Article
Forecasting the Bearing Capacity of the Driven Piles Using
Advanced Machine-Learning Techniques
Mohammed Amin Benbouras
1,2,
* , Alexandru-Ionu¸t Petri¸sor
3,4
, Hamma Zedira
5
, Laala Ghelani
5
and Lina Lefilef
6
Citation: Benbouras, M.A.; Petri¸sor,
A.-I.; Zedira, H.; Ghelani, L.; Lefilef, L.
Forecasting the Bearing Capacity of
the Driven Piles Using Advanced
Machine-Learning Techniques. Appl.
Sci. 2021, 11, 10908. https://doi.org/
10.3390/app112210908
Academic Editors: Nikos D. Lagaros
and Vagelis Plevris
Received: 10 October 2021
Accepted: 16 November 2021
Published: 18 November 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
1
Technology Department, École Normale Supérieure d’Enseignement Technologique de Skikda (ENSET),
Skikda 21001, Algeria
2
Central Public Works Laboratory (LCTP), Algiers 16006, Algeria
3
Doctoral School of Urban Planning, “Ion Mincu” University of Architecture and Urbanism,
010014 Bucharest, Romania; alexandru.petrisor@uauim.ro
4
National Institute for Research and Development in Tourism, 50741 Bucharest, Romania
5
Civil Engineering Department, University of Abbes Laghrour, Khenchela 40051, Algeria;
zedirahamma2003@yahoo.fr (H.Z.); ghilanilaala@yahoo.fr (L.G.)
6
Department of English Language and Literature, Mohamed Seddik Ben Yahia University,
Jijel 18000, Algeria; lefileflina@gmail.com
* Correspondence: mouhamed_amine.benbouras@g.enp.edu.dz
Abstract:
Estimating the bearing capacity of piles is an essential point when seeking for safe and
economic geotechnical structures. However, the traditional methods employed in this estimation
are time-consuming and costly. The current study aims at elaborating a new alternative model for
predicting the pile-bearing capacity based on eleven new advanced machine-learning methods in
order to overcome these limitations. The modeling phase used a database of 100 samples collected
from different countries. Additionally, eight relevant factors were selected in the input layer based
on the literature recommendations. The optimal inputs were modeled using the machine-learning
methods and their performance was assessed through six performance measures using a K-fold
cross-validation approach. The comparative study proved the effectiveness of the DNN model,
which displayed a higher performance in predicting the pile-bearing capacity. This elaborated model
provided the optimal prediction, i.e., the closest to the experimental values, compared to the other
models and formulae proposed by previous studies. Finally, a reliable and easy-to-use graphical
interface was generated, namely “BeaCa2021”. This will be very helpful for researchers and civil
engineers when estimating the pile-bearing capacity, with the advantage of saving time and money.
Keywords:
pile-bearing capacity; machine learning; deep neural network; K-fold cross-validation
approach; sensitivity analysis
1. Introduction
Pile foundations are used to transmit construction loads deep into the ground in
order to ensure structure stability [
1
,
2
]. Furthermore, computing the bearing capacity of
piles is essential when designing economic and safe geotechnical structures [
3
]. To date,
numerous approaches have been conceived for the sake of creating alternative methods
and techniques that contain numerical, experimental, and analytical approaches aiming at
predicting the bearing capacity of piles [
4
–
6
]. Among the most frequently used methods
is the Cone Penetration Test (CPT), known for producing accurate results in a variety
of situations [
7
,
8
]. This is probably due to the fact that CPT-based methods have been
modeled in harmony with the CPT results, which were proven to estimate more effective
different geotechnical properties, and make more precise pile capacity predictions [
6
].
Other semi-empirical methods have been widely utilized, such as Meyerhof’s formula,
which could yield an acceptable pile-bearing capacity [
4
]. On the other hand, the High-
Strain Dynamic Load Test (HSDLT) and the Static Load Test (SLT) have been employed
Appl. Sci. 2021, 11, 10908. https://doi.org/10.3390/app112210908 https://www.mdpi.com/journal/applsci
