A small but productive force of nearly 600 software engineers, data scientists, and customer representatives comprising GE Aviation’s Digital Solutions team continues to introduce and refine digital data solutions for aircraft operators. With products geared directly toward four groups of customers—commercial airlines, military, business and general aviation, and lessors—the GE Aviation Digital team uses aggregate data from thousands of engines and aircraft to provide asset lifecycle and performance management, flight-risk management, and other data-driven services directly to operators.
For the business and general aviation segment, GE Aviation Digital applications such as Corporate Flight Operation Quality Assurance (C-FOQA) Centerline and Prognostic Health Management Plus provide a way for small operators to reap predictive data benefits similar to airlines through aggregate data collected and analyzed by GE.
“If you are a major carrier with 800 aircraft, you have a big data set within your own airline,” Jon Dunsdon, GE Aviation Digital chief technology officer, told AIN. “Whereas if you operate a single business jet, the fact that data is being shared across the industry allows you to be more aware of what might be happening to your aircraft, how your aircraft is different than all of the others.”
C-FOQA can be traced back to 2004 when Flight Safety Foundation and NBAA committees worked with a handful of operators to collect and process flight data that corporate flight departments could use to analyze their operations. Handed over to Austin Digital before GE Aviation’s acquisition of the company in 2012, the C-FOQA Centerline database encompassed approximately 60 aircraft in 2012, but now includes data from more than 500 aircraft and 800,000 flight hours, thanks in some part to GE Aviation’s partnership with Gulfstream.
“The work we’ve done with Gulfstream, being part of their solution, aggregating that data across the Gulfstream fleet, allows Gulfstream to provide much better service for their customers,” Dunsdon said. “C-FOQA is a fantastic example where we not only provide the tool but also have our experts look at the data and advise you on your own asset."
Customer involvement and communication with GE Aviation representatives are key to getting the most out of the service. “We recently had a Gulfstream aircrew in flight radio back with an issue,” said Dunsdon, “We established two-way communication with the aircraft to interrogate what was going on, and that flight continued because of the crew’s confidence in what we could detect.”
Currect Tech Enhancements
Dunsdon listed three areas of digital technology that are beginning to make a significant impact on aircraft operations: on-aircraft processing of data using edge computing, connectivity solutions including bringing broadband to the cockpit, and application of artificial intelligence/machine learning processes to flight and component data analytics.
“The ability to run processing at the edge on the aircraft, update your analytics and not have to go through a long, complex, recertification process to make those changes is a huge differentiator,” said Dunsdon. “Then there’s connectivity: the ability to get data on and off the aircraft. We’re looking at ways to manage satellite communications, multiple satellite links, ACARS messaging, Wi-Fi, and how to put broadband over those links. You need a very flexible solution to connectivity because companies keep launching more types of satellites and you need to keep up, which comes back to edge computing becoming really important; because you’ve got to be able to change software much faster.”
There has been some movement toward applying artificial intelligence (AI) and machine learning (ML )for asset performance management and failure prediction, especially in the MRO arena. Dunsdon said the challenge in applying AI/ML in this arena is that while most technology companies are building their AI systems to recognize images or implement video facial recognition, aircraft component predictive systems must detect failures in critical items that perhaps have not failed before, or the failure may not look the same each time.
“It’s a very domain-specific approach to analytics,” said Dunsdon. “The accuracy that you need in our predictive systems is much higher than simply identifying whether something is a dog or a cat, and you must determine whether your analytics are actually confident in the question they’re being asked. Sometimes we’ve never seen a failure like that before, so we cannot be sure, but we can say, ‘Look, we think something is wrong.’”
While applying AI/ML to predicting component failures is a technologically daunting challenge, the GE Aviation Digital team can build on the expertise and technology developed by GE’s Digital division, which has its own set of asset performance management services for industrial applications.
“One of the benefits of GE being a global conglomerate is that we get to share technology between the business units,” said John Mansfield, GE Aviation chief digital officer. “Healthcare has actually been leading the push on AI image processing. They do it from CT scans, MRIs, etc., determining how to detect tumors more quickly by looking at images. A lot of that technology is making its way into the material science world.”
As CDO, Mansfield says that his job is to look ahead five to eight years to predict where technology is going and how to coordinate research activities to “get ahead of the curve.”
“We're even thinking about things like taking information off of a pilot’s smart watch and correlating that with the actual flight path,” said Mansfield. “I used to be a pilot and it was three hours of boredom and 10 minutes of terror. You just didn't know when the 10 minutes was going to happen. That was the problem. So you start marrying that structured and unstructured data together, you can get some great insights that you just couldn't get before. It was all gut instinct before. Now we're going to base it on data.”