AI-assisted Concrete Mix Design for 3D Construction Printing
We are currently developing an AI-powered mix design tool specifically for 3D printable concrete. This innovative tool utilizes data-driven algorithms to optimize and predict key properties of concrete mixes, particularly focusing on strength. By inputting specific mix compositions, the tool can accurately forecast the resulting material strength, helping engineers and researchers to design high-performance, customized mixes efficiently.
Our first iteration of the tool has demonstrated promising results, with a high level of accuracy in predicting concrete strength based on mix designs. This marks a significant step toward reducing trial-and-error approaches in concrete formulation, thus saving time and resources.
The methodology and performance results, as illustrated in the accompanying image, show the reliability and potential of this tool in real-world applications. We are committed to continuously enhancing this project, incorporating more data, refining our algorithms, and expanding the tool's capabilities to include additional properties beyond strength, such as durability, workability, and sustainability metrics. This ongoing development will further improve the tool's robustness and applicability across a wide range of construction and engineering projects.
The development of this tool is supported by NSF as a part of larger project, OpenMatFlo.
J. Gao, C. Wang*, J. Li, S.H. Chu (2024) Data-driven rheological model for 3D printable concrete. Construction and Building Materials, 447, 137912, https://doi.org/10.1016/j.conbuildmat.2024.137912
J. Gao, C. Wang*, S.H. Chu (2024) Mix design of sustainable concrete using generative models. Journal of Building Engineering, 110618, https://doi.org/10.1016/j.jobe.2024.110618
···
