Machine learning-driven prediction and digital twins implementation for VOC emissions in VAT photopolymerization additive manufacturing

Abstract With the growing adoption of additive manufacturing (AM) technology across various industries, concerns regarding the possible release of hazardous volatile organic compound (VOC) emissions have surfaced, particularly in VAT photopolymerization (VPP) processes. This study investigates VOC emissions in VPP AM by implementing machine learning (ML) and advanced digital twins to monitor, predict, and mitigate VOC release. An Industrial Internet of Things (IIoT) sensor network, integrated with an Anycubic Mono X 6 K 3D printer, captured data on critical parameters, including layer thickness, exposure time, and light intensity. Subsequent ML model analysis identified exposure time as a principal factor influencing VOC emissions. A Unity-based digital twin was developed to support proactive process optimization, offering real-time visualization and predictive analytics of emission trends. The system aligns with Industry 4.0 objectives, showing considerable potential to enhance operational efficiency and environmental sustainability in VPP AM. This integrated approach significantly advances environmentally responsible AM practices in industrial settings.