You said that predictive maintenance is not new, but the application of digital technologies into the business is different. Can you elaborate please?
We’ve done a lot of predictive maintenance at Pratt & Whitney and we’re working with our customers to find solutions for them. With new digital technologies, it enhances our capabilities to react faster and to develop more enhanced solutions for them. For example, our eFAST data acquisition system went into service last year on the Bombardier CSeries platform. That changes how we can access data and get visibility into our engines. So it’s about how you extract value, and different technologies can help us get there.
Today, we have 90,000 engines in the market and we expect this to grow to more than 100,000 by 2020. We are going from receiving a handful of parameters to thousands of parameters per flight on our latest engine platforms with our eFAST ecosystem. To help highlight the amplitude of this statement, I would like to highlight that today, the data generated by one flight on our CSeries platform is more than the data generated in flight from our entire existing fleet.
Obviously a lot has happened in the IT world—the Cloud, storage and acquiring data—and at the core, it’s all about security. Customers trust us to access their data. We add our smarts to it, we make recommendations but it is a mutual relationship. Every piece of the information needs to be protected.
We’re also making significant investment on the manufacturing floor: how do we better connect the machines and how do we better connect the data in that space to truly enhance the predictive maintenance. To truly enhance the customer experience, you need to collect all of these pieces of information so you can react better to any situation. The Connected Factory concept works. You can connect the data from when you design the parts to when you manufacture them and then get them into the field. The true value is when you connect each piece. Use Amazon for example. You can track you parcel—you know that it’s going to arrive on Wednesday morning—and you’re confident it will arrive. We need to do the same thing with engine shop visits. So that’s another opportunity where we can apply new technologies. How do we better connect all of these pieces of information to enhance the hardware availability, the engine turnaround and support customers the way they expect us to.
How do you parse all of that data coming from the geared turbofan—and filter what you need to, with the right algorithms?
There are different thoughts out there about whether or not you need all of the data. We store all data in its raw form. When we do convert and process it through our algorithms and analytics, we might not have used all of the data to find a conclusion. But it’s there. So six months or five years later, something new happens and you’d like to go back to the historical raw data to confirm your assumptions or analytics; the data is available. If you don’t collect it all initially, you’ve lost that opportunity. It’s nice to have the people and technology to go from diagnostic to predictive to eventually prescriptive. It’s going to be an evolution. But with the data matched with the technology, we can do it.
Where are we in the path to get to prescriptive?
It’s important to understand the airlines’ perspective, because at the end of the day, they’re responsible for the aircraft, flight and health of the engine. To go to prescriptive, there are instances where we can do that today. We know enough today to say, not only did I diagnosis it, I came up with a root cause, I see a problem coming and can see that in about four weeks you should replace it and do ABC in a more prescriptive way. The airline in return says, ‘I need to match it with when the aircraft is available.’ But they say, ‘At least I know and it’s in my court, and I can manage it with my maintenance.’
Airframers already are trying to adjust to that new need of airlines, which say they don’t want to be constricted to A or C checks. ‘My aircraft is available for the next two nights: what can I do then? Instead of having an aircraft down for 30-40 days, can I reduce it to two weeks or even one week because I’ve done all of those small interventions ahead of time.’ That’s the response: Diagnostics will always be there, but we need more predictive and prescriptive because that could help airlines better manage their business.
Looking at data in a holistic way, you mentioned during your panel during Aviation Week’s MRO Middle East that the industry should work together. Is it realistic?
I’ve seen good change during the last two years. We’ve gone from companies saying ‘these are my solutions and I’m the best’ to hearing more about ‘this has to be an integrated solution.’ Does it mean that one body controls it all? I don’t think we’re there yet. But I think we need to find a way to all be able to offer our expertise at the end of the day to the airline.
I’ve launched a ‘working together team’ with an airline, a parts manufacturer and ourself, Pratt & Whitney. We were trying to work through the challenge of how to share data amongst ourselves, which system do you use and how do you allow data to go from the airline to the airframer to the engine OEM to the part level. Most of the time, it’s the part level, like a valve, that will cause the cancellation. The part maker would love to get into the game. How do they get the data? That was what the working together team did: see how to flow the data.
Once you do that, the airline asked, when you return that part to me, can you share the learning too instead of just sending me the part with the tag. So it created new opportunities to better serve the customer.
Data ownership always comes up in these conversations. How can you do this while protecting IP, respecting different business models and serving the customer?
Any time you have a conversation about data ownership it becomes very emotional. Even if you say the raw and converted data belongs to the airline, you quickly get, ‘once I add my smarts to it, to whom does it belong?’ We want to stay away from those conversations. It doesn’t add value. We’re focused on whether we can get access to the data. I’m not interested in taking your data for the sake of data. If you allow me access, we can focus on finding the benefit for your fleet, and that’s when you find a mutual benefit. Then I can develop solutions for you, you give me data to accelerate my learning on my products, so it’s a win-win combination.