Recipe Based Anomaly Detection with Adaptable Learning: Implications on Sustainable Smart Manufacturing

The advent of Industry 4.0 has significantly transformed the manufacturing sector, bringing advancements in quality control efficiency, environmental sustainability, and production development etc. These changes have led to the development of intelligent technologies such as artificial intelligence(AI). However, implementing AI solutions in manufacturing processes still faces challenges in many aspects, particularly in handling irregular datasets influenced by diverse manufacturing settings. Especially in the field of injection molding, quality inspection often occurs in a batch level rather than singular data, providing only the overall defective ratio of the batch production instead of each individual labeling result. These issues limit the general application of AI and data-driven decision making. To address these limitations and promote product efficiency, this study proposes a novel anomaly detection framework for a specific manufacturing process. For Recipe-Based Learning, we first apply K-Means to take the flexible manufacturing process relying on diverse settings into account. The injection molding data are classified into specific-setting based recipes to secure data normality and uniqueness. Then, autoencoders for anomaly detection are trained with normal data from each recipe. With this data-driven AI approach, 61 defective products are predicted compared to the existing 41 defects. Meanwhile, the integrated model without considering the change of settings only predicted 2 defects which imply a poor and distorted quality inspection. For Adaptable Learning focused on new inputs with unseen settings, we apply KL-Divergence to find the closest trained recipe data and its model. In this case, it also exceeded the predictivity compared to the integrated and additionally trained models. This leads to continuous prediction without further training, which also successfully implies the enhancement of process optimization. In the aspect of smart factory in the injection molding field, such improvement of process management can first improve the overall productivity and decision making mainly with data-driven AI approach. Furthermore, it is expected to enhance energy efficiency such as reduction of environmental waste and computational cost, ensuring sustainable manufacturing.