“A digital twin is a virtual representation of an object or system that spans its lifecycle, is updated from real-time data, and uses simulation, machine learning, and reasoning to help decision-making.”
In plain English, this just means creating a highly complex virtual model that is the exact counterpart (or
twin) of a physical thing. The thing could be a car, a building, a bridge, a jet engine, or even a whole city.
Connected sensors on the physical asset collect data that can be mapped onto the virtual model. Anyone looking at the digital twin can see crucial information about how the physical thing is doing out there in the real world.
It is, in essence, a computer program that uses real-world data to create simulations, which can predict how a product or process will perform. These programs can integrate the internet of things (IoT), industrial internet of things (IIoT), artificial intelligence (AI), and software analytics to enhance the output. It uses virtual and augmented reality (AR) as well as 3D graphics and data modeling to build a virtual model of a process, system, service, product, or other physical objects. With the advancement of machine learning and factors such as big data, these virtual models have become a staple in modern engineering to drive innovation and improve performance.
The digital twin can be broken down into three broad types, which show the different times when the process can be used:
- Digital Twin Prototype (DTP) – This is undertaken before a physical product is created
- Digital Twin Instance (DTI) – This is done once a product is manufactured in order to run tests
on different usage scenarios - Digital Twin Aggregate (DTA) – This gathers DTI information to determine the capabilities of a
product, run prognostics, and test operating parameters
These overarching types can offer a variety of uses including logistics planning, product development and re-design, quality control/management, and systems planning. So, it can be used to save time and money whenever a product or process needs to be tested, whether in design, implementation, monitoring, or improvement.
Is digital twin a simulation?
Although simulations and digital twins both utilize digital models to replicate a system’s various processes, a digital twin is actually a virtual environment, which makes it considerably richer for study. The difference between a digital twin and a simulation is largely a matter of scale. While a simulation typically studies one particular process, a digital twin can itself run any number of useful simulations in order to study multiple processes. The differences don’t end there. For example, simulations usually don’t benefit from having real-time data.
With a simulation, engineers can run tests and conduct assessments on a simulated version of a physical asset. The simulation is static, however. In other words, it doesn’t keep pace with the physical asset unless the engineer inputs new parameters into the simulation.
A digital twin, on the other hand, receives real-time updates from the physical asset, process, or system. Therefore, the tests, assessments, and analysis work conducted by engineers are based on real-world conditions. As the state of the digital twin dynamically changes as it receives new data from the physical world, it matures and produces outputs that are more accurate and valuable.
By having better and constantly updated data related to a wide range of areas, combined with the added computing power that accompanies a virtual environment, digital twins are able to study more issues from far more vantage points than standard simulations can — with greater ultimate potential to improve products and processes.
How does digital twin work?
The life of a digital twin begins with experts in applied mathematics or data science researching the physics and operational data of a physical object or system in order to develop a mathematical model that simulates the original.
The developers who create digital twins ensure that the virtual computer model can receive feedback from sensors that gather data from the real-world version. This lets the digital version mimic and simulate what is happening with the original version in real-time, creating opportunities to gather insights into performance and any potential problems. The twin could also be designed based on a prototype of its physical counterpart, in which case the twin can provide feedback as the product is refined; a twin could even serve as a prototype itself before any physical version is built.
The skill sets are demanding, and require specialized expertise in machine learning, artificial intelligence, predictive analytics and other data-science capabilities.
A digital twin design is made by gathering data and creating computational models to test it. This can include an interface between the digital model and an actual physical object to send and receive feedback and data in real-time.
Data
A digital twin requires data about an object or process in order to create a virtual model that can represent the behaviors or states of real-world objects or procedures. This data may relate to the lifecycle of a product and include design specifications, production processes, or engineering information. It can also include production information including equipment, materials, parts, methods, and quality control. Data can also be related to the operation, such as real-time feedback, historical analysis, and maintenance records. Other data used in digital twin design can include business data or end-of-life procedures.
A digital twin comprises three main elements:
- Past data – historical performance data of individual machines, overall processes, and specific systems.
- Present data – real-time data from equipment sensors, outputs from manufacturing platforms and systems, and outputs from systems throughout the distribution chain. It can also include outputs from systems in other business units, including customer service and purchasing.
- Future data – machine learning as well as inputs from engineers.
Modeling
Once the data has been gathered it can be used to create computational analytical models to show operating effects, predict states such as fatigue, and determine behaviors. These models can prescribe actions based on engineering simulations, physics, chemistry, statistics, machine learning, artificial intelligence, business logic, or objectives. These models can be displayed via 3D representations and augmented reality modeling in order to aid human understanding of the findings.
Linking
The findings from digital twins can be linked to create an overview, such as by taking the findings of equipment twins and putting them into a production line twin, which can then inform a factory-scale digital twin. By using linked digital twins in this way it is possible to enable smart industrial applications for real-world operational developments and improvements.