Digital Twins, Explained in 1 Level of Difficulty
And what the Hong Kong Airport, Manchester City, and Chattanooga are using them for.
In 2023, McKinsey reported that 70% of C-suite technology executives (ie. Chief Technology Officers or CTOs) at large enterprises were exploring and investing in digital twins. The football club Manchester City is using Sony’s Hawk-Eye to also create a digital twin of its stadium and those in it. So, what is a digital twin, how does it work, and why should you spend 3 minutes reading about it?
(Because they’re good. See, Medium, I’m promising to provide value to the reader!)
At its most basic level, a digital twin is a replica of a system — a wind turbine, a factory, or even a car. Data from sources on the actual/physical system (e.g. sensors, LiDAR, photogrammetry) is taken in via a processing system and used to perform simulations, study performance issues, etc. on the digital twin. From this, insights or improvements can be identified and applied to the actual system.
For example, at the Hong Kong Airport, manager Francesco Tizzani, says a digital twin had data on “when we last drilled a pile, poured concrete, struck the formwork, or did a quality check.” This data was used to generate animations on the digital twin “for erecting the scaffold, breaking the ends of the footbridges, and removing the debris.”
The benefits across industries vary, though some examples are more effective Research and Development (digital twins are a cheaper and more effective product testing option because it’s…digital. There are no materials.), increasing revenues by 5–10% in some categories, and 25% fewer quality issues (Geez, how many quality issues did we have to begin with?)
IBM notes that digital twins are also more advanced than simulations, as they have more computing power, can study multiple processes, and leverage real-time data. This real-time data allows for a two-way flow of information as the data is given to the system processor and then insights are shared back with the original source.
In the case of the Hong Kong Airport, the hope is to eventually have a smart airport, where a digital twin will pick passengers up so they can head straight to the departure lounge (or browse shops and eat) without check-ins. As of now, they’re using a digital twin to build Terminal 2 for the airport.
In Chattanooga, Georgia, US, digital twins are being used to examine the safety of traffic intersections by mimicking near-miss activities (e.g. when a biker just barely misses an accident with a car, miscommunication at intersections, or whenever I’m on the road). Some are hoping though that this can be scaled for bigger problems, such as reducing construction costs for cities, identifying the risks of building on certain terrains, and even monitoring air pollution.
As of now, Microsoft and Nvidia have entered the market (which is expected to reach $73.5 billion by 2027) with Azure Digital Twins and the Nvidia Omniverse platform.
Note: Digital twins can be hard to integrate and create (they take 3–6 months to build), so they may start out as a data product (e.g. data that just optimizes an employee’s schedule) before capabilities are expanded (e.g. a real-time AI that examines and improves an employee’s performance as a whole).
