Dynamics Knowledge Graph-Based Digital Thread for Machine Tool Dynamics Digital Twin

Machine tool dynamic digital twin has the property of being multidimensional and multiscale, subject to the working position and processing speed. Thus, the data-driven twin models exhibit both discreteness and finiteness. It requires a method to fuse various twin models to build a complete digital twin. However, the digital thread-driven digital twin lacks efficient information fusion methods to provide it with evaluation and prediction capabilities. Therefore, the paper proposed the dynamic knowledge graph (DKG)-based digital thread for constructing a complete dynamic digital twin. Dynamic knowledge graph -based digital thread guides the construction of multidimensional and multiscale twin models, and enable information fusion for twin models. At the same time, the knowledge fusion capability based on machine learning provides the knowledge extension for the complete dynamic digital twin (DGT), so that the dynamic digital twin has the ability of prediction and evaluation. It is the first time that we propose the digital thread construction method based on dynamic knowledge graph. This approach has shown excellent predictive performance in dynamic digital twin of the machining process at the machining center. The analysis results show the effectiveness and performance of our proposed method.