V3I7P85

Strength Prediction of Recycled Aggregate Concrete Using Machine Learning Techniques

Aakanksha Panditrao Ingle^1*,  V. B. Chavan²  , S. P. Sharma³

Abstract

The increasing depletion of natural aggregates and environmental concerns associated with construction and demolition (C&D) waste have accelerated the development of sustainable recycled aggregate concrete (RAC). This study investigates the mechanical properties, microstructural characteristics, and machine learning-based strength prediction of RAC containing 0–100% recycled coarse aggregate (RCA). Sixty concrete mixtures with varying cement contents, water–cement ratios, and superplasticizer dosages were experimentally evaluated. Results indicated improved workability with higher cement content but reduced mechanical strength with increasing RCA replacement. Among the prediction models, Extreme Gradient Boosting (XGBoost) achieved the highest accuracy, demonstrating that machine learning offers a reliable and cost-effective approach for predicting RAC compressive strength.

Keywords:

construction and demolition (C&D) waste, ML, Strength, concrete