Buildability-Assessment of 3D Printed Concrete Elements Through Computer-Vision (2024-06)¶

Concrete 3D printing is a digital fabrication technology that has the potential to increase the level of automation in construction. However, getting consistent output quality is a challenge in concrete 3D printing because of the change in material properties with time and the influence of environmental parameters. A robust quality monitoring and control system is required to control the variations and obtain good-quality output. In this study, computer vision techniques are used to monitor the 3D printing process. Image features such as temporal variations in layer thickness and textural changes are used to assess the buildability properties. Two metrics have been developed for quantifying these features: entropy standard deviation and maximum layer thickness deformation. A significant correlation is found between the two metrics, and this relationship can be used to re-confirm the buildability assessment. For a given concrete mix, limiting values can be computed for the metrics to effectively classify an element into a stable type or one that is likely to collapse. This data can also be used as feedback to the printing system to make corrective actions to increase the quality of the print output. Thus, a real-time, nonintrusive buildability assessment system for concrete 3DP elements is demonstrated in this study.