As an Airline/MRO, AFI KLM E&M is well aware of the value of IFE systems. As one of the most important investment items in an aircraft's configuration, they are an essential part of the passenger experience. An IFE that suddenly breaks down means a dissatisfied customer with a direct impact on Net Promoter Score (NPS) or a seat that has to be taken off sale - with the resulting loss of revenue.
BlueLink
Prognos for IFE: anticipate failures and keep your passengers happy!
Prognos for IFE
Anticipate failures and keep your passenger happy!
Developed in-house by AFI KLM E&M, the exclusive predictive maintenance suite Prognos® expands once again. After aircraft, engines, components and APUs, in-flight entertainment (IFE) systems will soon benefit from the power of Big Data and AI technologies to predict when to replace devices before they actually fail.
Close supervision for in-flight entertainment systems
Following on from the approach developed with Prognos® for Aircraft, Air France-KLM's R&D/Data Science teams have now turned their attention to the field of IFEs. In October 2022, an experimental project was launched to explore the predictive potential of the flow of data generated by IFE and content servers in all classes of an A350 aircraft.
“The aim is to analyze this mass of data, such as the error messages transmitted by the IFE, and through feature engineering to feed the right information to a Machine Learning model and learn failure patterns from it. By taking hundreds of parameters into account, it is able to characterize the state of the IFE and, above all, identify the weak signals that could lead to failure. The project has delivered convincing results since the technology is capable of anticipating a breakdown between 15 and 25 flights in advance, enabling teams to be alerted and the appropriate maintenance action to be organized.”
Jad Naciri
Data Science Team Leader
A high-performance model
Still in the experimental stage, the system has achieved a significant level of performance. This is evidenced by a high "recall" score, i.e. the number of events correctly predicted (true positive) in relation to the total number of events. With an average recall superior to 90% for all classes combined, the Prognos® for IFE model is already demonstrating its high capacity to identify units likely to experience a breakdown. Although IFE reliability is not a critical component of flight safety, it is nonetheless essential to the airline's offer to its customers. With this innovation, AFI KLM E&M introduces a welcome dose of predictability, to preserve the integrity of IFE systems and, by extension, the quality of the flight experience for passengers.
