Data-driven additive manufacturing with concrete: Enhancing in-line sensory data with domain knowledge, Part II: Moisture and heat

Abstract

A data-driven approach to achieving first-time-right manufacturing in digital fabrication with concrete (DFC) relies on in-line sensors to capture real-time measurements. To extract meaningful information (features) from raw sensory data, these sensors must be integrated with knowledge-driven feature engineering (KDFE) strategies. The first paper describes the methodology and its application to geometric data. This paper continues by detailing features related to the dosing, mixing, and pumping of material, the quantification of surface dehydration by employing near-infrared spectroscopy, the measurement of surface temperature using infrared thermography, and the monitoring of atmospheric conditions. Although these in-line sensors provide valuable data, much of it is complex or indirectly related, which requires KDFE. The sensors and features are demonstrated using real-world data, while controlled laboratory experiments are used for their validation. In combination with the features presented in the first paper, this paper offers a comprehensive set of features for 3DCP process monitoring.