A digital crystal ball: Boeing’s analytics team predicts maintenance issues before they happen.
For commercial airlines, few setbacks are more significant than an “AOG” — airplane on ground — which occurs when a mechanical issue causes an airplane to be unable to fly. With an AOG, passengers are delayed, flight schedules are upended and maintenance can take days or even weeks.
That’s where alerts developed by Boeing Global Services’ Digital Solutions and Analytics (DS&A) Engineering team in partnership with the Managed Analytics team, can save the day. These alerts detect and prevent common maintenance issues before they cause an AOG. This capability recently prevented maintenance issues for a customer’s airplane, though the alert was developed well before the airplane ever took off or landed.
Several years ago, an operator discovered a cracked fuel manifold on their aircraft. A fuel manifold is a peripheral main pipe that distributes fuel from the main tanks to the engine of the airplane. A crack may occur in the fuel manifold as a result of fatigue over time, causing the need for repair.
Previous to this event, Boeing was able to obtain flight data for a different airplane experiencing the same issue. As a result of analyzing that information, engineering data scientists discovered a pattern or “signature” in the sensor data that indicated a tiny crack in the fuel manifold was growing over time. They created an alert that would indicate whether flight data matched the signature pattern which could mean a similar issue was developing on other airplanes.
In 2020, because the alert notified Boeing’s DS&A Engineering team that the data pattern was a match, the Managed Analytics team informed the customer, which enabled them to schedule the required maintenance without causing flight interruption.
“Once the maintainers were in the fuel tank, they confirmed the crack was developing, which validated our alert model,” said Seth Cohen, BGS Engineering customer focal. “Thanks to the partnership between DS&A Engineering and the Managed Analytics team, we will be able to notify future customers of this rare problem, and help them avoid a costly delay.”