Modeling of Compressive Strength of Self-Compacting Rubberized Concrete Using Machine Learning.

Miljan Kovačević,Marijana Hadzima-Nyarko,Emmanuel Karlo Nyarko,Silva Lozančić

doi:10.3390/ma14154346

Miljan Kovačević, Marijana Hadzima-Nyarko + Show 2 more

Open Access

https://doi.org/10.3390/ma14154346

Copy DOI

Journal: Materials	Publication Date: Aug 3, 2021
Citations: 33	License type: CC BY 4.0

Affiliation: University of Pisa, University of Osijek

Abstract

This paper gives a comprehensive overview of the state-of-the-art machine learning methods that can be used for estimating self-compacting rubberized concrete (SCRC) compressive strength, including multilayered perceptron artificial neural network (MLP-ANN), ensembles of MLP-ANNs, regression tree ensembles (random forests, boosted and bagged regression trees), support vector regression (SVR) and Gaussian process regression (GPR). As a basis for the development of the forecast model, a database was obtained from an experimental study containing a total of 166 samples of SCRC. Ensembles of MLP-ANNs showed the best performance in forecasting with a mean absolute error (MAE) of 2.81 MPa and Pearson’s linear correlation coefficient (R) of 0.96. The significantly simpler GPR model had almost the same accuracy criterion values as the most accurate model; furthermore, feature reduction is easy to combine with GPR using automatic relevance determination (ARD), leading to models with better performance and lower complexity.

Highlights

The aim of this article is to estimate the compressive strength of self-compacting rubberized concrete (SCRC) specimens using multilayered perceptron artificial neural network (MLP-ANN), ensembles of MLP-ANNs, regression tree ensembles, support vector regression (SVR) and Gaussian process regression (GPR)
The following variables are defined as input variables of the model: water, cement, fine natural aggregate, coarse natural aggregate, fine rubber, coarse rubber, superplasticizer, slag, silica fume and fly ash, and these input variables determine the number of neurons in the input layer of the ANN model
This paper gives a comprehensive overview of machine learning methods that can be used for estimating SCRC compressive strength, including MLP-ANN, ensembles of MLP-ANNs, regression tree ensembles, SVR and GPR, with different covariance functions

Summary

Introduction

Waste rubber has a major impact on the properties of fresh concrete. The use of recycled rubber decreases the entry of aggressive substances into the material, guaranteeing that the concrete has less permeability and is more durable. It enhances impact, wear resistance and durability as well as other mechanical characteristics. Wear resistance and durability as well as other mechanical characteristics It decreases compressive strength; shrinkage; and thermal conductivity coefficient, while increasing freezing resistance and sound absorption coefficient, depending on the amount and size of the rubber portion

Objectives

Methods

Results

Conclusion