An auxiliary power unit (APU), as its name suggests, is not as essential as an engine. It is not as much of a differentiating factor as an avionics suite and so, it has not benefited from much product improvement effort by manufacturers. APUs are not at the center of the race for ever-greater aircraft efficiency.
But an APU can be seen as a mini-engine running on the ground. And it just happens that engines are at the spearhead of predictive maintenance —the holy grail in repair planning. In addition, an APU failure, although unlikely to impact safety, can force an operator to delay or cancel a flight. At the very least, to cope with an APU failure, operators have to order ground power equipment and start carts, which can be expensive, if they are even available.
Manufacturers and maintenance service providers are thus applying big data analytics to APUs in a bid to increase their reliability.
Before the engines start, an APU supplies power to the aircraft’s electrical and pneumatic systems. This means the cabin environment is cooled or heated thanks to the APU. The crew performs its preflight checks with power supplied by the APU. Then it assists the main engine start. That is why the first APU was developed in the 1950s, Rich Douglas, senior director of large APU marketing and product management, tells Inside MRO—to provide autonomy to military jets in the field.
There is no requirement for an APU to run in flight, except in two instances. First, if both engines fail, the APU has to restart them. If they do not restart, the APU has to provide electricity and compressed air to the cockpit and cabin, like it does on the ground. “It gives you control of the aircraft,” Douglas says. The second exception is testing for extended twin-engine operations.
A few airlines keep the APU running in flight, according to Douglas. In a short-haul flight lasting half an hour, “switching it on and off would add cycles,” he says.
Apart from those isolated examples, APU manufacturers have failed to persuade their airline customers to use them in flight. In the industry, OEM officials sometimes can be heard suggesting that it would make sense to have each system doing what it is primarily designed for on the aircraft. The APU should therefore provide the flying aircraft with electricity and bleed air, they say, leaving the engine to provide thrust only. It seems the idea of having an extra turbomachine running in flight is not attractive to operators.
Unlike engines, which have enjoyed major architecture and technology changes this decade, APUs have seen few improvements recently. “It is a very classic design; not a lot is happening, nothing extreme,” Dennis Wetjens, managing director of Epcor, says. Epcor is a subsidiary of KLM and is the Air France Industries-KLM Engineering and Maintenance (AFI-KLM E&M) specialist in APU and pneumatic system maintenance.
Lufthansa Technik’s head of APU maintenance, Ole Gosau, concurs. He has seen no significant evolution in design. “It has been essentially the same over recent years,” he says. A special case may be the Boeing 787’s Pratt & Whitney Canada APS5000 APU, which supplies electrical power only to match the aircraft’s “more electric” architecture. The absence of a bleed air subsystem has no bearing on maintenance, Gosau says.
As a manufacturer, Honeywell’s Douglas has a slightly different viewpoint. The new generation of airliners have increased needs in electrical power, which APUs have to meet. “The generator has grown, to 120 kVA for a narrowbody and 160 kVA for a widebody, from 90 and 120, respectively,” which translates into a larger gearbox in the APU, he says.
Another evolution can be seen at the heart of the system. Although environmental regulations do not cap gaseous pollutants, Honeywell has designed a low-emission combustor, Douglas says.
The rotation speed of the APS5000 and the Airbus A350’s Honeywell HGT1700 can vary automatically, depending on the power needed.
Scheduled maintenance on an APU is limited to oil changes and the replacement of some life-limited parts. Usually, it can be operated “a very long time,” Lufthansa Technik’s Gosau says, between two unscheduled maintenance occurrences. “It runs until it fails,” Douglas says, smiling.
So what kinds of damage and failures are typical for APUs? Oil leakage can cause an APU removal, Gosau answers, because of the unpleasant oil smell in the passenger cabin. “Air quality for the passenger is important,” he points out.
Foreign object damage (FOD) is all the more frequent as an APU usually runs on the ground for an extended period of time. A plastic bag in the exhaust recently caused a performance drop on a Boeing 777 APU, Epcor’s Wetjens recalls. “Some environments are harsher than others,” Douglas says. Sand can be ingested, causing excess wear on the blades.
In winter, an impact with a chip of ice can bend an impeller blade, Gosau says. Even worse, anti-icing fluid can burn blades and lead to a failure. Due to the combustion point of alcohol, a blade’s barrier coating does not protect the blade against anti-icing fluid.
Both Epcor and Luft-hansa Technik have repair workshops for APU blades. Work sometimes has to be outsourced to the APU manufacturer —Honeywell or Pratt & Whitney Canada, says Wetjens.
Predictive maintenance is ineffective to counter sudden and unpredictable FOD. But, apart from those instances and because scheduled maintenance is nonexistent, it is an important issue for APUs.
Epcor last September launched a new APU service that “uses big data technologies to monitor the performance of APUs and predict faults,” AFI-KLM E&M says. It builds on a previously existing set of tools, called Epcor Trend Monitoring. “Based on this expertise, and after upgrading its databases and IT platforms, Epcor has now developed Prognos for APU, a real-time APU performance-monitoring and analysis solution that can identify weak signals and predict potential failures,” the company explains.
Prognos has already proven its effectiveness, according to AFI-KLM E&M. Before it was made available for APUs, the first systems monitored by Prognos for Aircraft figured in the top five causes of flight delays and cancellations. Thanks to data from Prognos, these causes are no longer so problematic, the company says.
Mostly, an APU has to be removed from the aircraft—to be repaired in a workshop—when its performance drops, says Wetjens. “That’s why we invest heavily in predictive maintenance; we would like to advise airlines to remove the APU before it is too late or costly.” The work for Prognos for APUs started early in 2016, and “we keep improving the tool with incoming data,” Wetjens adds. AFI-KLM E&M is being helped by companies specializing in big data analytics, as well as universities.
A couple of months before the Prognos for APUs launch, Honeywell rolled out its GoDirectMaintenance program for APUs, which uses connectivity and data analytics to “improve maintenance operations and reduce equipment downtime.”
Customer airlines provide Honeywell with access to data already being generated on the aircraft. “We use sophisticated analytical tools to examine and analyze the information in almost real time and provide predictive maintenance alerts before the failure occurs,” the OEM says.
During a trial program with Cathay Pacific—one of the first customers— the service demonstrated that it can reduce inoperative systems by up to 35% and reduce APU-related delay minutes by 51%, Honeywell says.
The company emphasizes that, for an operator, signing up does not involve adding any sensor or “a single ounce of weight to the aircraft.” The service works by using an existing data connection on the aircraft to download APU maintenance and fault data. That fault data is then shared with Honeywell and analyzed before being presented to the carrier’s maintenance team. The data identifies what maintenance actions are required to fix the APU and avoid an unscheduled maintenance event.
The goal is to predict impending failures in line-replaceable units, providing at least three days’ notice. “We tell the operator when to pull the parts and why; we take responsibility for our recommendation,” Douglas stresses.
With this new form of data-sharing, the program heralds the aerospace industry’s digital transformation. It is now available on the Airbus A320 and A330. Next in line are the Boeing 737 and 777. Later on, it will be extended to newer platforms like the Bombardier C Series and Airbus A350. “You want to have a lot of fleet data to ensure the reliability of your decisions,” Douglas notes.
The Boeing 767 and 757 will not be added to the list, as they do not generate enough data to download and analyze.
In addition to connected maintenance for APUs, Honeywell is looking to expand the service to other parts and systems of an aircraft, including wheels, brakes and environmental control systems.