AI-Augmented Predictive Quality Control in Additive Manufacturing Supply Chains

Abstract

Additive manufacturing (AM) has transformed the present-day supply chains with on-demand manufacturing, adaptive design, and digital inventory management. The distributed and heterogeneous nature of AM networks, however, induces inherent problems with stable product quality across distributed networks. Reactive and inspection-based quality control practices common with traditional quality control prove insufficient to handle variability in processes, material variability, and machine variability characteristic of AM networks. This paper proposes a conceptual framework of AI-augmented predictive quality control (PQC) applicable to additive manufacturing supply chains. The framework uncovers multi-faceted components: in-situ monitoring-based real-time data capture, AI-based analytics for defect prediction and anomaly recognition, decision support systems for adaptive intervention, and closed-loop continuous feedback facilities aided with digital twins. With the help of machine learning, deep learning, and reinforcement learning-based strategies, predictive frameworks are able to forecast defects, reduce scrap to a minimum, and transform supply chains with increased resiliency. The paper also elaborates on the theoretical benefit of AI-augmented PQC, including improved traceability, economy of costs, and increased congruity with just-in-time logistics. Data heterogeneity, scalability, cybersecurity, and workers' adaptability are also paramount challenges discussed in the paper. The future research directions are also enumerated in terms of hybrid AI-physics models, standardizable datasets, integration with blockchain, and human-AI teaming. This research enlists the transformative potential of AI-augmented PQC in making AM supply chains more reliable and viable.