Intelligent multi-objective optimization of 3D printing low-carbon concrete for multi-scenario requirements

This study presents, for the first time, a universal theoretical framework based on machine learning (ML) algorithms for the multi -objective optimization (MOO) design of 3D printing low-carbon concrete to meet the sustainability requirements of the construction industry. The study focuses on the simultaneous reduction of carbon emissions and costs during the design process. Specifically, the applicable scenarios are primarily classified into two categories: meet aesthetics requirements (printing layer height error-HE and printing layer width error-WE approaching 0) and strengths requirements (compressive strength-CS approaching 60 MPa and interlayer bonding strength-IBS as high as possible). Through a comparison with the traditional design method, 3D printing concrete based on the intelligent design method achieves a 44% reduction in costs and a 19% reduction in carbon emissions while meeting aesthetics requirements. Under strengths requirements, there is a 57% reduction in costs and a 22% reduction in carbon emissions. This demonstrates the indispensable and crucial role of the MOO method employed in this study in achieving the lowest costs and carbon emissions. Additionally, the results of experimental verification demonstrate that the models developed in this study have successfully achieved a close alignment between design values and experimental values within the allowable error range (The errors are 3.46%-9.83%). This not only holds the potential to drive the widespread application of 3D printing technology in the construction industry but also presents new possibilities for cleaner production, contributing to the achievement of sustainable development goals.