Evolution of Cement Concrete Modeling: A Review of Conventional Modeling Techniques and Application

Abstract

Modeling of cement concrete has been on the rise over the past few decades due to the growing need to understand its constituents, improve its properties and predictability, and reduce failure risks. With increasing housing and infrastructure demands, cement concrete modeling techniques have evolved from experimental, mathematical, and statistical models to computerized approaches including machine learning. This review presents studies on cement concrete modeling to identify research gaps and foster improvement. It employs scoping literature review and social network analysis using VOS Viewer version 1.6.20. Publications were accessed from Google Scholar and Dimensions data bases covering the period from 2000 to 2024, and analyzed with a focus on the years 2015 to 2024. The results show a notable rise in cement concrete modeling, with computerized methods especially machine learning-gaining prominence due to their superior accuracy and ability to capture complex behaviors. However, few studies consider detailed material characteristics such as aggregate size, distribution patterns and water pH as inputs in predictive models. The study advocates for incorporating more detailed material input parameters in modeling to enhance the accuracy and reliability of cement concrete performance predictions. A comparative study based on multiple model types and techniques can also yield better results.