AI ‐powered development of sustainable lightweight partition wall systems using neural networks

Abstract The construction industry is increasingly transitioning toward sustainable and energy‐efficient materials. Lightweight expanded polystyrene (EPS) concrete, with its superior thermal insulation, soundproofing, and lightweight properties, is an emerging alternative for block walls. In this study, an artificial neural network (ANN) model was developed to predict the compressive strength of EPS concrete, addressing the challenges of material variability and design optimization. A comprehensive experimental program consisting of 30 mixes was conducted to calibrate and validate the ANN model by leveraging parameters such as the water content, EPS, ordinary Portland cement (OPC), and air content. The results demonstrate the high accuracy of the ANN model ( R 2 >0.98, RMSE≈0.12 MPa, and MSE≈0.014 MPa), underscoring its reliability in capturing the nonlinear relationships between input variables and compressive strength. The study further proposes a precast EPS sandwich panel system optimized for strength, thermal performance, and ease of installation, achieving low U‐values of approximately 0.66 W/m 2 K for a 100 mm panel and 0.42 W/m 2 K for a 150 mm panel. This study highlights the potential of EPS concrete in advancing nearly zero‐energy buildings, offering a scalable solution for reducing energy consumption and environmental impact in the building sector. Future developments that integrate advanced coatings can further enhance the sustainability and thermal efficiency of the system.