Estimation of load-set behavior of driven concrete piles using artificial neural network and cone penetration test

Abstract

Pile foundations are among the most common types of deep foundations. Predicting their draft is a fairly standard engineering goal. There are many techniques for this. Many methods for determining the settlement of piles are based on various empirical relationships. In particular, the cone penetration test allows you to determine the settlement and bearing capacity of a pile using the developed empirical dependencies. However, some models must be used to develop such correlations. This study proposes to use a slightly different approach. Instead of using a pile model and a set of empirical dependencies, it is suggested to use machine learning methods. Artificial neural networks are one of the machine learning methods. The article presents a description of the development of a neural network that allows estimating pile settlement using data from the CPT test.