Prediction of the Compressive Strength of Alkali-Activated Mortar for Block-Work Binders Using Adaptive Neuro-Fuzzy Inference System

This study employs an Adaptive Neuro-Fuzzy Inference System (ANFIS) to predict the compressive strength of alkali-activated mortars (AAMs) designed for blockwork applications. The model, configured with 225 fuzzy rules, evaluates key mix parameters including activator ratio, alkali concentration, and liquid-to-solid ratio, optimizing strength predictions. Experimental validation confirms that ANFIS achieves high accuracy, with an RMSE of 0.97 and an a20-index of 0.74, indicating strong predictive reliability. The research utilizes waste cellular lightweight concrete (CLC) as a precursor, demonstrating its feasibility for sustainable construction. By reducing the need for extensive laboratory testing, ANFIS improves mix design efficiency, making AAMs more viable for commercial applications. The findings highlight the effectiveness of ANFIS in optimizing AAM formulations, ensuring a balance between material efficiency, sustainability, and structural performance. This research supports the transition toward more eco-friendly, high-performance construction materials, facilitating the broader adoption of alkali-activated mortars in real-world applications.