Evaluation of net-zero materials in mortar bricks with predictive modelling using random forest and gradient boosting techniques
G. Uday Kiran, N. Ganesan, Dipankar Roy, Sumant Nivarutti Shinde, M. Indumathi, George Uwadiegwu Alaneme, Val Hyginus Udoka Eze, Kuzmin Anton
Abstract
This study investigates the potential of incorporating alternative waste materials—limestone (LS) powder and construction and demolition waste (CDW)—in sustainable brick production under real-time curing conditions. Cement was partially replaced with LS powder in the range of 0–20% by weight, while natural aggregates were substituted with CDW at 0–100%. Experimental evaluations were conducted to assess the mechanical properties of the fabricated bricks. The results demonstrated a notable enhancement in performance compared to conventional bricks, with a maximum compressive strength of 14.99 N/mm 2 observed at 2% LS and 10% CDW replacement levels. Additionally, a significant reduction in water absorption was recorded. Machine learning models, including Random Forest and Gradient Boosting, were employed to predict mechanical behavior. The findings underscore the feasibility of producing eco-efficient bricks by valorizing industrial waste within a circular construction framework.