How do you feel about this article? Help us to provide better content for you.
Thank you! Your feedback has been received.
There was a problem submitting your feedback, please try again later.
What do you think of this article?
Digital twinning removing uncertainty from modern system deployments.
Digital twinning is acknowledged as being a major technology disruptor, with numerous types of industrial endeavour set to benefit from its use in the years ahead. By being able to accurately construct intricate virtual representations of real-world systems, then look at what will happen over the course of their lifetime or if certain scenarios occur, a great deal of valuable insight can be derived. Through this approach, it will be possible to gain a better understanding as to how different elements comprised within a complex engineering implementation will interact with one another and also with regard to the environment in which they are situated. Issues that could have the potential to seriously impact on the system’s long-term operational efficiency or its ongoing reliability might thus be identified and then subsequently addressed.
Figure 1: An example of an avionics based digital twin
The basic principles of digital twinning were first outlined by Dr Michael Grieves, an academic then working at the University of Michigan, some twenty years ago. At the time this was only really a concept though, since the computational technology needed to actually create digital twins simply didn’t exist. It was a contemporary of Grieves, John Vickers of NASA who is credited with coining the phrase ‘digital twin’ in his report discussing the subject in 2010. In the past decade, interest in this area has kept on gaining further and further momentum. Analyst firm Markets & Markets now estimates that, by 2025, the global business being generated annually by the digital twinning sector will have reached a staggering $35.8 billion (with other bodies giving similar projections).
One of the places where this technology will also be pivotal is in accelerating design and development work. Following the conventional approach of having to create and then test various different physical prototypes can be extremely time consuming, slowing projects down significantly. Being able to simulate all the facets of these different options and carry out thorough investigations on each of them, then making any necessary modifications, while remaining within the virtual domain throughout is highly advantageous. It means that development cycles will be completed a lot quicker, lowering the human resources and overall engineering effort that have to be assigned to a project. Among the places where development based on digital twins is already starting to be applied is in relation to next generation avionics and automotive ventures. By using this technology, elevated levels of dependability can be achieved, thus assuring the safety of aircraft and automobile occupants.
As well as aiding how systems will manifest themselves before they have been built, digital twins can also be used for systems that are already in existence. In an industrial context, for example, digital twinning may be applied to entire plant installations - providing a platform upon which far more detailed levels of analysis, diagnostics and troubleshooting might be undertaken. This means that it will be at the very foundation of future Industry 4.0 activities, allowing workflows to be enhanced and better optimising important processes. The throughput attained by equipment and the quality of products that such equipment fabricates can both thereby be improved. The consequences that an alteration made at one point might have on other points in the workflow/process downstream may be evaluated before any actions are actually taken. So, for example, it would be possible to play out a scenario where a factory assembly line was run at a more rapid pace. Using the digital twin that had been constructed, it would be possible to look at the implications that could result. This might be in relation to the quality of the product being fabricated, the rate at which components inside the production equipment would fail, or the power consumption involved. From this, it would be possible to make a fully informed decision as to what should be done to get the best outcome - i.e. whether an increase in operational expenses or in product returns would be justified to boost production throughput.
Digital twinning will also permit predictive maintenance to be applied to systems - so the wear on particular constituent components (in relation to different operating conditions, prospective situations, etc.) may be estimated. The upshot of this will be that it is possible to accurately determine when component parts are going to need replacing, meaning that the servicing of equipment can be planned out much more effectively. Hence, the risk of downtime (should a failure occur) or prolonged delays while repair work is carried out, may be avoided.
The essential technology
In order to create a digital twin of a given system, access to huge quantities of data is going to be required. Consequently, arrays of distributed sensors will need to be continuously providing data streams covering various diverse operational parameters. This is why, in many cases, the advent of the Internet of Things (IoT) will be a key enabler for the construction of digital twins. The data utilised can either be collected and then actioned in real time, or in some circumstances historical data might be referred to. The availability of all this data is what differentiates digital twins from conventional computer simulations, as there isn’t any reliance on assumptions about how a system will work, instead everything is constantly being backed up by hard facts. Once datasets have been compiled, sophisticated algorithms can be applied to them.
Artificial intelligence (AI) is another key technology through which data will be leveraged by digital twins. After the necessary data has been acquired, it can then be processed via extensive cloud-based computational resources, with the opportunity for AI models to be subsequently applied. A multitude of different possibilities can then be explored. It could be conceivable to use augmented reality (AR) too - superimposing a digital twin implementation onto a physical system, so that a better understanding of what is happening might be derived. This would allow specialist engineers to advise on the repair or upgrading of equipment in a remote site from a completely different location (with time and money thereby being saved).
Standardisation and interoperability
Creating digital twins in isolation using proprietary technology will not be enough of course, as this will mean that the benefits of integration cannot be realised. There is, as a result, an urgent need for standardisation within this sector. The ISO 23247-1 standard has been formulated accordingly. Via this standard, a regulated framework is provided for digital twin construction within the manufacturing sector. It will enable consistency in terms of the architectures employed by different equipment vendors, so that the digital twins produced for various items found on the factory floor may be combined together.
Bodies like the Digital Twin Consortium, whose membership includes GE, Microsoft, Dell and Ansys, are likewise looking to accelerate the adoption of this technology. They are attempting to do this by promoting the sharing of a common taxonomy across the digital twin landscape, alongside the development of open source code and the establishment of industry recognised application programming interfaces (APIs).
The scope for digital twins to interoperate with one another will help to encourage collaboration between all the different equipment manufacturers and service providers involved in system installations. This will be universally applicable - whether it is in relation to an aircraft turbine, an electric vehicle, a fabrication facility, an oil refinery or the infrastructure within a smart city.