Exploring elastic properties of fly ash recycled aggregate concrete: Insights from multiscale modeling and machine learning

Abstract

The primary focus of this study is to analyze the impact of different components of fly ash recycled aggregate concrete (FARAC) on its elastic properties, aiming to suggest it as a potential alternative to traditional concrete. For this study, a variety of techniques were employed, such as multiscale modeling, finite element methods, and Mori-Tanaka theory. The study employed supervised machine learning methods to conduct classification and regression analyses, with the goal of investigating the impact of mix proportion on elastic properties. The results obtained from the machine learning models suggest that the XGBoost and K-Nearest Neighbors algorithms have shown exceptional precision when compared to the other algorithms, achieving an accuracy of over 0.9. The current investigation reveals that the elastic characteristics of FARAC are primarily affected by fly ash and calcium silicate hydrate (C-S-H) in comparison to other components. According to the research findings, incorporating recycled aggregate content in the range of 0–100% resulted in a 25% decrease in Young's modulus and an 8% increase in Poisson's ratio of FARAC. In addition, two web-based applications have been developed to calculate the elastic properties of composite materials with spherical inclusions using the Mori-Tanaka theory, and to predict the elastic properties of FARAC through machine learning.