Verification of Numerically Controlled Manufacturing Processes, Toward Identifying Cyber-Physical Threats

Motivated by cyber-physical vulnerabilities in precision manufacturing processes, there is a need to externally examine the operational performance of computer numerically controlled (CNC) manufacturing systems. The overarching objective of this work is to design and fabricate a proof-of-concept CNC machine evaluation device, ultimately re-configurable to the mill, lathe, and 3D printing machine classes. This device will assist in identifying potential cyber-physical security threats in manufacturing systems by identifying perturbations outside the expected variations of machining processes and comparing the desired command inputted into the numerical controller and the actual machine performance (e.g., tool displacement, frequency). In this paper, a requirement-driven prototype device design method is presented and tested, and the results will be used to improve future iterations of the design. The first design iteration is tested on a Kuka KR 6 R700 series robotic arm, and machine movement comparisons are performed ex situ using Keyence laser measurement sensors. Data acquisition is performed with a Raspberry Pi 4 microcomputer, controlled by custom, cross-platform python code, and includes a touch screen human–computer interface (HCI). A device design adapted for a CNC mill is also presented, and the Haas TM-2 is used as a case study, which can be operated by CNC operators to assess machine performance, as needed, before a critical manufacturing process. This research will enable the broader goal of recognizing repeating advanced manufacturing performance deviations, which occur outside of a machine's normal variability, that could signal a malicious “attack.”