A Review of Current Progress and Application of Machine Learning on 3D-Printed Concrete

Abstract

3D-printed concrete is a special type of concrete that is digitally fabricated based on 3D-printing technologies without vibration and formwork. Statistical and empirical models have been used to predict the properties of concrete mixtures and structures and support printing processes. However, developing these models requires laborious experimental work and may provide inaccurate results when the complex relationships between the evaluation parameters of concrete mixtures and printed elements. Therefore, machine learning (ML) has become a potential solution in material optimization, manufacturing process management, and behavior prediction for concrete mixes and printed structures. Although advances in ML provide an opportunity to design and optimize 3-D printed structures and materials and achieve more cost-effective and sustainable designs, the number of studies applying ML in 3D printed concrete remains limited. Most of the research on 3-D printed concrete has so far been experimental, with little focus on computational simulations and prediction for the 3-D printing process. This review critically discusses and analyzes the applications of ML and its performance, thereby identifying practical recommendations, current knowledge gaps, and needed future research. © 2024, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.