newsmaker The world of aerospace engineering appears to share little in common with the bytes and pixels of information technology, but Ric Parker, director of research and technology at aviation giant Rolls-Royce Group, says IT has become even more instrumental for the U.K. company's success going forward.
"If you go right back to our tag phrase [which is] delivering excellence, that excellence increasingly comes from use of IT, from computer simulation to data analysis. That can take us far further than any human being can go."
Parker, who is based in Derby, was chatting with ZDNet Asia at the brand new 65,000 square meter Rolls-Royce Seletar Campus in Singapore, which was officially opened Monday and houses manufacturing, training and research facilities. He was in town also to speak at the Aerospace Technology Leadership Forum conducted by Singapore statutory board Agency for Science, Technology and Research (A*STAR) and attend the annual Singapore Air Show which started Tuesday.
Describing himself as someone who's "always been excited by technology", Parker joined Rolls-Royce back in 1978, holding various senior technical and management roles before becoming its research and technology director in 2001, which he calls "the best job I've had".
Technological innovation and disruption within the aviation industry moves a lot slower than in the IT realm, given that the former is an extremely "safety-conscious and tightly-regulated" field, Parker pointed out. Yet, there are compelling benefits to be had from embedding IT throughout the business--starting with the birth of an engine design right up till its operation in service for customers--which Rolls-Royce aims to reap, he said during the interview.
Q: How does information technology specifically relate to Rolls-Royce Group?
Parker: The starting point is always safety. The control system of an aero engine is designed to be inherently safe. They always say that about 10 percent of the software in a control system is to do the controls, 90 percent is to trap problems with the first 10 percent. IT also helps with fuel efficiency, being able to control fuel precisely and monitor that.
And then, there's IT modeling, which in the broadest possible term, is all about cost of ownership. We increasingly design the engine inside the computer before we start cutting metal to make something. It's about being able to simulate what's happening inside the engine, the gas flows, mechanical issues, and even complicated issues like what happens when a blade comes loose. For instance, how does IT help us design an engine that's cheaper in the first place, uses less metal or uses metal more intelligently? How does it help us extend the life of a component better by understanding exactly how it's going to operate, so that it becomes cheaper to operate? How does it help us reduce the fuel bill for customers?
The other part of IT that's increasingly important to us is data collection and mining. The sensors that we have in our engines can generate about 1 GB of data every flight. So it's about how we find the significant [insights] into that data and how do we use all that going forward to manage our fleet and help our customers.
So, IT's importance to the company has grown over the years?
Oh definitely. It's becoming fully embedded in the business at both ends. I talked about simulation; our design process is so dependent on IT now, right from day one. And if you're trying to model how the whole engine behaves in a very violent event, we can only model the first few milliseconds of such an event. We would like to follow it on over a longer period of time. Computing power is important as too is understanding how to get the best out of these big supercomputers. Destructive testing is extremely expensive. If you can get a computer to tell you, "Yeah, that's going to work", you still do the test, but you don't have to do it twice. When you blow up a real engine, it costs you S$30 million (US$23.7 million). The computer test is a fraction of that [cost]. That's the front end of the business process.
[At the other end], we have a very intimate relationship with our customers, where we're providing more and more services to them--from the normal services such as repair to, say, managing the fuel use for customers, so they can run the engines and aircraft more efficiently using the data we gather. So IT is becoming crucial to us at both ends of the process--birth to death as it were--from the design of a brand new engine to understanding how an old engine that's operating in service, how it's behaving, how it's deteriorating, and when it needs to come back to the shop.
You mentioned data mining, which analysts have cited as an opportunity as well as a challenge for companies to gain business insight. What is Rolls-Royce's take?
One of the key things is learning how your customer is actually using your product. You think you know how it's designed and how they're going to use it and even how they're supposed to use it, but are they doing that? How does the pilot fly the plane? How have they used the engine? Once you have that data, it then helps us design better for the future. At the same time, we can also use that data to advise the customer, "Maybe if you don't fly quite like this, you'll use less fuel", so you can build that into a fuel-saving regime for the customer. So the more you can extract from the data, the more you learn.
But the problem is the sheer volume of data. Sometimes you get too much. If you knew exactly what it was you wanted to measure upfront, then you just put one sensor on the engine and you sit and monitor that one single thing. That would be perfect. But you don't. So you put sensors all over the engine and often, the important information is coming from a combination of results from different sensors. So you have to understand all the permutations of the data.
Increasingly by using data mining, we can find what we call "novelty analysis". It is a self-teaching process, so it just learns as it goes along. Obviously in the early days, if it's only looked at five engines and something different happens, it can't describe any significance to that. [But] the more data it gets, the more it understands.
At the moment, we look as we go along. We gather data and analyze it for novelty as we go along, so there's some analysis, but we would really like to take this huge data sets [for data mining]. Where we're aiming is heavy duty data mining, taking the entire fleet data for a year, say one, two terabytes of data and put all that into one big machine and learn all the lessons from it. We haven't got there yet; that is the dream, to be able to crunch all this data. The ideal would be to take data direct from the plane via radio signals to the ground in real-time, understanding what's happening, and that again needs IT.
Given the benefits you've listed, has Rolls-Royce upped its IT investments? And is there good ROI (return on investment) or savings?
Yes. As soon as you tell your suppliers, 'Oh, we're happy this is the right shape, go ahead and start making it', then everybody incurs a lot of cost. If you can perfect it as far as possible in the computer before you start to cut metal, then you save a lot of money.
We are spending a lot more in recent years on IT, both in-house and using more high power computing. It's not a decreasing area of investment. As you know with IT, it's a constant race to keep the systems at the state of the art. As soon as you find the biggest supercomputer in the world, one year later, it's the tenth biggest. You're always chasing.
I also think there's a better understanding now throughout the company on what IT can do and why it's important. So it's not such an uphill struggle to get people to spend money on IT. There was a generation of senior management who maybe didn't grow up with IT, and every time somebody came and said I need a bigger computer, they say, I don't really believe you. That, I think, has taken a generation of senior executives who didn't grow up with computers to retire and be replaced by people who grew up understanding what computers could do, and are more easily convinced these really were valuable investments.
In 2010, Rolls-Royce was in a scrape after one of its jet engines on a Qantas A380 passenger plane failed. Could technology have helped mitigate it or has the company's technology strategy change because of this incident?
We understand quite well why and how that failure happened and it's very regrettable. We don't intend anything like that to happen. For this case, you couldn't point to anything and say 'Well, if we had better IT, that wouldn't have happened'. There were a lot of instances that were avoidable but not spotted early enough. It's not a case where I think IT would have gotten us out of the mess or avoided it in the first place. It was not up in the design stage, there were no fundamental design faults that led to it.
The only bit where IT came in was really a change to the control systems on the engines, so it could spot ahead of time if a particular component was going too fast. That's now embedded in the control systems. For a jet engine, it's not just 'Oh, that looks like a high speed'; you have to understand a lot of detail, is this component going faster than that component, and you make those decisions on a millisecond-by-millisecond basis. That's where the IT in the basic control systems helps. By having a better control system, you can certainly trap these sorts of problems in the future.
The IT world sees frequent and plenty of innovation and disruption, most recently with social media and mobility. Has there been as much technological disruption in the aviation space, and just as fast?
Not to [sound] conservative, but I mean this is a very safety-conscious and tightly-regulated industry. So if someone comes into my lab with a sticky black mess in a test tube and says, "I think this is going to be the next best material for making fan blades", then 20 years on, you might just be using that in service. That's the typical time scale. It's not an area where you want to take technology and try it out on aircraft. You don't want to put people's lives at risk by trying something brand new that nobody's used before. We do a lot of testing, and we have a very structured process for taking technology through what we call technology readiness levels, a scale which NASA (National Aeronautics and Space Administration) invented. It's not a quick process. Let's try it on a much safer application on the ground first and then build up experience.