Abstract
The article presents a systematic analysis of modern approaches, methodologies, and tools for creating digital twins of metal-cutting machine tools as a key element of cyber-physical production systems within the Smart Manufacturing paradigm. The relevance of transitioning from isolated technological systems to integrated intelligent environments, where continuous real-time synchronization between production equipment and its virtual representation is ensured, is substantiated. The essence of a digital twin is defined as a multi-physics, multi-scale, and data-oriented model with bidirectional information feedback. A classification of modeling approaches is carried out, including physics-based (White-box), data-driven (Black-box), and hybrid (Grey-box) models. The advantages of the latter for ensuring high accuracy, adaptability, and computational efficiency are demonstrated. The role of multi-physics and multi-scale modeling, as well as the Model-Based Systems Engineering (MBSE) methodology, in creating digital twins of complex mechatronic systems is examined. Architectural approaches to building digital twins are systematized in accordance with the ISO 23247 series of standards, including the levels of physical elements, communication, digital representation, and user services. The stages of creating a digital twin are described in detail: integration of sensor infrastructure, data processing and fusion, construction of mathematical models, implementation of virtual–real synchronization, and validation and self-calibration procedures. Special attention is paid to the role of the Industrial Internet of Things (IIoT), edge and cloud computing technologies, as well as standardized data exchange protocols (OPC UA, MTConnect) in ensuring interoperability and scalability of solutions. Practical cases of using digital twins for optimizing machining processes, reducing vibrations, predicting wear, and implementing predictive maintenance of metal-cutting machine tools are analyzed. The main implementation challenges are identified, including limited access to CNC data, the difficulty of integrating heterogeneous equipment, and the need to adapt models to changes in the state of machine subsystems throughout their life cycle. Prospects for development related to the creation of cognitive digital twins, the use of advanced artificial intelligence models and tools, and the formation of industrial metaverses that will provide a new level of autonomy and production efficiency are outlined.