Summary
A digital twin is a true-to-reality simulation of physics and materials — of a real-world physical asset or system, which is continuously updated. A digital twin can be a digital replica of an object in the physical world, such as a jet engine or wind farms, or even larger items such as buildings or even whole cities. It’s useful to break down the concept into level of product magnification: component twins, asset twins, system twins, and process twins. It is a concept rather than a specific technology and a digital twin requires multiple technologies including sensors, a wireless network, 3D modelling software, machine learning and others.
Viability (5)
The idea of digital twin technology was first voiced in 1991 by David Gelernter in his book “Mirror Worlds”. However, in 2002, Dr. Michael Grieves publicly introduced the digital twin concept in relation to Product Lifecycle Management (PLM). Digital twins are at commercialisation stage with a wide variety of vendors and customers across industries. All of the major cloud providers including AWS, Azure, and GCP offer Digital Twin tools as does NVIDIA making adoption easier. R&D is focused on increasing granularity and accuracy of the model as well as improving the prediction algorithms for specific deployments e.g. predictive maintenance for a factory or wake effect simulation for wind farm placements. The market was worth $7.5 billion in 2021 and expected to grow a massive 40% to over $180+ billion by 2030. With a 40% growth rate and adoption across almost all industries. A catalyst for faster growth came in 2022 when NVIDIA demonstrated its Omniverse digital twin and world simulation platform with clients including Amazon, Siemens Gamesa, Kroger, Pepsi and Lowes. In 2022 AWS with IoT TwinMaker, Google with Supply Chain Twin, and Microsoft with Azure Digital Twins have all beefed up their digital twin functionality.
Drivers (4)
Supply-side, digital twins are an application only possible with the advent of the Internet of Things. For digital twins to be useful, any deployment needs lots of sensors, a low-latency and high-bandwidth wireless network, and a lot of computing power and accurate prediction algorithms. Digital twins require application-specific software and therefore it’s only in the last few years as costs have fallen that it has been economical for vendors to serve specific industries or use cases (e.g. retail floorspace optimisation). On the demand-side, digital twins are part of the same digital transformation trend driving adoption of digital tools to increase efficiency. Increasing efficiency is always on the agenda somewhere.
Novelty (5)
There aren’t alternatives to digital twins as such. It solves the specific problem of replicating what is happening to a physical asset in the real-world. Simulation is complementary and helps to answer what might happen, but cannot answer what is happening. A future workflow for the production of digital and physical goods might include simulations pre-production and a digital twin post-production.
Diffusion (4)
Digital twins were at the top of the hype cycle in 2018 and have seen adoption over the past 4 years across a wide variety of industries. The first customers were in manufacturing, aerospace and transportation tapping into the Industry 4.0 trend, but now there are customers in every industry. The industry advocacy group the Digital Twins Consortium now has over 159 members including every major manufacturing company in the world including Northrop Grumman, Autodesk, and GE Digital. Digital twin usage does require a lot of preparatory work and looks closer to a 24-month implementation than a SaaS offering. Digital twins require business unit collaboration and multi-stakeholder input and as such adoption takes time. Beyond strategic vision and alignment, the interconnected nature of digital twins and they fact it sits between functions and business units means it requires new skills which need to be required or trained. However these costs can not be concretely compared against value with vendors now having reference examples and case studies yo point to to start ramping up sales with things like 15% increase in sales, turnaround time and operational efficiency and a 25% improvement in system performance.
Impact (4)
Digital twins are important interfaces between the world of atoms and bits. The ability for humans to map, simulate and monitor in real-time everything in the physical world is an inevitable end state. The journey will be from low-fidelity and not in real time, to high fidelity, real-time digital twins application by application starting for high-value manufacturing goods in aerospace, defence and transportation. Long-term, digital twins are part of a broader set of technologies that include simulation software, 3D Printing, Augmented Reality and Metaverse which are blurring the boundaries between digital and physical worlds.
Sources