Anomaly Detection for Catalyzing Operational Excellence in Complex Manufacturing Processes: A Survey and Perspective

The evolution of manufacturing processes, fueled by the transition towards Industry 5.0, has ushered in an era marked by the integration of sophisticated technologies such as the Industrial Internet of Things (IIoT), artificial intelligence (AI), and machine learning. This transformation has spawned smarter, more efficient production environments that, while advantageous, have led to the emergence of complex anomalies and deviations that traditional anomaly detection (AD) models struggle to effectively address. This chapter delves into the complexities of smart manufacturing, proposing advanced AD methodologies designed to cope with the challenges posed by the integration of cutting-edge technologies. These challenges include handling the vast volume, speed, variety, accuracy, and the inherently noisy, non-stationary, and high-dimensional nature of smart manufacturing data. By transcending simple data storage and analysis, the proposed approach links data directly to operational processes, allowing for a deeper understanding of end-to-end manufacturing activities. This not only speeds up the detection of anomalies but also aids in pinpointing the process parameters that cause quality deviations, bridging the gap between conventional models and the nuanced requirements of smart manufacturing.