Multi-domain fusion meta-model for digital twin in precision manufacturing: design, implementation and verification

Precision manufacturing involves complex multi-physical field coupling, and digital twins combined with AI enable rapid robot programming for customized requirements. However, heterogeneous data sources, incompatible cross-domain models, and dynamic demands hinder system integration. This paper proposes a multi-domain fusion meta-model and virtual-real fusion architecture, featuring a structure-behavior dual-semantic multi-domain ontology association meta-model, a domain-agnostic mapping mechanism, and automated model-to-model/model-to-code transformation algorithms. Validated via a robot elastic material milling system, the architecture achieves a minimum machining error of 0.0027 mm (600 rpm, 0.015 mm feed rate) and optimizes thermal stability through virtual-real closed-loop feedback. It effectively solves core integration challenges in precision manufacturing, providing reliable support for high-fidelity virtual-real mapping and process optimization.