Title:
Computer Vision to Predict Interlayer Bond Strength from Early-Age Properties
Author(s):
Rakesh Khan
Publication:
Web Session
Volume:
ws_S25_RakeshKhan.pdf
Issue:
Appears on pages(s):
Keywords:
DOI:
Date:
3/30/2025
Abstract:
Interlayer bonding in 3D concrete printing is significantly affected by time between layers and changing site conditions, including temperature, humidity, wind, and sun exposure. The variability introduced by these factors forces designers to heavily penalize these bond strengths when designing structural elements printed in the field, despite the high strengths often achieved. A computer vision sensor at the nozzle that can identify the effects of these dynamic construction conditions on each bead and predict the bond strength that will be achieved with the next bead will enable 1) documentation of as-built bond strengths, 2) user feedback to adjust processes for better bonding, and 3) intelligence for future closed loop systems to automate the production of elements with reliably high bond strengths. With this reliability, designers can take topology optimization to a new level for large field-printed structures, with substantial potential to reduce material quantity, construction time, and carbon footprint. This presentation discusses the basis of this computer vision approach: multispectral and hyperspectral imaging, the methods of mechanically testing bond strength of hardened printed walls, and the correlation algorithms under development as part of ongoing research in collaboration with multiple 3DCP practitioners and material suppliers. The presentation closes with a roadmap of future efforts planned to make such a technology economically feasible for the industry.