A few weeks ago, myself and the team at Hitachi Solutions had the pleasure of attending the Microsoft Inspire 2019 in none other than glitzy Las Vegas. The conference was a great opportunity to meet partners and clients from around the world and take stock of where Microsoft technology is heading. It was also a fascinating showcase of digital transformation success right across the Microsoft technology platform.
Of the multitude of sessions at Inspire, it was a case study from Unilever that most inspired me. During the Gavriella Schuster and Judson Althoff’s CoreNote speech, the Chief Engineer of Unilever, Dave Penrith, was invited to share his success in employing a technology known as digital twinning in a Unilever factory.
For me, this really resonated – because at Hitachi Solutions we’ve seen for ourselves just how powerful digital twinning can be. In this blog, I thought I’d take the opportunity to discuss this in more detail and consider how the technology can drive results and performance for other manufacturing companies everywhere.
A digital twin is a digital replica of a real-world asset, process or system. In most deployments of the technology so far, this involves creating a digital simulation of a machine, powered by real-time data from the physical machine itself. The technology uses the Internet of Things (IoT) to pull in information about the machine and build the digital twin. Then AI and machine learning algorithms can be deployed to analyse this information, using these insights to predict dangers and power improvements.
Twinning was first pioneered by NASA, long before the days of IoT, because they wanted a way to remotely analyse, detect and fix problems with their space-based machinery. But it’s only with the expanding potential of new technology like Azure, machine learning, IoT and analytics that digital twinning can become a reality for today’s manufacturing businesses.
The Unilever case study struck me particularly because it was such a fantastic example of how digital transformation can make a real and tangible difference to not only businesses but people. In their Valinhos factory, Unilever used Azure, IoT, Analytics and AI technology to create a digital architecture for the factory that produces Dove soap.
Before the digital twin was pioneered, power outages and machinery failures were a big drain on productivity. Together, the three production lines in the factory produce around 1,500 bars of soap a minute – meaning every second matters. With no data available, engineers would have to scale the 40-foot machines themselves looking for the fault – a task described as ‘searching for a needle in a haystack’.
Unilever used IoT and AI technology to create a digital twin of the soap-producing machinery. Powered by IoT sensors, the digital twin pulled in an enormous amount of data about the machinery – from the amount of soap bars produced each minute to the temperature of individual machine parts.
IoT sensors and controls on the physical parts send signals to digital simulations, allowing plant operators to almost immediately analyse the condition of every part, optimise power and figure out if maintenance is necessary. Feeding that information into machine learning and analytics algorithms allowed them to identify trends and obtain insights, which became richer and more complex over time as increasingly more information was added.
Processing and analysing this vast amount of data allowed them to understand more about the machinery in the Valinhos factory. It meant problems in the machinery could be more efficiently identified, and in many cases – pre-emptively prevented from occurring in the first place. And what were the benefits? Well, in just the first year, Unilever saved €2.5 million and boosted their productivity by 3%.
Digital twinning is becoming increasingly popular in the wider manufacturing industry – for all the same reasons that Unilever adopted it. Digital technology allows manufacturers to achieve faster, less expensive R&D cycles, creates safer, better products, and facilitates better decision making.
In fact, IDC goes so far as to predict that by 2020, 30% of Global 2000 companies will use digital twins and IoT-connected products, to improve product innovation and organisational productivity, achieving up to 25% profit gains.
But the benefits can come in many forms – not just gains in productivity. In the automotive industry, for instance, digital twinning has been used to help predict and resolve maintenance issues and measure the environmental impact of a production process.
General Electric, for instance, have also experienced success with the technology, creating a digital wind farm that allows them to maximise the efficiency of their wind turbines. As well as this, they’re even looking into the possibility of combining this technology with alternate reality (AR), using a HoloLens to look into a physical simulation of the digital twin.
At Hitachi Solutions, we’ve had the pleasure of being able to experience the potential that IoT-powered technology can provide. Working with Okuma, a Japanese machine tool manufacturer, we helped them accelerate factory cycles and visualise production for total optimisation. We did this by gathering and linking data on production progress and the operational status of their facilities, using IoT. Much like with Unilever, we discovered exactly how powerful IoT technology really can be to improve processes and drive business outcomes.
If you want to find out more about the impressive Unilever case study, I’d absolutely recommend that you watch David Penrith’s contribution to the Corenote speech, which begins about 20 minutes into the video. There’s more to the Unilever case study than we could ever fit into this blog – so it’s worth investigating further.
With the glitz and glamour of Las Vegas and Microsoft Inspire now a distant memory, all that’s left to do is start getting ready for 2020! In the meantime, if you want to find out more about how you can benefit from IoT and Digital Twinning technology in practice – get in touch with the team at Hitachi Solutions.