Rolls-Royce engines have powered aircraft for more than a century, but recent financial problems have forced the company to launch a transformation programme to cut costs.
Chief Digital Officer Neil Crockett is playing a key role in the recovery. Crockett is leading a digital strategy that uses data innovation to drive efficiencies and improve Rolls-Royce products and services.
"We can't just rely on the engines that we have built before," said Crockett at the AI Summit London. "We've got to think of all the different types of power that are available and we can actually utilise.
"We also have to think about being a pioneer. We have to think about being cutting edge. We've got to be a technology company not an engine company. We have to think wider about the climate and not just about the products that we make."
AS CDO, Crockett's thoughts tend to focus on digitalisation: building digital twins that improve physical assets, utilising machine learning to help engineering, and building a centre of innovation that is discovering new applications of technology.
When Crocket joined Rolls-Royce two years ago, he struggled to grasp the complexity of the company, until a trip to a Singapore factory brought the business to life for him.
Floating on pulleys above Crockett, 10 of the XWB jet engines that power Airbus planes were being assembled. When the engines' fans are in action, they suck in a squash court-sized gulp of air every second. There are around 50 turbine blades at the centre of each engine, which operate at temperatures three times above the melting point of the metal. The turbines collectively have the power of 55 Formula 1 cars.
Producing that blend of power and precision comes at a high cost, which Rolls-Royce can only afford through long-term service contracts. This business model presented an opportunity for data to show its value. Rolls-Royce now uses analytics to predict future faults and guide inspection scheduling around them, which helps the company cut downtime.
"Ninety-seven percent of the faults found on our engines are automatically predicted," said Crockett. "By planning and understanding how our engines work, we have reduced disruption to our customers by 40% in the last 13 years, and we've reduced our maintenance burden by 30% since 2012."
Rolls-Royce is driving further efficiencies through the use of digital twins. These scale replicas of the physical engines are used to digitally model the performance of the real machines. Rolls-Royce uses them to save time and money in the rigorous testing that's required before an engine is certified as fit to fly.
A particularly challenging test is on the performance of an engine when one of the fans breaks during a flight.
Previously, testers would rev up the engine as fast as it can go and then ignite a small explosive charge to release one of the blades. To pass, the engine has to stay in its casing and prove that it can survive for the rest of the flight.
However, with each engine costing tens of millions of pounds to produce, there's a high price to pay when a test goes wrong. The use of digital twins in virtual tests allows Rolls-Royce to predict how the design will behave, without the cost. Engineers can then iterate until they're certain that it will pass the real thing first time.
"This is incredibly advanced digital twinning," says Crockett. "In that 0.2 seconds when you know that you've failed the test, to process just that bit of data takes six weeks.
"This is intensely power and compute consuming. And of course, the great thing is that now we've got that data we can put it back into design, to understand our design better and understand our maintenance schedules better."
R2 Data Labs
The digital twins and data analytics have already proved their value, but to compete with startups and its traditional competitors, Rolls-Royce has to continuously innovate.
In November 2017, the company unveiled its plan to stay at the cutting edge. Rolls-Royce had moved 274 of its leading developers from eight global locations into a new unit for data innovation called R2 Data Labs, which Crockett heads.
The strategy is to create small hubs of talent surrounded by large ecosystems of experts from the private and public sectors and focus the research on industrial problems.
The approach has already yielded results. R2 Data Labs has used analytics to understand how an engine control system can be used to optimise fuel efficiency, machine learning to evaluate the terms of contracts with external contractors, computer vision to diagnose faults, and visual recognition to demonstrate autonomous shipping.
"We are now using machine learning to find the unknown design possibilities in terms of engine design," said Crockett. "We are finding that the amazing propositions coming up from those AI solutions are actually higher performance than the manual work we are doing."
Crockett advises other companies that want to create centres of innovation to focus on the business rather than the tech, aim for a self-service approach instead of relying on centralised teams of experts, and to concentrate on automation and emphasise collaboration.
They should choose a common collaborative analytics platform and develop their products in sprints, building skills through academies and working with small hubs to develop a big ecosystem.
"The benefits are going to be reduced costs," he said. "It's hundreds of experts, it's hundreds of millions of pounds. We're going to take seconds and minutes - not months - to get conceptions done. We're going to make compliance happy, we're going to make customers happy, we're going to reduce scrap and we're going to learn things for our design and for our maintenance going forward."