Credit:
NIST AMS 100-61
A digital twin is a 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 the system. They are used in five primary areas based on the sales of software for implementation: predictive maintenance (39.9 %), business optimization (25.3 %), performance monitoring (17.8 %), inventory management (11.9 %), and product design and development (3.4 %). The remaining applications represent 1.6 % of the sales. Digital twins primarily function to make predictions or as a status indicator for the system being modeled. The benefit of the broader category of data tracking and analytics, which includes digital twins, 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. Digital twins provide the potential for high level accuracy, precision, and flexibility in data tracking and analytics, where flexibility is the model’s ability to consider different types and levels of input and output factors. The cost effectiveness of investing in a digital twin is likely affected by the complexity and sensitivity (i.e., the level of system variation that affects economic outcomes) of the real-world system being modeled along with the cost consequence of having the non-optimal level of settings or design for the system.
The total benefits of all future data tracking and analytics investments in the U.S., including those for digital twins and those with less precision, accuracy, and flexibility, is estimated to be $88.6 billion. If digital twins account for data tracking and analytics investment costs that are above the 85th percentile, the total potential impact of the adoption of digital twins in the manufacturing industry is estimated to be $37.9 billion. Anecdotally, it is common for the highest performing category within a group to account for between the top 10 % and 20 %; thus, the 85th percentile is a significant but reasonable assumption, given the patterns in the costs and return-on-investment found in other similar investments that are discussed in this report. A Monte Carlo simulation varying key factors of this estimate by -50 % and +20 % (i.e., biasing it downwards) and assuming that digital twins account for between the 80th and 95th percentile of data tracking and analytics investments by cost, puts the 90 % confidence interval between $16.1 billion and $38.6 billion with a median of $27.2 billion annually. These industry level estimates are based on a number of datasets and calculations, including tendencies or patterns in the relationship between the costs and returns on investments entered in the Department of Energy’s (DOE) Industrial Assessment Center data. From the industry estimates in this report, one could reasonably surmise that the potential impact of digital twins is likely in the low tens of billions of dollars. Reasonable assumptions are made to calculate the estimates above and these assumptions are relaxed using a Monte Carlo approach. Despite these best efforts, there is a wide range of error in the estimates. Future research could increase the accuracy and precision of these estimates by collecting additional data from manufacturers. Understanding the potential impact affects the investment analysis of public investment in advancing digital twins and their adoption.
For more information, please see NIST AMS 100-61:
Credit:
NIST AMS 100-61