Ever since the publication of a paper by Michael Grieves, the concept of ‘digital twin’ has caught the imagination of those who engineer systems and release them into the wild.
The appeal is clear - a simple metaphor for how we can stay connected to and optimise the things we have engineered and built. Recent advances in sensors, cloud and ‘computing at the edge’ have made the vision of optimised performance, extended life or reduced maintenance costs more achievable than ever, and great strides are being made across a number of domains, from the ‘poster child’ wind turbine through to buildings, roads and ‘smart cities’.
However, there is another dynamic at work in the market – the shift from the provision of an engineered product that provides a functional effect, to an engineering-based service that provides a business outcome. Even a solidly engineering-based company like BAE Systems now earns more revenue from their services than from the products themselves – think in terms of a customer buying ‘flying hours’ rather than ‘aircraft’. And stand-alone wind farms are evolving into hybrid solutions that combine the products and services of multiple technologies to achieve greener energy outcomes.
So in this case, can I extend digital twin thinking to reflect these kinds of outcomes? And if so, what exactly should my digital twin be a twin of?
The answer is, in a way, simple - the real world ‘thing’ is simply the operation around the product that makes it a service: the supply chain, the support operations, the information that flows through these operations, the people that run these operations, the decision points and arguments that govern and assure delivery, the skills of the people, the training that provides these skills, the commercial agreements that provide a sustainable delivery of value …
The traditional digital twin modelling and control architecture still applies at this scope, but with each component having its own techniques and approach:
The model: the asset model at the heart of this twin is not a ‘bill of materials’, but a connected and ‘ontological’ business architecture
The data: data pipelines that feed the twin are engineered less for speed and bandwidth of IoT sensor data than for transformation of and inference from a wide variety of data sources
Learning and prediction: prediction is less to do with a learnt model of behaviour based on repeatable families of product-in-an-environment, and more to do with the operational flow of information, demand and authority, and learning how best to impact the speed, efficiency and value of a service
There are challenges, of course. Being only partly engineered, the real world operation that this all reflects changes in ways that are not entirely under our control; and the need to manage uncertainty and risk is a given.
But the good news is that you are probably close to having some of these building blocks in place already. With a grasp of the operating model of the business, data pipelines that connect into this, and with insights into the dynamics that drive outcomes, you can start to evolve a digital twin that reflects and optimises the services that a customer values.
Simon Smith IIG MooD is the Chief Architect at CACI Information Intelligence Group.
If you would like to learn more about techUK’s work in this area, or learn about our new Digital Twins Working Group (DTWG), then please feel free to get in touch with Tom Henderson (Tom.Henderson@techUK.org) today or follow our @techUK handle on Twitter & LinkedIn! You can also visit techUK's #DigitalTwinFuture campaign here!