Digital Twin of Cyber-Physical CNC for Smart Manufacturing

Due to the complexity of smart manufacturing systems, digital twins (DTs) can help manufacturers test and validate products as well as determine any process failures before production. Computer numerical control (CNC) machines are widely used and employed in several manufacturing industries because of their precision and efficiency. Despite the wide adoption of CNC machines, there is currently no CNC-based DT that simulates an entire production process. This paper demonstrates the implementation and validation of a DT that can 1) replicate the physical manufacturing operations of products on Tormach CNC machines by accurately generating the positional values of the machine's tool path along the X, Y, and Z axes, 2) employ MTConnect's communication protocol by utilizing an adapter and an agent that produces data such as timestamps, axis positions, and spindle speed into MTConnect's standardized XML format, 3) deploy an MTConnect utility, dataminer, to collect data from the agent and properly store it into a JSON format for further analysis. For validation purposes, an example coin geometry was simulated on the DT and manufactured in aluminum on a real Tormach machine. A comparison between the simulated operations and real system was made to validate the accuracy of the DT. Dynamic time warping (DTW) was used to compare process data from the DT with data from the real machine. Our analysis shows that the DT and real system have a high correlation as the DTW warping path is closely aligned at a 45 degree angle line and has minimal time warping.