Reinforcement bond performance in 3D concrete printing: Explainable ensemble learning augmented by deep generative adversarial networks

Integrating various reinforcements into 3D concrete printing (3DCP) is an efficient method to satisfy critical requirements for structural applications. This paper explores an explainable ensemble machine learning (EML) method to predict the bond failure mode and strength of various reinforcements in 3DCP. To overcome the problem of insufficient experimental data for the 3DCP technology, a deep generative adversarial network (DGAN) is integrated for data augmentation. The trained EML models augmented by the DGAN accurately and reliably evaluated the bond performance. The relative embedded length, concrete cover, and compressive strength of 3D printed concrete were identified as the three most important features based on the explanatory analysis. The effect of the amount of training data and integration of the proposed method in the additive manufacturing process were also investigated. The proposed method combining EML, data augmentation, and model interpretation can be extended for other aspects of digital fabrication with concrete.