Repairing or replacing avionics components is hardly the biggest cost in maintenance. Avionics do not wear out regularly like some other aircraft systems. They just break down, often unexpectedly.
Unexpected maintenance is more expensive than scheduled repairs. And there is another frustrating maintenance problem that seems to afflict avionics components disproportionately.
An avionics line replaceable unit may be removed because a pilot or mechanic finds something faulty in its performance. But when it gets back to the shop, it tests as perfectly good, or is “No Fault Found.” These NFFs cost labor time to remove, logistics costs to move—often long distances—expensive time to test on costly equipment and require burdensome increases in inventories. Far worse, NFFs may cause very expensive delays or flight cancellations, apparently for no good reason.
NFFs are common in avionics and can be a major portion of components sent to shops for ispection. A numerically smaller problem is the really bad behavers. Sometimes called “rogue” units, these are components sent to shops repeatedly and frequently for exactly the same reason, but they always perform like stars on the test bench.
The entire industry has been fighting NFFs and hunting rogue units for some time. Effective approaches require the collaboration of many parties, the airlines that remove NFF units, the internal or third-party shops that test them and manufacturers of both individual components and the aircraft of which they are parts.
Arinc 672, “Guidelines for the Reduction of NFF,” lays out the problem in conceptual terms. But NFFs and rogues must be tracked to their causal lairs by extremely detailed practical steps. Some airlines and shops work very hard to gain even incremental progress. Others don’t try very hard.
Jim Saltigerald is an Air Wisconsin reliability analyst who is active in drafting the next version of Arinc 672. “Speaking from an airline perspective, in some circles NFFs are regarded as an unavoidable phenomenon, whose financial impact is less than that of trying to find NFF reduction remedies,” he says.
Saltigerald disagrees with this passiveness. He estimates that avionics components removed as faulty for any reason and found NFF in airline shops can be half or more of avionics components sent to shops if no corrective actions are taken. Based on experience at Air Wisconsin, he believes that rate can be cut by half or two-thirds.
To do that, Saltigerald recommends initially focusing on low-hanging, or high-cost, NFF fruit. Concentrate first on NFF items that cost a lot, even if they are infrequent. Next, work on NFFs that are low-cost but frequent. The last priority is the low-cost, low-removal frequency NFFs. “Don’t chase stuff that has no value, you want to get the biggest bang for the buck.”
Air Wisconsin began using this three-tiered approach in 2005. It first found the top 10 NFFs in terms of total cost to its operations. Seven of the 10 had double-digit removal rates annually, up to 29 NFF removals per year for a single component type. The carrier has reduced those rates to four to six removals per year. Now only four out of its 10 worst NFFs are removed 15-20 times per year. The rest are in the single-digit NFFs.
So progress can be made, but it’s not easy. Better troubleshooting helps. Saltigerald says NFF problems can trace to the piece-part level, the line-replaceable-unit level, the system level or the aircraft level.
Air Wisconsin designated a core team to deal with NFFs, based at its headquarters and linked to the entire maintenance environment. Members came from flight operations, maintenance control, line operations, component management and shop operations. The team had five core members. They did not work full-time on NFF, but met weekly to address NFF problems. Other experts were also tapped. “You have to have a structured approach,” Saltigerald insists.
Arinc 672 gave NFF staff a common language and a common matrix to help locate root causes. Common language is essential because solving many NFFs involves at least three parties: an airline, airframe OEM and component OEM.
A structured approach and core team institutionalize NFF experience. “You need to build a knowledge base and good documentation because you don’t want one employee to take all the NFF knowledge when he leaves,” Saltigerald says.
Data on removals is essential. He says it should contain as many relevant parameters and cover as long a history as possible, back to birth if feasible. Airlines usually have the most important data, but component OEMs do not always get this data to help design out NFFs in the future. Contract language or mutual interest should ensure necessary data are shared.
Air Wisconsin was especially concerned with rogue units. “Rogue units can ruin an inventory,” Saltigerald says. “Good parts chase rogues and mean time between unscheduled removals [MTBUR] suffers.”
NFFs and rogues can occur in any avionics area. “Anything with buttons you touch and moving parts can be in the top 10,” Saltigerald notes. “But sometimes it’s computer chips and PC boards that you are not even touching. Each NFF situation is different and dynamic.”
Saltigerald has high hopes for the future. He says good work is being done on electrical wiring interconnection systems (EWIS) to find factors that contribute to NFFs and hidden-failure modes. He hopes RFID tagging will alleviate some problems by recording part and serial numbers, how long a part has been on an aircraft and possibly past findings by mechanics.
Other technology also looks promising. Better data analytics, fault isolation, troubleshooting and predictive maintenance tools should help. So should better flight data recorders that can download data in real time on exactly what pilots are seeing on cockpit displays. “Knowing the phase of flight when certain faults appear would help,” Saltigerald says.
He also hopes the next revision of Arinc 672, due out in 2015 or 2016, will help by providing a common syllabus of what everyone should know in dealing with NFFs. Online training in NFF management might also help. And Saltigerald would like to see innovators, designers and engineers collaborate more closely on preventing future NFFs.
Major airline-affiliated MROs also have extensive experience with NFFs, and some have made dramatic progress. Arinc 672 is still the best guide in principle for dealing with the problem, according to Taco Vingerhoed, avionics and accessories director at KLM Engineering and Maintenance. “The challenge is putting principle into practice,” Vingerhoed says.
The three steps urged by Arinc are assembling complete data on all NFFs, tracing their root causes and taking corrective actions. The AFI-KLM exec says the data are generally available; for example, on part number, tail number, serial number, flight hours and shop findings. And corrective actions are usually straightforward after root causes are identified. The hard part is finding those root causes.
Vingerhoed manages avionics repairs for Air France, KLM and all the airlines whose components depend on the MRO’s pool support. Avionics NFFs are now below five percent of all avionics components turned in for repair, and this portion does not differ much according to whether components come from external or internal customers. The NFF rate is down from seven years ago, but is still well above the 1-3% Vingerhoed is aiming at. “But we are very confident we will reach that figure in the near future.”
The top-three NFFs now are engine control units (ECU), display electronics units (DEU) and air data inertial reference units (ADIRU). To find root causes for these NFFs, the MRO collaborates with both internal and external customers. As AFI-KLM usually charges a flight-hour rate to its external airline customers, the MRO bears NFF costs and so has an intense interest in reducing the NFF rate. It obtains data from external customers and talks to their managers when investigating NFFs.
Modern IT systems help a lot with data, but investigation of root causes requires man-hours of work, much of it by highly skilled and well-paid engineers. Thorough investigations also require help from avionics OEMs, and Vingerhoed says most cooperate fully. Especially at annual aviation maintenance conferences, techs and engineers exchange information and look at common problems. Nevertheless, “not every OEM is as cooperative as it could be,” Vingerhoed says. “I want open books from OEMs.”
AFI-KLM E&M obtains avionics testing software from OEMs and usually uses this software. But where OEM software is unavailable or too slow, the MRO may develop its own. For example, some OEM software requires 48 hr. for a thorough component test. The MRO wants to test as fast as possible, so it developed software that needs only six hours to validate results. “It’s in our interest to make testing as fast and cheap as possible,” Vingerhoed stresses.
For almost all NFFs, the MRO analyzes data and looks for root causes using simple methods. But some avionics components show up as NFFs three or more times in 18 months. This one percent of NFFs prompts much more labor-intensive investigations.
Here the MRO will look at the station, the pilots, cabin crew, logistics, the whole chain of responsibility and possible influence. It first takes a high-level look and, if that does not yield conclusions, goes a step deeper and finally—if necessary—a further step. This last step may look at crews and staff at each station. “Normally, we do not go that far,” Vingerhoed says. The full, three-step investigations are done one to three times per year.
Root causes fall into three main categories: 60% originate in the component itself, 30% in the aircraft and 10% in AFI-KLM E&M’s own shops.
Vingerhoed hopes that new-generation aircraft will reduce the NFF rate, as their avionics components have more self-testing software built in. But KLM does not fly the 787 yet, and its only customer operating the 787 has been doing so for only a year and nine months, so there is not enough data to know if that hope has been realized.
A version of this article appears in the November 3/10 issue of Aviation Week & Space Technology.