Knowledge-based Digital Twin System: Using a Knowlege-Driven Approach for Manufacturing Process Modeling
Chang Su,Yong Han,Xin Tang,Qi Jiang,Tao Wang,Qingchen He
DOI: https://doi.org/10.1016/j.compind.2024.104101
IF: 10
2024-01-01
Computers in Industry
Abstract:The Knowledge-Based Digital Twin System is a digital twin system developed on the foundation of a knowledge graph, aimed at serving the complex manufacturing process. This system embraces a knowledge-driven modeling approach, aspiring to construct a digital twin model for the manufacturing process, thereby enabling precise description, management, prediction, and optimization of the process. The core of this system lies in the comprehensive knowledge graph that encapsulates all pertinent information about the manufacturing process, facilitating dynamic modeling and iteration through knowledge matching and inference within the knowledge, geometry, and decision model. This approach not only ensures consistency across models but also addresses the challenge of coupling multi-source heterogeneous information, creating a holistic and precise information model. As the manufacturing process deepens and knowledge accumulates, the model's understanding of the process progressively enhances, promoting self-evolution and continuous optimization. The developed knowledge-decision-geometry model acts as the ontological layer within the digital twin framework, laying a foundational conceptual framework for the digital twin of the manufacturing process. Validated on an aero-engine blade production line in a factory, the results demonstrate that the knowledge model, as the core driver, enables continuous self-updating of the geometric model for an accurate depiction of the entire manufacturing process, while the decision model provides deep insights for decision-makers based on knowledge. The system not only effectively controls, predicts, and optimizes the manufacturing process but also continually evolves as the process advances. This research offers a new perspective on the realization of the digital twin for the manufacturing process, providing solid theoretical support with a knowledge-driven approach.