ad16mroit-patrickdelapierreairfrance-industriesklmengineeringandmaintenancepromo.jpg.crop_display.jpg Air France Industries KLM Engineering & Maintenance/Patrick Delapierre

Big Data: Finding The MRO Nuggets

Avalanche of operating data coming from new-generation aircraft poses a new problem for MROs—how to mine it for the most useful information.

At the 2010 SITA Air Transport IT Summit, Boeing and Airbus executives extolled the virtues of new-generation, e-enabled aircraft providing much more operating data. But several airlines responded by questioning the benefits of existing aircraft data and cited the expense and difficulty of managing even more. 

Five years later, the new aircraft have entered or are entering fleets, and data is plentiful. For example, Airbus’s A320 generates 15,000 parameters per flight, the A330 30,000, the A380 250,000, and the A350 will generate 400,000 parameters. But where should carriers and MRO providers look first for real payoffs from all of this information?

Airbus itself is proceeding along two paths, explains Philippe Gourdon, vice president for engineering and maintenance services. First, it is using design expertise to add two modules to its Aircraft Health Monitoring system: Expert, and Prognostics and Risk Management (PRM).

Expert uses real-time data to start troubleshooting faults during flight, so these can be remedied quicker at gates. The module will support A380s in December 2015, A350s in the second quarter of 2016, then A330s in the fourth quarter. Airbus is still determining whether the A320 generates sufficient data to make Expert worthwhile. Carriers, often low-cost airlines with short turn-times, are interested in further minimizing delays for the highly reliable Airbus narrowbody.

PRM uses sensor data to predict future failures and help decide when to repair problem components. It should be up and running for A330s and A380s before the end of 2015 and A350s and A320s in the second quarter of 2016.

These engineering techniques apply to both electromechanical and electronic components. For example, valves that are part of many aircraft systems are designed to open or close in certain time spans. PRM can detect slower movements through its sensors, then flag degradation and future problems. This logic already has been validated for Delta Air Lines’ bleed valves.

Airbus is also using a statistical approach to predictive maintenance. It is working with EasyJet data on four years of operations from more than 240 A320s to develop algorithms that predict the likelihood and timing of failures. The research correlates operational information such as sensor data, aircraft age, routes and operating conditions with actual aircraft events. The technique is close to practical application and can be applied to many components.

Gourdon says Airbus is committed to exploiting Big Data and that the data and technology to interpret it are available. Most important, airlines both big and small are demanding it.


Boeing has argued aggressively for the benefits of more ample data. The OEM is setting up a Data Analytics Lab with Carnegie Mellon University to improve its capabilities.

Major MROs also have been active. Air France-KLM E&M has just begun servicing its first Boeing 787, and its parent airlines are not operating A350s yet. So it is using predictive maintenance capabilities and experience on 747-400s and A380s to prepare for 787s, says Prognostics Manager Wouter Kalfsbeek. Similarities between the A380 and A350 also should help with the A350.

As both an operator and MRO, AFI-KLM E&M is confident in the benefits of using data from the new aircraft. It has been managing a spares pool for the 787 for three years and knows which systems have the most frequent problems. It is now prioritizing the most critical components, analyzing the available parameters and their potential data sources. It also is asking pool members for more data.

Kalfsbeek gives one example of successful analytics on the A380: the fuel system. Here, the MRO already uses predictive maintenance for feed pumps, transfer pumps and pressure switches. For the 787, AFI-KLM E&M will start applying predictive tools to the cabin air system.

Lufthansa Technik is working along similar lines. But a spokesman says LHT researchers are reluctant to discuss their work, since “it would be of high interest to competitors.”

In tackling Big Data, knowing the right questions to ask can be as valuable as knowing the correct answers.

So where should airline and MRO managers look first for predictive gold in Big Data? Start with the most expensive problems and work back from there. Oliver Wyman partner Chris Spafford points to “high-cost, high-failure components in actuation, power generation, APUs, engine accessories such as starter and fuel pumps, [and] landing gear and hydraulics.”

Spafford also urges attention to avionics, where self-diagnostics can replace boards rather than entire boxes. He estimates there are 150-300 expensive, frequently faulty rotables that cause about two-thirds of delays and cancellations

This article was originally published on December 24, 2015.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.