Digital twins

The Forecasting Revolution: Digital Twins and the Bottom Line

Credit: Image by Vlad Aivazovsky from Pixabay

A digital twin is a particular type of computer model of a physical system, such as a machine or building, that has the potential for high accuracy, precision, and flexibility to model various aspects of a system. Forecasting is a foundational element across all functions of digital twins. Whether through simulation, monitoring, optimization, or decision support, digital twins rely on models that predict future states, behaviors, or outcomes. 

The prediction abilities of digital twins can aid in reducing inefficiencies and losses. As discussed in NIST AMS 600-16, estimates of the percent of planned production time that is downtime can range from 8.3 % to 13.3 % and amounts to $245 billion in losses just for U.S. discrete manufacturing. Further, defects result in an additional $32 billion to $58.6 billion in losses, again just for U.S. discrete manufacturing. These losses are significant and and many are preventable. Product design, factory optimization, and machinery settings all affect costs that can potentially be reduced by digital twins.

Digital twins primarily function to make predictions or as a status indicator for a system that is being modeled. The benefit is being able to identify more optimal design and/or settings for a particular system, such as when to conduct maintenance or where to place machinery. NIST AMS 100-61 approximates the potential aggregated manufacturing industry benefits of digital twins to be $37.9 billion annually if they are adopted throughout the U.S. manufacturing industry. 

Charting the Digital Twin Frontier: NIST’s Leadership Role

NIST's Digital Twins for Advanced Manufacturing project has invested in research and development to further develop the capabilities of digital twins and demonstrate their implementation. This project provides technical contributions to standards to help manufacturers systematically identify digital twin requirements; formulate the digital twin problem; collect and manage relevant data; develop, validate, and maintain digital twin models; analyze digital twin results; and provide actionable recommendations. The research in this project includes three thrusts. First, the project develops implementation and testing methodologies to help manufacturers create and validate digital twins. Second, the project performs research contributions to relevant standards development and testing. Third, the project establishes, maintains, and utilizes a Digital Twin Testbed to support the R&D of digital twins within NIST.

This Digital Twins for Advanced Manufacturing project aligns with recent advancements in smart sensors, the Industrial Internet of Things, artificial intelligence, and modeling and simulation to realize digital twins of manufacturing systems and processes. Digital twins help observe, diagnose, predict, and optimize the manufacturing system in near real-time and gain the insight needed to decide how to improve overall system performance. A digital twin can help monitor the status, detect anomalies, predict system behaviors, and prescribe future operations. Applications in manufacturing include analyzing machine health, evaluating alternative plans and schedules, setting up maintenance, and performing virtual commissioning.

To address the digital twin complexities, a system of systems approach needs to be taken to integrate and coordinate all appliable subsystems to ensure the value and credibility of digital twins. In addition, to overcome the siloed digital twin challenges, the lifecycle approach also needs to be taken to integrate digital twins for different lifecycle stages. This will provide an integrated view of the physical asset for its digital twin development, avoiding redundancy of information exchange. Combining both systems of systems and lifecycle approaches on digital twins would help establish a marketplace for digital twin users and technology providers and help improve the agility and flexibility of manufacturing systems and the competitiveness of the US manufacturing base. Linked below are products of this project.

Driving the Next Generation of Digital Twins