Paper Contents
Abstract
Abstract :This study investigates the integration of Diorite stone and recycled coarse aggregate (RCA) in rigid pavement concrete, with a focus on mechanical and durability performance validated through machine learning (ML) models. The concrete mix designs were developed at M50 grade, varying RCA (030%) and Diorite Stone (DS) content (060%). Laboratory tests were conducted to assess compressive strength, flexural strength, water absorption, and acid resistance. The mix containing 20% RCA and 30% DS was identified as optimal, demonstrating superior strength and durability.Three ML modelsRandom Forest (RF), XG-Boost, and Multi-Layer Perceptron (MLP) Regressorwere trained and evaluated to predict experimental outcomes. XG-Boost yielded the best performance with R 0.9607 and MSE 0.03, outperforming RF (R 0.9461) and MLP (R 0.1219). Cross-validation standard deviation and model training times were also analyzed. Results confirmed that ML can reliably model complex relationships between input mix proportions and output performance metrics.The proposed approach offers a sustainable and cost-effective methodology for optimizing concrete mixes using waste materials, while reducing the need for extensive physical testing. These findings are relevant for infrastructure development aiming at resource conservation and performance assurance.
Copyright
Copyright © 2025 VIPIN KUMAR YADAV. This is an open access article distributed under the Creative Commons Attribution License.
