How machine learning can transform the future of concrete

Abstract

The concrete industry is confronted with persistent challenges, such as the need for extensive experimentation, time limitations, and high costs. Machine learning (ML) has become an extremely useful tool, providing diverse applications to tackle these challenges. This paper reviews the growing influence of ML on the concrete industry, highlighting its potential to revolutionize different aspects of concrete research and practical applications. The review explores the evolution of ML in this field, identifying key techniques, algorithms, and data sources commonly used in concrete related studies. It discusses the diverse applications of ML, including material characterization, mix design optimization, prediction of concrete properties, enhancement of nonlinear finite element analysis, crack detection, improvements in sustainability, and structural health monitoring. Additionally, the paper addresses challenges faced in the implementation of ML and offers recommendations to enhance its accuracy and effectiveness for concrete researchers, engineers, and practitioners.

Graphical abstract