Performance optimization of eco-engineered waterproof concrete blocks using machine learning and industrial by-products

The construction industry is a significant contributor to environmental degradation, primarily due to the extensive use of conventional concrete, which is associated with high carbon emissions and resource depletion. This study explores the development of sustainable waterproof concrete blocks through the partial replacement of natural fine aggregate by mass with industrial steel slag and reinforcement using polypropylene fibers (PPF) and high-density polyethylene (HDPE) sheets. HDPE sheets were integrated as effective barriers against water ingress, significantly enhancing the waterproofing and long-term durability of the blocks. The mechanical properties, durability, and environmental impact were thoroughly evaluated. The experimental program demonstrated that Mix C1 (10% slag replacement) achieved the highest 28-day compressive strength of 49.67 MPa, representing a 16.2% gain over the control mix (42.75 MPa). Mix C2 (15% slag replacement) achieved optimal durability and sustainability, reducing water absorption by 23.5% (from 8.5% to 6.5%) and the Global Warming Potential (GWP) by 10.9% (from 320 to 285 kgCO2​ eq./m3). Advanced machine learning models (Random Forest and XGBoost) were developed; the XGBoost model proved to be the most effective, demonstrating superior predictive accuracy across all metrics, achieving R2 values consistently above 0.90 for all target metrics, with a peak R2 of 0.96 for compressive strength. The integration of experimental data, environmental metrics, and predictive modeling establishes a holistic framework for producing eco-efficient, high-performance concrete blocks, providing a sustainable approach toward reducing the ecological footprint of the construction sector.