It’s been a tough year or so for Air Force maintainers. High-profile aircraft failures plagued the service recently, including emergency landings of C-5 cargo aircraft, the grounding of the B-1 bomber fleet, and the loss of a C-130 propeller in mid-flight.
The immediate causes of these accidents vary, the but the root cause is the same: age. The average age of an Air Force aircraft is 28 years, and many planes are significantly older. Crews fly still fly the B-52 bomber, after all, with its average age of 56.
Will Roper, the USAF’s assistant secretary for acquisition, technology, and logistics, oversees $40 billion in programs. But even though 70 percent of the money USAF spends on acquisition goes to “sustainment,” he says, few people talk about the issue. “There are rarely hearings on aircraft sustainment, but it’s the reason we’re able to go fight and win a war. There are amazing men and women that keep airplanes able to fly, but here’s the sad thing: They’re using technology that is decades old.”
Roper’s job is to bring new tools to the Air Force, and in this age that means data mining. “We’ve brought in a lot of artificial intelligence experts to advise us on how to use A.I. to predict when planes are going to fail, and I believe we’re the first service to have A.I. operational on its flying fleet,” he says.
Late in 2018, the Air Force (with help from Delta) retrofitted its aging C-5 and B-1 fleets to perform predictive maintenance. “It’s already doing amazing work, telling us things that we need to look at before they become critical,” Roper says. “The data is there but it’s not in a discoverable format that you can layer in machine learning on top of it. A lot of what we had to do was reverse engineering, so that that data can be exposed in an algorithm friendly way.”
While performing a training mission,
an U.S. C-130 “Hercules” cargo plane from Puerto Rico Air National Guard crashed about 11:30 a.m. today near the Savannah Hilton Head International Airport with 5 people on board.
All are deceased. pic.twitter.com/ilx1f9saDr
— David Begnaud (@DavidBegnaud) May 2, 2018
He says there are more than 100 algorithms running on the C-5 systems, and more than 40 examining the B-1. Each algorithm parses the information generated by specific systems, like the landing gear, wheels, temperature sensors, and anything that is deemed mission-critical.
So far, the A.I. found three maintenance actions on the C-5 “that we wouldn’t have found through traditional processes, that affect 36 different aircraft,” Roper says. Maintainers also removed 17 parts that were showing subtle signs of wear well before those parts had issues.
The larger the fleet, the larger the time and money savings. This year’s test focused on large aircraft with small fleets, which are easier to handle. But the brass ring would be to bring A.I. scrutiny to fighters, both old and new. ”We’re working very hard to make sure that our newest fighter is going to be able to do the same kind of smart maintenance approach,” Roper says. “The more that we can see into the future the better.”
The adoption of predictive maintenance is one modern tool of many that Roper hopes to bring to the Air Force. He holds an abiding hope for 3D-printed parts, which he sees returning some balance to defense procurement. “Historically, militaries often made their own stuff. For Julius Cesar, the idea of not being able to make your own arrows or swords would be weird, right? He’s his own manufacturer,” Roper says. “The last 20 years military hasn’t made much of anything. We’ve had to rely completely on industry. But additive manufacturing machines put us back in the seat of being able to make our own parts, and we are. We’re making hundreds of parts today.”
There is a subtle morale issue at work as well. Many of the airplane groundings and emergency landings of 2018 were traced back to mistakes made by maintenance personnel. It can be hard to summon up much enthusiasm for a job focusing on dead technology. Using new tech tools to keep its fleet flying would send an opposite signal to civilian and uniformed maintainers who are working on vintage aircraft. For all the A.I. algorithms and 3D-printed parts, there’s always a human being working on the airplane who stands as its keeper. “I want people to think if you’re a maintenance officer in the Air Force, you’re exposed to some of the greatest technology that exists in the world,” Roper says.
Originally posted on Popular Mechanics