Machine learning approach for predicting the compressive strength of biomedical waste ash in concrete: a sustainability approach

Abstract

Abstract This study investigates the effects of strength and durability of concrete for different water-cement ratios, aggregate contents, and partial replacement of biomedical waste ash at 5%, 10%, 15%, 20%, and 25% by weight of cement. At 7, 14, and 28 days, the control mix showed inferior mechanical properties, particularly compressive strength, compared to concrete mixtures containing Biomedical Waste Ash (BWA). The replacement of cement by 5% and 10% increased the compressive strength but it is decreasing from 15%. Additionally, BWA modified concrete demonstrated a slower water absorption rate and minimal weight loss under acid test curing conditions, indicating enhanced durability. The economic and environmental benefits of incorporating biomedical waste into concrete promote sustainable construction practices. Using three machine learning approaches—K-Nearest Neighbors (KNN), Random Forest (RF), and CatBoost—the compressive strength of concrete with biomedical waste ash was simulated. Cement, biomedical waste, water absorption, slump, and the water-to-cement ratio were key input variables. Among the models tested, the RF model emerged as the most accurate, with a predictive performance of R 2 = 0.9945 and RMSE = 0.7080. Its unparalleled reliability, consistency, and accuracy in predicting compressive strength make it a top choice for this task.