Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

From Experimental Studies to Predictive Machine Learning Modelling : Polypropylene Fibre Reinforced Concrete

Bayat Pour, Mohsen LU (2025) In Results in Materials
Abstract
This study investigates the influence of polypropylene fibres (PF) on concrete performance across a finely resolved dosage spectrum (0.0–2.0% by cement mass, in 0.1% increments) using 2,100 laboratory specimens. The experimental programme evaluated compressive strength, tensile strength (flexural and splitting), modulus of elasticity, and water penetration depth. Predictive modelling was conducted using Random Forests (RF) and Support Vector Regression (SVR), trained and evaluated with simple cross-validation and benchmarked using mean absolute error (MAE) and coefficient of determination (R²). The results reveal distinct optima for strength indices and a threshold behaviour in permeability, with PF dosages between 0.2% and 0.6% balancing... (More)
This study investigates the influence of polypropylene fibres (PF) on concrete performance across a finely resolved dosage spectrum (0.0–2.0% by cement mass, in 0.1% increments) using 2,100 laboratory specimens. The experimental programme evaluated compressive strength, tensile strength (flexural and splitting), modulus of elasticity, and water penetration depth. Predictive modelling was conducted using Random Forests (RF) and Support Vector Regression (SVR), trained and evaluated with simple cross-validation and benchmarked using mean absolute error (MAE) and coefficient of determination (R²). The results reveal distinct optima for strength indices and a threshold behaviour in permeability, with PF dosages between 0.2% and 0.6% balancing mechanical enhancement and substantial reductions in water penetration, although accompanied by a pronounced reduction in elastic modulus at very low PF contents. The RF models exhibited superior predictive performance, consistently outperforming SVR across properties. The experimental outcomes demonstrate that incorporating PF into the concrete mixture enhances its mechanical properties. However, the optimal fibre-to-cement ratios differ for various properties: compressive strength (0.3% to 0.4%), tensile strength (0.2% to 0.4%), modulus of elasticity (0.1%), and permeability (0.6%). The overall optimal fibre range is identified as 0.1% to 0.6%, which satisfies all specified criteria. Notably, the inclusion of PF results in a 60% increase in compressive strength, a 115% increase in tensile strength (bending test), a 288% increase in tensile strength (Brazilian test), a tenfold reduction in modulus of elasticity, and a twenty-fivefold reduction in permeability. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Fibre reinforced concrete, Polypropylene, Machine learning, Compressive strength, Tensile strength, Modulus of elasticity, Permeability
in
Results in Materials
pages
16 pages
publisher
Elsevier
external identifiers
  • scopus:105018179345
ISSN
2590-048X
DOI
10.1016/j.rinma.2025.100777
language
English
LU publication?
yes
id
2b126449-fad7-48d0-a838-5b766cea11a9
date added to LUP
2025-10-24 20:58:21
date last changed
2025-11-03 10:48:47
@article{2b126449-fad7-48d0-a838-5b766cea11a9,
  abstract     = {{This study investigates the influence of polypropylene fibres (PF) on concrete performance across a finely resolved dosage spectrum (0.0–2.0% by cement mass, in 0.1% increments) using 2,100 laboratory specimens. The experimental programme evaluated compressive strength, tensile strength (flexural and splitting), modulus of elasticity, and water penetration depth. Predictive modelling was conducted using Random Forests (RF) and Support Vector Regression (SVR), trained and evaluated with simple cross-validation and benchmarked using mean absolute error (MAE) and coefficient of determination (R²). The results reveal distinct optima for strength indices and a threshold behaviour in permeability, with PF dosages between 0.2% and 0.6% balancing mechanical enhancement and substantial reductions in water penetration, although accompanied by a pronounced reduction in elastic modulus at very low PF contents. The RF models exhibited superior predictive performance, consistently outperforming SVR across properties. The experimental outcomes demonstrate that incorporating PF into the concrete mixture enhances its mechanical properties. However, the optimal fibre-to-cement ratios differ for various properties: compressive strength (0.3% to 0.4%), tensile strength (0.2% to 0.4%), modulus of elasticity (0.1%), and permeability (0.6%). The overall optimal fibre range is identified as 0.1% to 0.6%, which satisfies all specified criteria. Notably, the inclusion of PF results in a 60% increase in compressive strength, a 115% increase in tensile strength (bending test), a 288% increase in tensile strength (Brazilian test), a tenfold reduction in modulus of elasticity, and a twenty-fivefold reduction in permeability.}},
  author       = {{Bayat Pour, Mohsen}},
  issn         = {{2590-048X}},
  keywords     = {{Fibre reinforced concrete; Polypropylene; Machine learning; Compressive strength; Tensile strength; Modulus of elasticity; Permeability}},
  language     = {{eng}},
  month        = {{10}},
  publisher    = {{Elsevier}},
  series       = {{Results in Materials}},
  title        = {{From Experimental Studies to Predictive Machine Learning Modelling : Polypropylene Fibre Reinforced Concrete}},
  url          = {{http://dx.doi.org/10.1016/j.rinma.2025.100777}},
  doi          = {{10.1016/j.rinma.2025.100777}},
  year         = {{2025}},
}