When fixing Navy readiness problems, money helps. But so does data.

Like the other military services, the Navy has been working in recent years to improve the readiness of its aviation fleets, a problem that was exacerbated when cuts under the Budget Control Act almost a decade ago dealt a serious blow to aircraft availability.

Bigger maintenance budgets over the last few years have helped, but money isn’t everything. The Navy is trying to innovate its way out of the readiness problem too, and with some success, thanks to a combination of processes borrowed from the commercial airline industry, data analytics and emerging technologies like artificial intelligence.

The first big idea the service adopted from the commercial world was the concept of reliability control boards. In the Navy’s case, their most important function is to get the entire maintenance enterprise focused on the biggest readiness “degraders” — the parts of any given system that do the most to keep aircraft grounded.

“The idea is fairly simple, and that is you have to all be on the same page when it comes to what’s hurting your head the most,” Robert Smith, the head of the reliability control board data analytics team at the Naval Air Warfare Center Aircraft Division (NAWCAD) said in an interview for Federal News Network’s On DoD. “We can’t have one part of the enterprise working A through F and the rest working X, Y, and Z. We all now agree that the one through 20 are top head-hurters.”

The approach appears to be bearing fruit. The Navy is now cataloging those “head-hurters” in a new degrader tracking system. And after having identified the top maintenance issues for a variety of Navy aircraft, the team has managed to reduce the impact of those “degraders” by 34% in the two years the process has been in place.

But identifying those top issues isn’t as easy as it might sound. Finding the root cause for a particular problem takes real work by skilled diagnosticians, and the Navy generates hundreds of maintenance action forms for each type of aircraft it owns every month. Collecting that data and identifying common issues across an entire fleet of airframes is not a trivial task.

But it’s easier with artificial intelligence algorithms that can process all of that data all at once and draw conclusions from it. NAWCAD has just started using an Army-developed system called the Composite Learning Algorithm for Records Evaluation (CLARE) for that purpose.

“It’s this enormous database we have to go through to identify what’s actually hurting our heads, what’s actually going on,” said Jason Thomas, the data analytics team’s principal analyst.  “So the artificial intelligence solution looks at relationships in the data and what’s documented to say, ‘Hey, when you said [the problem is] this, it’s actually this. And that helps us score and correct all that maintenance data so we don’t have some of the inaccuracies that get perpetuated through the system.”

Using technology to identify the real root causes of aircraft downtime has its obvious advantages, but the ultimate goal is to use similar approaches to begin a course of predictive maintenance, so that the right parts can be replaced before they turn into maintenance problems.

“We’re heading there as quickly as we can as an organization,” Smith said. “The one thing that’s going to get us there is what’s called the Failure Reporting and Corrective Action System. It will contain all the data behind a degrader, including what we’ve identified with CLARE. It will actually provide us with the ability to optimize our maintenance to predict failures before they occur. It will give us the confidence that the time on wing for a particular component is X. And before we get to X, and that aircraft is inducted for some type of maintenance, we will have the confidence to remove that component and replace it before it fails.”

For any of those ideas to work, the Navy first needs to trust its own data. NAWCAD is now in the process of validating and verifying the Navy’s vast stores of maintenance records and flight performance information. The goal, Thomas said, is to prioritize which of those datasets will be most useful for creating real-world impacts on Navy flight lines.

“The question that always comes up is, ‘What projects do I do? Where do I fund? Where do I spend my energy and effort?’ So we’ve partnered with IBM to help us optimize that decision logic to say, ‘Hey, if you’re going to do your projects, do project one, six, and 18, don’t do one, two, and three,” he said. “Those may be engineering change proposals or they may be buying more spare parts. But we really have to look at it from a systems view and not just buy more parts. We can’t buy our way out of this. We’re too smart, and we can’t keep doing that.”

Related Stories

Comments