Evaluation of Foamed Concrete Properties Containing Engineered Pozzolans and Compressive Strength Prediction Through Artificial Neural Networks

Abstract The performance of foamed concrete (FC) is significantly influenced by the type and proportion of supplementary cementitious materials (SCMs). However, comprehensive studies investigating the combined effects of engineered pozzolans on the thermal, mechanical, durability, and microstructural properties of FC are limited. This study addresses the scientific problem of optimizing the replacement of Ordinary Portland Cement (OPC) with blast-furnace slag (BFS), silica fume (SF), and surkhi (SK) at varying substitution levels (5–25%) to enhance both thermal and mechanical properties of FC. The scientific novelty lies in the integration of an experimental-analytical framework that combines material characterization, scanning electron microscopy (SEM), and artificial neural networks (ANNs) trained on 570 datasets to predict compressive strength. Sixteen FC mixtures were evaluated for compressive and flexural strength, thermal conductivity, thermal diffusivity, specific heat capacity, water absorption, porosity, and workability. The results indicate that engineered pozzolans significantly improve performance, with the optimum substitution levels identified as 20% BFS, 15% SF, and 10% SK. At 28 days, the 20% BFS mixture enhanced compressive strength by 23.25% (from 17.2 to 21.2 MPa), flexural strength by 26.52%, and reduced thermal conductivity and thermal diffusivity by 27.5% and 9%, respectively, showing enhanced insulating properties. Additionally, the specific heat capacity increased by 7.29%. SEM analysis confirmed denser matrices and improved pore structures, while the ANN model demonstrated high predictive accuracy for compressive strength. This study provides novel insights into the role of engineered pozzolans in FC and offers practical guidelines for developing sustainable, thermally efficient, and mechanically enhanced concrete mixtures.