Data-driven fault diagnosis has received significant attention in the era of big data. Most data-driven methods have been developed under the assumption that both training and test data come from identical data distributions. However, in real-world industrial scenarios, data distribution often changes due to varying operating conditions, leading to a degradation of diagnostic performance. Although several domain adaptation methods have shown their feasibility, existing methods have overlooked metadata from the manufacturing process and treated all domains uniformly. To address these limitations, this article proposes a weighted domain adaptation method using a graph-structured dataset representation. Our framework involves encoding a collection of datasets into the proposed graph structure, which captures relations between datasets based on metadata and raw data simultaneously. Then, transferability scores of candidate source datasets for a target are estimated using the constructed graph and a graph embedding model. Finally, the fault diagnosis model is established with a voting ensemble of the base classifiers trained on candidate source datasets and their estimated transferability scores. For validation, two case studies on rotor machinery, specifically tool wear and bearing fault detection, were conducted. The experimental results demonstrate the effectiveness and superiority of the proposed method over other existing domain adaptation methods.
5 December 2025 in The International Journal of Advanced Manufacturing Technology
Continuous workpiece localization is critical for ensuring traceability, quality control, and process coordination in hot forging environments, where heavy workpieces are repeatedly transferred by specialized handling equipment such as overhead cranes among storage yards, furnaces, and forging stations. However, traditional workpiece-centric approaches are often unreliable under extreme temperatures, surface degradation, and irregular routing conditions, leaving a critical gap in continuous localization. To overcome these challenges, this study presents an equipment-centric framework that infers workpiece locations indirectly by analyzing equipment operations captured through video streams from multiple static 2D cameras. This paradigm shift from workpiece-centric to equipment-centric localization enables robust inference under harsh conditions by coupling equipment behavior with workpiece-handling events. The framework estimates floorplan-space 3D coordinates of handling equipment and recognizes workpiece-handling activities such as grasp and release. Event-driven finite state machines (FSMs) validate and detect discrete handling events based on equipment coordinates and recognized activity cues, enabling continuous inference and updating of workpiece states and floorplan-space coordinates. Experimental validation in an operational hot forging factory demonstrated the effectiveness and practicality of the proposed framework, achieving 100 % event detection accuracy within a 33-second tolerance window, a mean localization error of 317.8 mm, and a mean system latency of 21 seconds. Ablation studies further confirmed that integrating the Keypoint-Guided Attention (KPGA) mechanism improves activity recognition performance compared to baseline 3D convolutional neural networks (3D-CNNs) and transformer-based models. Beyond localization accuracy, the structured outputs of the framework bridge vision-based perception and event-driven reasoning by representing handling operations as interpretable state transitions. This capability enables data-driven visualization of workpiece transfers and quantitative evaluation of equipment handling states, contributing to more intelligent and traceable forging operations.
26 June 2026 in Journal of Intelligent Manufacturing