ORNL/TM-2026/4400 Machine tool data analytics for digital twin and machine predictive maintenance Akash Tiwari, ORNL Jaydeep Karandikar, ORNL Vladimir Orlyanchik, ORNL Chris Hamlin, Scytec Josh Davids, Scytec January 2026 CRADA FINAL REPORT NFE-23-09833 DOCUMENT AVAILABILITY Reports produced after January 1, 1996, are generally available free via US Department of Energy (DOE) SciTech Connect. Website www.osti.gov Reports produced before January 1, 1996, may be purchased by members of the public from the following source: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone 703-605-6000 (1-800-553-6847) TDD 703-487-4639 Fax 703-605-6900 E-mail info@ntis.gov Website http://classic.ntis.gov/ Reports are available to DOE employees, DOE contractors, Energy Technology Data Exchange representatives, and International Nuclear Information System representatives from the following source: Office of Scientific and Technical Information PO Box 62 Oak Ridge, TN 37831 Telephone 865-576-8401 Fax 865-576-5728 E-mail reports@osti.gov Website http://www.osti.gov/contact.html This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. ORNL/TM-2026/4400 CRADA/NFE-23-09833 Manufacturing Science Division Energy Science and Technology Directorate MACHINE TOOL DATA ANALYTICS FOR DIGITAL TWIN AND MACHINE PREDICTIVE MAINTENANCE Akash Tiwari, ORNL Jaydeep Karandikar, ORNL Vladimir Orlyanchik, ORNL Chris Hamlin, Scytec Josh Davids, Scytec January 2026 Prepared by OAK RIDGE NATIONAL LABORATORY Oak Ridge, TN 37831-6283 managed by UT-BATTELLE, LLC for the US DEPARTMENT OF ENERGY under contract DE-AC05-00OR22725 CONTENTS CONTENTS .................................................................................................................................................iii ACKNOWLEDGEMENTS ..........................................................................................................................iv ABSTRACT...................................................................................................................................................5 1. SUMMARY...........................................................................................................................................5 1.1 Multiple Sensing Modality...........................................................................................................5 1.2 Method and Application for toolpath level analytics...................................................................6 1.3 Real-time data monitoring in a modernized GUI.........................................................................7 2. CONCLUSIONS ...................................................................................................................................8 iii ACKNOWLEDGEMENTS This CRADA NFE-23-09833 was conducted as a Technical Collaboration project within the Oak Ridge National Laboratory (ORNL) Manufacturing Demonstration Facility (MDF) sponsored by the US Department of Energy Advanced Materials and Manufacturing Technology Office (CPS Agreement Number 24761). Opportunities for MDF technical collaborations are listed in the announcement “Manufacturing Demonstration Facility Technology Collaborations for US Manufacturers in Advanced Manufacturing and Materials Technologies” posted at http://web.ornl.gov/sci/manufacturing/docs/FBO- ORNL-MDF-2013-2.pdf. The goal of technical collaborations is to engage industry partners to participate in short-term, collaborative projects within the Manufacturing Demonstration Facility (MDF) to assess applicability and of new energy efficient manufacturing technologies. Research sponsored by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Advanced Material and Manufacturing Technology Office, under contract DE-AC05-00OR22725 with UT-Battelle, LLC. iv ABSTRACT The primary objective of this project is to improve machining process performance using in-process machining data from the machine tool controller and external sensors. Advances in the Industrial Internet of Things (IIoT) enable monitoring of machines using controller data. Examples of the data provided by a controller include execution status of the controller, part count, block of code being executed, door status, tool position, the spindle and axis load, etc. MTConnect and OPC-UA are the two common protocols for capturing machine information. In this collaboration, methods for retrieving the machine controller data from selected machine tool controls and making these data accessible in different subsystems (such as digital twins and machine maintenance portals, etc.) will be developed and tested. In addition, analytics to improve machining process performance (by increasing productivity and reducing downtime) will be developed. 1. SUMMARY Phase I of the project successfully established the foundations for data collection from multiple machines. The success was demonstrated by application of the collected data to monitor machine productivity at the part level by identifying partial and full runs of the cycle in a discrete part manufacturing scenario. In phase II, external high-speed vibration sensor was introduced, and the analytics was further advanced from the part-level to the toolpath level. 1.1 MULTIPLE SENSING MODALITY A vibration sensor was identified, procured and installed to the stationary casing of the spindle illustrated in Fig. 1(a). For safety a separate insulated electrical box was prepared to house the PLC which is responsible for collecting and controlling the sensor acquisition start and stop. The insulated electrical box is illustrated in Fig. 1(b). Figure 1. Vibration sensor installed. (a) Sensor placement on spindle (b) Insulated electrical box housing the PLC with logic for sensor data acquisition and coordination with controller PC. Three data modalities in total became available for advanced data analytics. The three data modalities include: (1) Machine data served by standard industrial protocols like MTConnect and OPC-UA (2) External vibration sensor (3) Part program 5 The additional sensor installed is an independent sensing system and was therefore required to be integrated with the machine data to provide synchronized. The synchronization was performed using a common clock for timestamping the Machine data and vibration sensor data. The analytics developed in phase I can summarize part-cycles as a complete or a partial run and the time taken to complete the runs using only the machine data. In comparison, the analytics in phase II uses all three data modalities to provide a detailed breakdown by G-code and toolpath. A comparison of the advancement in analytics is illustrated in Fig. 2 using only Machine data and external vibration sensor. Figure 2. Comparison of analytics details in phase I which used only machine data to advanced analytics in phase II which uses machine data and vibration data. (a) Phase I analytics (b) Phase II analytics which shows breakdown of vibration data distribution and time-to-complete breakdown by blocks in the G-code. It was noted that when using only machine data augmented with external sensor data, certain machines, which have slow sampling rate for the machine data in the sub-10 Hz range, up to 90% of the machine information can be lost due to high-speed movements. This limit is noted by the illustration in Fig. 2(b) where duration and vibration information is organized based on only the sampled G-code. This limitation significantly reduces the resolution of information available. 1.2 METHOD AND APPLICATION FOR TOOLPATH LEVEL ANALYTICS A method was developed to address the limited sampling rate of machine controller data by incorporating contextual information from the part program. The method enables reconstruction of toolpath-resolved process information even when controller data is sampled at sub-10 Hz, where a large fraction of executed G-code blocks may not be directly observed. The method was evaluated using experimental data collected from two CNC machines – a Makino a51nx and a DMU 60 eVo – executing a common test geometry (Fig. 3(a)(i)) and corresponding toolpath (Fig. 3(a)(ii)). Raw machine controller data, external vibration data, and the part program were combined to organize process information in the context of the programmed toolpath. As shown in Fig. 3(b)(i), the raw machine data is sparse and non-uniform, providing limited visibility into toolpath execution. 6 Figure 3. Experimental demonstration of the developed method for toolpath-level analytics. (a) Test setup consisting of (i) a representative test geometry and (ii) the corresponding programmed toolpath executed on a Makino a51nx and a DMU 60 eVo CNC machine. (b) Comparison of raw and processed machine data organized in the context of the toolpath: (i) sparse and non-uniform raw machine controller data sampled at sub-10 Hz, (ii) upsampled toolpath-level representation showing duration spent along the toolpath, (iii) upsampled toolpath-level representation showing spindle load distribution, and (iv) visualization of feedrate override application along a different toolpath segment. The results illustrate the significant increase in effective toolpath resolution and the ability to localize process variables along the programmed path. Application of the developed method produces an upsampled, toolpath-level representation that substantially increases the effective resolution of the available data. Figures 3(b)(ii) and 3(b)(iii) illustrate the ability to visualize duration and spindle load continuously along the toolpath, while Fig. 3(b)(iv) demonstrates localization of feedrate override behavior on a different toolpath segment. Compared to the raw controller data, the resulting representations provide a consistent and detailed view of machining behavior along the programmed path. These results demonstrate that the developed method advances analytics from part-level summaries to localized toolpath-level insights, supporting improved process understanding and diagnostics in machining operations. 1.3 REAL-TIME DATA MONITORING IN A MODERNIZED GUI In Phase II of the project, the data collection developed using Scytec DataXchange were demonstrated through a modernized, real-time graphical user interface (GUI) designed for shopfloor use. The GUI was implemented to visualize live machine data streamed simultaneously from multiple CNC machines, enabling intuitive monitoring of machine status and operational context in a single environment. The demonstration included four CNC machines – DMU 60 eVo, Makino a51nx, HAAS TM-0P, and a FANUC-controlled concrete base machine – represented within a virtual shopfloor layout illustrated in Fig. 4. Each machine was displayed with real-time color-coded operating status (running, idle, or stopped), along with key execution information such as the active program and current G-code line (Fig. 4(a)). This provided immediate situational awareness at the shopfloor level. 7 Figure 4. Modern GUI for machine tool monitoring using Scytec DataXchange as data source. (a) Virtual shopfloor environment with four machines and color indicating operating status of the machine (Green – running; Yellow – idle; Red – Stopped) and live information on current program being executed and instantaneous G-code Line being executed. (b) Detail view of the Makino a51nx is displayed when clicked/touched to reveal additional real-time details about the machine. This picture illustrates a 10 second history of the spindle load (c) A real-time instance of the GUI running in a touchscreen monitor located in the shopfloor environment with detail view of the DMU 60 eVo showing Y-position. In addition to high-level monitoring, the GUI supported machine-level inspection through interactive selection. Upon selecting a machine, users could access short-term historical trends of critical variables such as spindle load and axis position, enabling rapid assessment of machine behavior during operation (Fig. 4(b)). The system was deployed on a touchscreen interface, allowing shopfloor personnel to interact directly with the visualization without specialized training (Fig. 4(c)). 2. CONCLUSIONS Overall, this demonstration showed that real-time machine data served through Scytec DataXchange can be effectively integrated into an intuitive, modern GUI to enhance operational visibility, support decision- making on the shop floor, and provide a foundation for future integration of advanced analytics developed in this project. 8