Digital Transformation

Digital Twins

Get To Know Your Digital Twins

What if analysis was always on, near-real-time and could provide automated, instant feedback?

Rob Gear and Peter Durant from PA Consulting Group explain the value of creating digital versions of real-world objects. . .

Previous waves of digital disruption have been characterised by replacing an existing product with a digital version that can be produced and distributed at lower cost. For an example that shattered a long-established industry, look no further than the digitisation of print media.

As digital disruption continues to transform more and more industries, one of the hot trends for 2018 is ‘digital twins’, with digital enhancing physical products and processes instead of completely replacing them.

The term ‘digital twin’ is widely credited to Dr Michael Grieves, who used it in the 1980s while researching product lifecycle management at the University of Michigan. However, its true origins trace back to NASA’s work in pairing digital objects with physical ones in the early days of space exploration. At the time, this was achieved solely through simulation and modelling and did not benefit from communication between pairs of physical and digital objects.

Today, a digital twin is a near-real-time digital representation of a physical product, process or system. Internet of Things (IoT) sensors ensure changes in the physical system are immediately reflected in the digital twin, which is carefully analysed to improve performance of the actual product.

A hot trend for 2018

Why should we care about this? Firstly, serious industrial players are investing big money and resources. Siemens has invested over $10bn in its digitalisation efforts in the last decade, acquiring TASS International, MRX Technologies Group, Mentor Graphics and CD-adapco in the last two years alone. And it’s not the only one to do so. In the last two years, GE has spent more than $1bn on Meridium, Wise.io, Bit Stew, ServiceMax and IQP Corporation in a bid to strengthen its own digital capabilities.

Then there is the trend towards ‘servitisation’. Driven by concerns about only being seen as suppliers of commodity products, organisations are increasing looking for ways to provide ‘products-as-a-service’ (page 40). Consider how established manufacturing companies Rolls Royce and GE have been exploiting digital twins to make a success of their aero engine business models. Indeed, Rolls Royce no longer sells engines and it is now a ‘power-as-a-service’ providers, charging aircraft owners for each hour of powered flight.

Thirdly, now’s the right time for digital twins. The key facilitating technologies are falling into place and Industry 4.0, with the explosion of the IoT, is ready with the remote sensing capabilities and near-real-time data needed to link physical assets to digital models. Costs continue to fall for the computing power, data storage, bandwidth and data communications needed to handle everything. We are also seeing growing confidence in the ability of machine learning to mine and analyse the mountains of data being produced.

Finally, there’s a lot of hype fuelling interest in digital twins. We don’t want to add to that hype, although we do believe there are a couple of uses where digital twins have potential for many organisations and which will become viable soon…

1. Maintenance and servicing

Consider the digital twin of a car. It can be fed with near-real-time data from IoT sensors on the vehicle and monitored remotely. The monitoring system can look at many digital twins at the same time, all of which are fed with data from many similarly-equipped cars. It can learn how cars are used, how they wear out over time and how this can lead to degraded performance and component failure.

An intelligent monitoring system can then work out when to make necessary maintenance and service interventions for each car based on model, condition and usage. This would predict and prevent failures before they happen, improving safety and customer satisfaction. If all cars on the road had such a digital twin, the data could also benefit highways agencies, oil companies and vehicle and component manufacturers.

2. Improving performance, efficiency and quality

Think of the production lines that make components for cars. Each one, from lightbulbs to dashboard displays, involves complex, high-speed processes, with numerous different direct and indirect variables affecting throughputs, yields, quality and line stoppages.

Digital twins of these production lines, fed with near-real-time data, would learn how the changing variables impact the system. They would also be linked to actuators capable of making fine adjustments and taking action when needed, keeping things running optimally and in specification.

The digital twins would do this tirelessly, continually learning and getting better, and do it across changeovers for all variants of components made

The benefit to customers

Imagine if your car received a personalised update every week that set it up perfectly for you. Each update is tailored to the current condition of your engine, gearbox, brakes, steering, suspension and so on. But it’s also tailored to the roads and journeys you’re currently making, as well as the ones you’ve told your car you intend to make over the next week. It optimises your car according to the weather forecast and road conditions. It even looks at driving data to factor in your own personal driving style and habits. This update is created by the digital twin of your car.

On top of all this, you will be advised on a maintenance and service programme personalised to your priorities, be they maximising high-end performance, maximising resale value by minimising component wear or even, at a bare minimum, ensuring that every aspect of your motoring remains legal.

Digital twins require hard work

There are a lot of challenges to overcome before we reap the benefits of digital twins. Standardised data models and end-to-end channels are needed to exchange and share information. Many businesses will discover their internal systems are not yet prepared for this level of data sharing and collaboration.

It will also be necessary to think carefully about the design of the digital twin and its level of granularity. Designs that are too complex will likely lead to an overload of data that will be hard to process, manage and analyse. Yet overly simplistic designs are unlikely to deliver the analysis capable of unlocking the real benefits we’ve just described.

We can already explain and understand the clear opportunities and benefits of this technology. As it matures, it’s very likely that the scope, reach and uses of these twins will increase. We can start to imagine how interface technologies like speech, haptics and augmented reality might lead to new ways of interacting with and interrogating digital twins.

This in turn could increasingly blur the boundaries between the physical and the digital worlds. Additionally, blockchain could be used to establish immutable records of provenance, usage and performance that would span products, processes and supply chains.

As AI and cognitive technologies continue to evolve, the ability to analyse data from ever more sophisticated digital twins will emerge.

As more twins are connected, along with more data sources relating to processes and the environment, the insights and analytics will increase exponentially, leading to many new possibilities and business models.

It’s an inevitability that digital twins will advance. Perhaps in the future we might all even have personal digital twins that record and reflect the biological processes happening within our own bodies. These could be fed with data from wearables and be accessed by trusted healthcare professionals to diagnose problems and treat us in better ways.

Beyond that, if one day we could harness the properties of quantum entanglement, what might that lead to in terms of more deeply linking digital twins with their physical counterparts?

As is often the case with emerging and disruptive technologies, you should begin by thinking big about the potential to transform your organisation. Our advice is then to start small, experiment, learn and scale fast. With such a broad disruptive potential, can you afford not to get to know your digital twins?

Peter Durant is a technology and innovation consultant based at PA’s Cambridge Technology Centre. Rob Gear is a futurist at PA Consulting Group. Find out more about PA’s Digital Innovation Lab here.

PA Consulting are a founding partner of The HUB, D/SRUPTION’s recently launched membership network.