Digital Twins
We have already encountered the concept of digital twins, when we discussed your digital shadow in the first chapter.
As we touched on there, digital twins are digital counterparts of something that exists in reality—whether they be physical objects, places, processes or people. Another way to look at it is that the digital twin is a simulation of what happens in the given system, process, or object.
It's worth pointing out that this isn't a specific technology, but rather a way to apply technology. The digital twin can be based on computational models, machine learning algorithms, sensor data, and other data.
Think of it as a way to gather all the data you have about a process or an object, so that you can use this to monitor what's happening—and to predict what's going to happen.
How do digital twins work?
The basic idea behind digital twins is as follows:
- You have an object or a process (theoretical or real) that you're interested in observing, optimising, testing, predicting the behaviour of, or similar.
- You create a virtual, digital representation of the object or process in question.
- In many cases, you may want to enable an exchange of data between the digital twin and the relevant object, process or system.
When the twin and its real-world counterpart share data, this may allow tests, simulations and monitoring to be done largely on the digital twin instead of the real thing, saving money and reducing risk.
In industry, this is used to optimise production and streamline operations, for example by running tests without interrupting operations, or by simulating possible outcomes to make better decisions and optimise processes. In other words, both saving money, and reducing the impact on the climate and the environment.
Digital models of physical things are by no means a new phenomenon. But the fact that a digital twin utilises data from the physical world for analysis—often with the help of machine learning—means that the twin is constantly learning from the real objects, often in real time.
What are digital twins used for?
The power industry is among those that use digital twins extensively. Statnett, the system operator in the Norwegian energy system, for example, feeds digital twins with real-time data on electricity consumption in Norway to ensure that they can deliver power. They also need good models to regulate a market that can be very unpredictable, so they can make better decisions.
The technology can also be used in the development of physical objects: By using digital twins in prototyping, you can run tests virtually—and avoid challenges and costs associated with physical prototypes. Car manufacturers, for example, use sensor data from cars on the road to develop everything from better self-driving technology to battery saving and better traction.
But it's not just large businesses that could benefit from using digital twins. Imagine that you can save electricity by running simulations of the consumption in your home. Then you can see how your routines affect your use of electricity, and add variables such as outdoor temperature to make more cost-saving choices.
The technology is also used by researchers, for example those working with issues related to climate and the environment. Here, digital twins of natural ecosystems, the sea, or, the entire world, for that matter, can be created. This can be done in order to monitor and simulate different scenarios around, for example, extreme weather, light pollution and global warming.
To sum up, we can say that digital twins are a good way to ensure that we really take advantage of simulations, modelling, computation, the Internet of Things, and data streaming (which you will learn more about in later chapters). It's also a good tool to use with artificial intelligence.