Pile Driving Records Reanalyzed Using Neural Networks

Abstract

Pile driving formulas are commonly used to estimate the load capacity of driven piles. The formulas assume that there is a correlation between the pile set and the ultimate load capacity of the pile. The important factors influencing the load capacity include the hammer characteristics, the properties of the pile and soil, and the pile set. The present technical note investigates the feasibility of using neural networks to predict the load capacity of driven piles. Neural networks attempt to simulate the process by which the human brain learns to discern patterns in arrays of data. The data used in this study were derived from actual pile driving records. First, the neural network concepts are reviewed, then the neural network model for predicting the pile capacity is presented. The neural network predictions were found to be more consistent and reliable than other, more conventional pile driving formulas.