An article written by Wlodek Mista
What is a Digital Twin?
According to the Digital Twin Consortium established in 2020, A Digital Twin is a virtual representation of real-world entities and processes, synchronized at a specified frequency and fidelity.
- Digital twin systems transform business by accelerating holistic understanding, optimal decision-making, and effective action.
- Digital twins use real-time and historical data to represent the past and present and simulate predicted futures.
- Digital twins are motivated by outcomes, tailored to use cases, powered by integration, built on data, guided by domain knowledge, and implemented in IT/OT systems.
The concept has been known for many years and has been used extensively during the space flight projects, especially the space shuttle. The use of digital models of spacecraft components was the only way to remotely diagnose and find solutions once the spacecraft was in space.
As the computing power of computers and the capacity of databases increased, it became more efficient and more accessible. Nowadays, we have the possibility to collect all the historical data about an object and, thanks to this, we can simulate and check the impact of the changes made on the functioning of the object or the whole process. This has a key role to play in reducing risks and minimizing the costs of implementing changes, especially in manufacturing companies.
Digital Twin can be used to simulate multiple options for potential changes, optimize, and select the best solution. This allows us to avoid many unforeseen problems with the production implementation and the consequences of directly implement an incorrect solution in production.
On the other hand, we must not forget the limitations of the Digital Twin solution. A virtual model will only be as good as the good quality data that feeds it. Classical methods of data collection are fraught with technical limitations – the quantity, quality of sensors and their costs but also the difficulty of interpreting the data and assessing their real impact on simulation results.
Very high computing power and adequate disk resources are required to analyze and store the data.
With help comes AI and ML. With these tools, the Digital Twin concept can be made significantly more accessible. It is much simpler, cheaper, and therefore more accessible to interpret the data, obtain the valuable information necessary for proper simulation and assess the impact of changes on the process. Thanks to the accessibility of AI and ML through cloud solutions, as well as the removal of significant computing power constraints, even smaller companies can now harness the advantages of these technologies. It’s no longer exclusive to the largest corporations.
Digital models of real objects can be used to:
- Visualize process data in an accessible way.
- Testing different solutions and making changes to products.
- Testing equipment prototypes for functionality, robustness, and user expectations.
- Analyzing errors and inconsistencies in design concepts and eliminating them from the earliest stage of a commercial project.
- Carrying out operator training before the device is built in the factory.
How do you create a digital twin?
A digital twin is a combination of the virtual world and reality. Its model consists of a physical object, a digital representation, and a combination of the two, which takes place through real-time data exchange.
What is the process of creating a digital twin?
Firstly, experts must prepare a model that carefully analyses the functioning of the physical device and learn the principles of how it works. This information is used to create a mathematical model that operates in a digital space that replicates the environment of the system or device in question.
The digital twin then receives data such as the parameters of the mathematical model from, for example, sensors and sensing devices that surround the physical version of the device. The data can be collected in real time, analyzed, and visualized. Once the tool has been developed, it is possible to test the operation in practice to verify the correctness of the model. Importantly, we do not need to create a digital twin of an existing device – we can use a prototype. This will give us information that will help us improve the first version. More and more companies are choosing to first create a first prototype as a virtual model. The physical prototype, on the other hand, is created at the next stage.
It is up to us what the complexity of our digital twin will be. We can create a simple model or design, or a more complex one. However, it is important to remember that the accuracy of the data on the performance of the physical counterpart will depend on how much information and data is fed into the system.
According to Forbes, the market size for digital twins exceeded $5 billion in 2020 and is expected to grow at a compound annual growth rate of over 35 percent between 2021 and 2027. A few examples of digital twins developed in recent years:
- The human brain: The EU-funded Neurotwin project aims to simulate specific human brains to build models that can predict the best treatments for conditions such as Alzheimer’s and Epilepsy. Clinical trials using the model are due to start in 2023.
- Every Tesla ever sold: Tesla creates a digital simulation of all its cars, using data collected from sensors on the vehicles to upload to the cloud. This allows the company’s AI algorithms to determine where faults and breakdowns are most likely to occur and minimize the need for servicing.
- Los Angeles transportation: The Los Angeles Department of Transportation has partnered with the Open Mobility Foundation to create a data-driven digital twin of the city’s transport infrastructure. It will model the movement and activity of micro-mobility solutions, such as the city’s network of shared-use bicycles and e-scooters.
- 3D-printed bridge: The steel bridge in central Amsterdam is remarkable for two reasons: it’s the first pedestrian bridge to be entirely constructed via 3D printing and it has a digital twin. A network of sensors, placed across the structure, gather data that is used to build the twin. The data can then be used to analyze its performance as it comes under stress during everyday use.