Hybrid Knowledge-Based Digital Twin for Cutters of Machine Tools

Abstract

It is necessary to conduct a real-time, intelligent, and self-adaptive computer numerical control machine tool system due to its basic executor role in the manufacturing process. The digital twin (DT), evolved from the cyber-physical system, provides a solution to this demand. However, most commercial DT software focuses on simulation, lacking online diagnosis and prediction of the manufacturing process. Meanwhile, conventional research concentrates on component modeling of machine tools rather than cutters and workpieces, leading to poor correlation to product quality. Furthermore, tool wear compensation is seldom considered for quality improvement. To address these issues, this article proposes a hybrid knowledge-based DT for cutters of machine tools. A novel DT architecture of a machine tool is demonstrated and explained first. Then, the implementation methods for how to construct digital shadow via the machining process, generated and real-time measured signals, to integrate knowledge from physics and data domain for diagnosis and tool wear prediction, and to compensate for the tool wear for machining process online optimization are discussed. Finally, a case study shows the shadow fidelity, prediction accuracy, and compensation effectiveness. Results show that the proposed method can be implemented in the practical machining process and improve the workpiece quality efficiently.