Research on Key Technology of Cloud-edge Coordinated Digital Twin Manufacturing Platform Towards Customized Production
Contemporary manufacturing is characterized by large-scale and highly customized production. The diversity of equipment and processes, coupled with fluctuating demand, continually alters the allocation of production resources. Consequently, a more flexible, efficient, and real-time intelligent-manuf...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | Chinese |
| Published: |
Harbin University of Science and Technology Publications
2025-02-01
|
| Series: | Journal of Harbin University of Science and Technology |
| Subjects: | |
| Online Access: | https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2398 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Contemporary manufacturing is characterized by large-scale and highly customized production. The diversity of equipment and processes, coupled with fluctuating demand, continually alters the allocation of production resources. Consequently, a more flexible, efficient, and real-time intelligent-manufacturing paradigm is required. This study targets the perception, decision making, and execution stages of manufacturing and proposes a cloud-edge collaborative digital-twin platform architecture together with an implementation roadmap. By deeply integrating digital-twin technology with an ontology-driven knowledge base and transfer-learningmethods, the platform enables accurate sensing of production entities, data-driven decision-making, and real-time, adaptive execution of control strategies. First, a virtual-physical synchronous digital-twin model is constructed within a hierarchical cloud-edge framework to tackle the dual challenges of dynamic sensing in complex production systems and the fusion of heterogeneous data sources. Second, an ontology-driven methodology is introduced for building a knowledge base whose continuous evolution and optimization broaden the generalizability of knowledge representations and enhance the accuracy of customized process planning. Third, a transfer-learning-based scheme for personalized calibration and real-time adaptation of control strategies is proposed to suppress production disturbances. The paper also summarizes the team’s key advances in each of these domains. Collectively, the methodology delivers a flexible and robust solution for large-scale, customized manufacturing and serves as a reference for deploying intelligent-manufacturing systems in complex, dynamic settings. |
|---|---|
| ISSN: | 1007-2683 |