Smart machines that learn on their own are making all our lives easier. Smart machines hold enormous promise for making helicopters safer and more efficient, while improving asset availability and mission effectiveness.
At Honeywell, we’re constantly looking for ways to improve the capabilities of our industry-leading Health and Usage Monitoring System (HUMS) technology. We’re applying the latest Industrial Internet of Things (IIoT) and machine learning techniques to make our HUMS smarter and a lot more valuable to helicopter operators.
“We’re migrating from providing condition indicators to providing health indicators,” said Andrew Vechart, principal research and development scientist in the Vehicle Health Management group. “Our HUMS will soon be able to not only tell you whether a particular mechanical component is ‘good’ or ‘bad,’ it will indicate the health of the whole line replaceable unit so you won’t have to spend a lot of time interpreting the data.”
Looking further down the road, Honeywell sees a future when HUMS will apply Honeywell’s Connected Aircraft and advanced analytics capabilities to an even greater extent.
“Predictive analytics are the future,” said Raj Bharadwaj, staff research and development scientist, Vehicle Health Management. “Before long we’ll be able to tell an operator that a specific unit is likely to fail, within a particular timeframe, with a great level of confidence. This capability will enable a leap forward in condition based maintenance, which will provide enormous safety, operational and cost benefits to operators.”
This summer, Honeywell completed its work on a critical program for the U.S. Army Aviation Applied Technology Directorate (AATD) and prime contractor Boeing. The program was called ASTRO, which stands for Autonomous Sustainment Technologies for Rotorcraft Operations. The program focused on using machine learning algorithms to improve the way future Army rotorcraft HUMS systems will be designed and maintained. Our efforts centered on prognostics health indicator development for the drive systems gearboxes.
“This was an ideal program for us because of our expertise in analytics and vehicle health management, and the vast amount of HUMS data we’ve collected over the years,” Bharadwaj said. “Our data, combined with data the Army already had from post-failure teardowns and other sources, was instrumental in the success of the program. With this extensive database, the Army hopes to improve availability, maintenance and fault-diagnostics, as well as improve hardware reliability.”
Honeywell customers will benefit, too, from our participation in ASTRO. “Machine learning is the wave of the future and a real opportunity to enhance the capabilities of our HUMS products,” Vechart said.
“We now have a method to configure a HUMS box with the appropriate diagnostics and prognostics health indicators more rapidly. And we demonstrated that with machine learning we can add capabilities to our HUMS products that make them smarter and more intuitive for our customers.”
Machine learning and predictive analytics, coupled with our HUMS expertise, is a great example of how Honeywell is bringing the Power of Connected to the helicopter market, said Josh Melin, HUMS product line leader.
“The ASTRO program clearly demonstrated that these leading-edge disciplines have enormous potential to improve the safety and efficiency of rotorcraft operations,” Melin said. “The results clearly show that diagnostic and prognostics performance was improved using machine learning approaches to find a wide range of gearbox failures."
“But it doesn’t stop there. We’re actively working to apply our learnings from ASTRO to other types of aircraft and components beyond the gearbox as we improve the breadth and depth of Honeywell’s vehicle health management capabilities.”