Leveraging its in-depth knowledge of A380 systems, its ability to innovate and its Big Data skills, AFI KLM E&M has developed a predictive analysis program used to replace a fuel system component before it fails.
"The MRO Lab" is a program grouping innovative, high value-added solutions that meet operator requirements to optimize their fleet maintenance. Developed in-house, then tested and proven on the Air France and KLM fleets, these innovations are subsequently marketed to AFI KLM E&M clients, who benefit from this unique know-how and its concomitant scale effects, improving their operational performance.
This was the framework in which, two years back, AFI KLM E&M began to explore the use of Big Data applied to predictive maintenance. What does it involve? "Using the aircraft's history to automatically predict and prepare for any future part failures before they occur", says James Kornberg, Director Innovation. This exercise is made possible by the large amounts of data now supplied by new-generation aircraft.
With the assistance of the operational research specialists at the Group's Information Systems department and external partners, AFI KLM E&M's engineering teams have developed a series of algorithms that use the data automatically downloaded by WiFi from the aircraft during their stopovers at Paris-CDG. The results are transferred in real time to the Maintenance Control Center where the specialists receive alerts on the components detected via a sophisticated graphical user interface and, where appropriate, trigger maintenance operations.
One example of the application is PROGNOS for ENGINE. This software is being designed to carry out statistical analyses of engine data to enable dynamic monitoring and predict failures using an early warning system, for the fleets of AIR FRANCE and KLM as well as client airlines. PROGNOS for ENGINE is part of a series of projects and initiatives focused on Big Data that have already led to operational solutions such as PROGNOS for AIRCRAFT, the early warning and failure monitoring software used for the A380's systems, which extends the solution to bigger data volumes. Solutions based on the same approach are currently under development to support a range of other critical aircraft systems and components, including on the 787.
The predictive maintenance program has been trialed since early July on the fuel circuits of the Group's ten A380s and has now confirmed its reliability. "The predictive detection rate stands at 70%", says a delighted Mr Kornberg. "In the space of three months, four pumps and switches have been replaced using this new system in predictive mode and also in corrective mode, which makes us more effective when carrying out repairs."
As experienced in the field by AFI KLM E&M while using PROGNOS to support AIR FRANCE KLM's engines, components and aircraft, fleet operators can reap significant benefits from these new predictive maintenance technologies, which contribute to avoiding operational disruptions, resulting in increased passenger satisfaction as well as reduced maintenance costs.