Digital Twins: distilling science from fiction

Once the preserve of critical infrastructure or high-value assets such as gas turbines, digital twinning is now a technique coming to a production line near you.

Danny McMahon of the University of Strathclyde’s Advanced Forming Research Centre (AFRC) discusses the application of digital twins to traditional industries such as whisky production.

Digital Twin Digital Twins Digital Twinning - Stock Image
Digital twins aren’t necessarily a new concept – the term has been around since the turn of the millennium.

Digital twins: it’s a concept that sounds like something out of science fiction. It makes you think of people recreating themselves as avatars in computer games like Second Life, The Sims, or World of Warcraft and living out an alternative reality in the digital realm.

Another interpretation might even resemble films such as The Matrix, in which those living in ‘the real world’ engage with the planet as we know and see it today through computer-rendered versions of themselves.

That might be the world of fiction, but digital twins aren’t that far off what the creators of The Matrix envisaged.

Essentially, they are digital renderings of physical equipment, displayed on a headset, computer, tablet, or other device, showing real-time data about a range of measures: from its condition and historic performance, to its current position and any external factors acting on it.

Digital twins aren’t necessarily a new concept – the term has been around since the turn of the millennium. Since then, it’s grown from an idea into a relatively common way for businesses to monitor their assets. Recently, however, it’s begun to change focus.

Moving down the value curve

To date, digital twinning has primarily been used to monitor high-value assets. Wind turbines, aircraft engines, and other highly-technical and expensive engineering systems are the most common places the technology has been implemented, quietly tracking all the data they produce and delivering it to technicians in remote locations.

In the latter example, they tend to monitor engines while they are in flight, notifying their proprietor about performance and any issues – sometimes even before the pilot realises.

Yet, in recent years we’ve seen a small but significant change in digital twinning – a variety of factors has combined to open up the technology to other types of businesses, particularly further down the value curve.

This article first appeared in the September issue of The Manufacturer magazine. To subscribe, please click here.

It’s being driven by a better understanding of what’s available and what can be achieved, as well as cheaper and more accessible sensors, and the increasing ubiquity of connectivity.

Outside the manufacturing industry, we’ve seen digital twinning used in construction to help contractors, architects, and facilities managers understand the lifecycle of buildings. It’s allowing them to enact new, more prescriptive ways of working, such as condition-based servicing, rather than checking at set intervals for faults and repairs.

The same sort of concept has been applied in engineering, to turbines and engines. A specific new trend for digital twinning has recently begun to emerge from manufacturing. Instead of being the sole preserve of high-value assets, such as turbines and engines, it is being scaled down to a lower level, increasingly being considered for use in assets that aren’t necessarily high value in themselves, but create high value products – think luxury goods, automotive, or premium food.

Savouring every drop

Barrels Brewing Distilling Whiskey Whisky Alcohol Scotland - Stock Image
The company involved has a specific challenge with them: its current valve system only fills the casks to 95% of their total volume.

A project AFRC has recently taken on in one of Scotland’s oldest and most-celebrated industries is a case in point. We’re working with a range of partners to develop a digital twin for a whisky production line, predominantly focussing on a very specific and crucial part of the process – its valve.

In this context, valves are mostly used to control the flow of whisky into casks. The company involved has a specific challenge with them: its current valve system only fills the casks to 95% of their total volume.

While 5% of a cask might not sound like a huge burden, when it is extrapolated over the volumes of whisky that this company produces, it adds up.

The savings could either be achieved through maximising the amount of whisky casked and ultimately sold, or saving the business the inventory space during its future expansion. This works out as the equivalent of an entire warehouse.

The consortium involved is aiming to tackle this challenge head-on through an Industrial Internet of Things system that optimises time standards and controls, thereby ensuring that casks are filled to 100% of their capacity.

Spillage from turbulent flow is relatively common with the current setup, so it will also be developed to factor this in and react accordingly. Essentially, we’ll be able to count every drop that has gone into each cask.

Starting with a taster

To test the system’s potential implementation on the production line, we’re creating a new system design that will be modelled and simulated to allow zero-physical prototype product development – the entire test setup is digitally rendered and built to be absolutely correct, using data drawn from actual conditions in the plant.

That means the digital twin will be placed in conditions as close to reality as possible, rather than relying on nominal data. It will also reduce the resources required to build the prototype, make it quicker to develop, and eliminate the waste associated with constructing a physical system.

Using CAD software, the digital twin will simulate monitoring the spirit’s density and flow as it passes through the line, understanding in real-time what blend is being poured into each barrel and to what level they’re being filled. It will also set the parameters for its implementation in real conditions.

Danny McMahon, senior manufacturing engineer and team lead for metrology and digital manufacturing, The University of Strathclyde’s Advanced Forming Research Centre (AFRC).
Danny McMahon, senior manufacturing engineer and team lead for metrology and digital manufacturing, The University of Strathclyde’s Advanced Forming Research Centre (AFRC).

When overfill occurs, the system should automatically recover the excess and recalculate the volume in the cask and return the excess to recirculate into the filling process.

When it’s implemented, the twin should help the company understand how the valve is filling the cask, how it’s performing and, when there are any issues, issuing alerts on the associated virtual reality (VR) headset or tablet.

A user will be able to keep an eye on all the casks on multiple lines simultaneously. This is typically a demanding manual process.

Aiming for a new blend

After the first phase, this system will go further. Everything will be built on a one-to-one scale on the VR headset and get to the point of delivering ‘on-the-fly simulations’ – in other words, being able to change variables and, then, ultimately the production throughputs remotely, before tinkering with the physical apparatus in any manner.

Eventually, the goal is to be able to make changes in the virtual realm that apply in real life. Imagine being anywhere in the world and being able to log in remotely to drive machinery and make adjustments, taking it yet another step closer to the concept developed by the creators of The Matrix.

Digital twin technology is advancing rapidly, and increasingly seems to reflect some of the inventions of games developers and science fiction writers – perhaps it’s a by-product of the former being so heavily involved in their development.

It’s exciting to see where this technology will go next and what industries, even further down the value chain, could see the benefits of twinning. It could be coming to a production line near you soon.