PSI - Issue 70

Available online at www.sciencedirect.com

ScienceDirect

Procedia Structural Integrity 70 (2025) 424–431

Structural Integrity and Interactions of Materials in Civil Engineering Structures (SIIMCES-2025) Enhancing Concrete Properties through Supplementary Cementitious Materials and Predictive Modeling R. Ashwathi a* , R. S. Soundariya b , M. Nivaashini c , R. M. Tharsanee d , A. Dinesh e

a Department of Civil Engineering, Bannari Amman Institute of Technology, Sathyamangalam 638401, India b Department of Computer Technology, Bannari Amman Institute of Technology, Sathyamangalam 638401, India c Department of Artificial Intelligence and Data Science, Sri Eshwar College of Engineering, Coimbatore 641202, India e Department of Civil Engineering, Kumaraguru College of Technology, Coimbatore 641049, India e Department of Civil Engineering, Sri Ramakrishna Engineering College, Coimbatore 641022, India

Abstract The growing demand for high-performance and sustainable concrete necessitates the incorporation of supplementary cementitious materials (SCM) into the concrete matrix to reduce cement consumption and mitigate the associated carbon footprint, a substantial contributor to greenhouse gas emissions. The proposed research explores the utilization of different industrial and agricultural byproducts including fly ash, silica fume, metakaolin, and rice husk ash, as cementing material to improve performance and sustainability. These materials exhibit higher pozzolanic behavior by reducing the porosity and contribute to the strengthening of the Interfacial Transition Zone leading to improved strength. Mechanical properties such as compressive, flexural, and tensile strength are obtained. Machine learning (ML) techniques reduce the need for extensive experimental trials and streamlines the process. The proposed research adopts a Graph Neural Network (GNN) model that analyzes experimental data, gets trained with the laboratory results and predicts the mechanical properties of concrete and identifies key factors influencing concrete performance. Testing results indicate that the GNN model exhibits higher 2 values and lesser statistical error values when compared to the other existing models in the literature. This clearly implies that advanced ML models like GNN can be utilized in feasible, efficient and rapid prediction of the strength properties of concrete. © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under the responsibility of International Conference on Structural Integrity Organizers

Keywords: Concrete; Supplementary cementitious materials; Strength prediction.

* Corresponding author. Tel.: 7904056447 E-mail address: ashwathi@bitsathy.ac.in; ashwathisingh@gmail.com

2452-3216 © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under the responsibility of International Conference on Structural Integrity Organizers 10.1016/j.prostr.2025.07.073