Task incremental learning-driven Digital-Twin predictive modeling for customized metal forming product manufacturing process
Jie Li,Zili Wang,Shuyou Zhang,Yaochen Lin,Lanfang Jiang,Jianrong Tan
DOI: https://doi.org/10.1016/j.rcim.2023.102647
IF: 10.103
2023-09-09
Robotics and Computer-Integrated Manufacturing
Abstract:Customized metal forming products entail personalized requirements in terms of dimensions, materials, and other specifications, while the processing conditions involved are subject to dynamic changes. Digital-Twin (DT) predictive models have become essential tools for optimizing the complex manufacturing process. However, the traditional approach exhibits limitations in handling dynamic data, capturing complex nonlinear relationships, and leveraging multi-source information. Additionally, retraining predictive models for novel tasks with unique operating conditions in specific scenarios can lead to substantial time and resource inefficiencies. Therefore, a task incremental learning-based approach for DT predictive modeling is proposed in this paper. Firstly, a DT framework and a comprehensive information model are established for real-time monitoring and integration of multi-source information. Moreover, the pre-trained Temporal Fusion Transformer model is utilized to capture valuable knowledge from historical tasks. Subsequently, task incremental learning is adopted to fine-tune the model using new task data, thereby enhancing adaptability and enabling rapid and scalable modeling. Finally, the effectiveness of the proposed method is validated on a customized metal tube bending forming platform, demonstrating accurate prediction of tube cross-section deformation.
robotics,computer science, interdisciplinary applications,engineering, manufacturing