Realizing real-time monitoring of the machining state and product quality during the manufacturing process is essential for ensuring part quality and improving production efficiency. Machining state fluctuations often indicate production process instability, resulting in increased costs, including accelerated tool wear and degradation of workpiece quality. However, the complexity and time-varying nature of machining processes make it challenging for traditional methods to achieve real-time, accurate monitoring. This paper establishes a data-mechanism hybrid-driven digital twin system (DMH-DTS) framework through the internal real-time data of the machine tool to achieve high-precision, real-time visualization monitoring of machining states and quality without the need for any additional sensors. The composition and working principle of the DMH-DTS are elaborated in detail, and the key enabling technologies of the DMH-DTS are introduced. The motion model and environment model of the DTS, alongside the geometric model, material model, and behavior model of the material removal process are established from perspectives of the machining mechanism and data fusion, respectively. Through the mutual flow of energy, material, and information in the DMH-DTS, large-scale multi-source heterogeneous data fusion is realized, ensuring real-time symmetry between the physical and virtual systems. The cross-scale online visualization monitoring method of the grinding state proposed in this paper enables real-time monitoring of both the grinding state and the shape accuracy of the workpiece. The visual prediction of workpiece machining accuracy can be achieved in the crossscale range (mm, mu m, and nm). Furthermore, this method facilitates the traceability of machining errors, enabling the identification of the causes of workpiece errors, thereby providing a theoretical basis for further improving product accuracy and quality.
