Machine learning-driven analysis of nanoparticle performance on concrete mechanical properties

Nanoparticles as raw material additive are substances that modify the concrete product. This study presents a comprehensive analysis of nanoparticle effects on concrete mechanical properties using advanced machine learning (ML) algorithms. We examine various nanoparticle types, including multi-walled carbon nanotubes (MWCNTs), graphene nanoplatelets (GNPs), nano-SiO 2 (silica), and nano-TiO 2 (titanium dioxide), investigating their impact on concrete’s flexural ( f b ), compressive ( f c ), and tensile ( f t ) strengths. We use ML algorithms such as decision tree (DT), Pearson correlation coefficient, and the hierarchical clustering algorithms to analyze their mechanical properties. Results show that there is a significant increase in mechanical strength when nanoparticles are incorporated into concrete. For example, adding nano-Fe 2 O 3 (iron oxide) can increase the control concrete sample of f c and f b from 105 MPa to 140 MPa and 16 MPa to 23 MPa, respectively. The study identifies five primary enhancement mechanisms: filler effect, nucleation site provision, pozzolanic reaction, nano-reinforcement, and C-S-H structure modification. However, Pearson correlation analysis reveals significant inconsistencies in strength improvements, with correlation coefficients ranging from 0.87 for tensile-compressive strength relationships to −0.26 for flexural strength improvements. The DT analysis reveals that nanoparticle concentration is the decisive factor in determining the improvement of concrete strength. On the other hand, the hierarchical clustering analysis identifies distinct groupings of nanoparticles based on their enhancement mechanisms, with MWCNTs forming an independent cluster due to their unique concrete f b (23 MPa) and f c (140 MPa) strengths. In addition, the cost analysis reveals that nanoparticle additions can improve concrete qualities, but their selection and dosage optimization should be considered to balance performance increases with economic viability in practical use. This research provides useful information for developing optimized nanoparticle-enhanced concrete formulations while highlighting the complexity of strength enhancement mechanisms.