Load-Deflection Prediction of Helical Piles Using Hyperbolic & Machine Learning Methods
Availability
On-Demand
Cost
$0.00
Credit Offered
No Credit Offered

Abstract:

As the sophistication of modern structural design methods has improved over time, the ability to accurately predict and model foundation stiffness has become increasingly more important. In complicated structural systems with deep foundations, correct calculation of the distribution of load and serviceability requires proper modeling of axial load-deflection characteristics. Load is attracted to stiffness. To keep pace with advances in structural analysis, there is much need for more practical and scientific work on the load-deflection prediction of deep foundations. This is especially true for the emerging technology of Helical Piles, which are being used for increasingly more applications. The load-settlement mechanism of an axially loaded Helical Pile is a complex soil-structure interaction problem influenced by Helical Pile dimensions, installation method, and ground conditions. Due to uncertainties and a general lack of consistent methodology, pile load tests are often usedto verify load-deflection design assumptions. In this study, a comprehensive data set of over 350 full-scale tension and compression tests are analyzed and summarized. These data points were collected for the AC358 evaluation of individual Helical Pile manufacturers by CTL Thompson, Inc. The data includes Helical Pile sizes ranging from 1.5” square shaft to 4.5” round shaft. Load test data are evaluated using a statistical hyperbolic fit technique introduced here and a previously published graphical method (Perko 4 Point Method). Trends in the data are compared with 11 categories of geometry and geotechnical factors such as pile shape, shaft size, number of helix, theoretical capacity by torque, and soil type. Empirical factors are developed to enable better prediction of Helical Pile load-deflection characteristics.
Andre; Dr. Andre Hawks; Dr. HOWARD A. PERKO
Seminar Proceedings
Powered By