Please use this identifier to cite or link to this item: https://hdl.handle.net/20.500.14279/33019
Title: Developing predictive models for the load-displacement response of laterally loaded reinforced concrete piles in stiff unsaturated clay using machine learning algorithms
Authors: Braun, Kevin T. 
Markou, George 
Jacobsz, S. W. 
Calitz, Duan 
Major Field of Science: Engineering and Technology
Field Category: Computer and Information Sciences;ENGINEERING AND TECHNOLOGY;Civil Engineering;Other Engineering and Technologies
Keywords: Soil-structure interaction;Machine learning algorithms;Predictive models;Reinforced concrete pile;Horizontal load-displacement response
Issue Date: 1-Jun-2024
Source: Structures, 2024, vol.64
Volume: 64
Journal: Structures 
Abstract: The design of pile foundations that are expected to develop significant lateral loading is a complex procedure that requires the development of objective and accurate design formulae that will not be based on semi-empirical know-how. For this reason, the main objective of this research work is to develop predictive models that will be able to compute the overall mechanical response of reinforced concrete (RC) piles embedded in unsaturated clay. To achieve this goal, experimental data, and advanced nonlinear 3D detailed finite element (FE) modelling were used to construct datasets comprising multiple results related to the ultimate capacity and corresponding horizontal deformation of RC piles that are loaded horizontally until failure. In total, three datasets were developed and then used to train and test predictive models through the use of various machine learning (ML) algorithms. After successfully developing various predictive models, an out-of-sample dataset was developed and used to further validate the accuracy and extendibility of the predictive models. Finally, the most accurate ML-generated predictive model was used to predict the mechanical response of a RC pile embedded in unsaturated clay that was experimentally tested. The ability of the proposed predictive model is demonstrated through this pilot research work.
URI: https://hdl.handle.net/20.500.14279/33019
ISSN: 23520124
DOI: 10.1016/j.istruc.2024.106532
Rights: Attribution 4.0 International
Type: Article
Affiliation : University of Pretoria 
Neapolis University Pafos 
SRK Consulting 
Publication Type: Peer Reviewed
Appears in Collections:Άρθρα/Articles

Files in This Item:
File Description SizeFormat
Developing predictive models.pdf10 MBAdobe PDFView/Open
CORE Recommender
Show full item record

Page view(s)

30
Last Week
7
Last month
checked on Nov 6, 2024

Download(s)

16
checked on Nov 6, 2024

Google ScholarTM

Check

Altmetric


This item is licensed under a Creative Commons License Creative Commons