not yet entirely understood. Most available methods for predicting axial capacity of driven piles have failed to achieve consistent success in relation to accurate pile capacity prediction. However, among available methods, the cone penetration test (CPT) based models have shown to give better predictions in many situations. In an attempt to obtain more accurate axial pile capacity predictions from CPT test results, the genetic programming (GP) technique is used in this study. GP is a relatively new artificial intelligent computational technique that has been recently used with success in the field of geotechnical engineering. The data used for development of the GP model are collected from the literature and comprise a number of 78 pile load tests and CPT results. The model robustness is further investigated via a sensitivity analysis, and the performance of the GP model is compared with three of the most commonly used CPT-based traditional methods. The results indicate that the GP model provides more accurate axial capacity predictions of driven piles and outperforms the traditional methods.