The third annual survey conducted by Aviation Week and the International Air Transport Association on airline spending plans for maintenance IT shows investment intentions are shifting, with areas such as data analytics becoming a greater priority. In responses from 52 carriers and four independent shops, less than one-fifth stated they plan to replace their maintenance or enterprise resource planning (ERP) systems in the next five years. Nearly half plan to enhance their MRO or ERP systems, and less than one-third will spend chiefly to sustain current systems.
Of the carriers planning to enhance IT, data analytics has emerged as a priority—up significantly from last year. About 73% of respondents are planning to invest in this area, followed by maintenance planning and document management, both tied at 69%. Also important at more than 60% were e-signatures and mobility, both of which were also popular priorities in 2015.
More than 94% of airlines seeking better analytics want to move toward predictive maintenance, and more than 82% wish to monitor aircraft health as well. This emphasis on predictive maintenance may indicate that some airlines have already begun monitoring engine and component health and now want to take the next step, converting unscheduled to scheduled maintenance.
Significantly, the most intense area of interest, shared by all airlines seeking better data, is improving airframe maintenance. Here again, the slightly lesser number of respondents concerned with engines and components may simply mean that OEMs and others are already working on these maintenance challenges.
When asked what they expect to gain from better analytics and the improved maintenance it would enable, almost all expect cost reductions and better reliability, while clear majorities also expect lower inventories and shorter turn-times.
Data analysis to improve maintenance is clearly a joint effort. Most respondents said they would improve their own systems, but a majority also said they would rely on OEM systems, MROs and other third parties to help as well.
Given the importance of collaboration in doing better analytics, it is not surprising that the major challenge to improvement—mentioned by the great majority of carriers—is insufficient standardization of data. Substantial numbers also see marginal economic returns and other IT priorities as hurdles. Interestingly, for all the worries about it, only a minority said data ownership itself is a hindrance.
In any case, OEMs and MROs are upping their game as both become collaborators on and providers of maintenance analytics.
OEMs Have Design Experience and Scale
Caroline Pereira-Garcia, marketing manager for maintenance and engineering applications, says that Airbus’s Prognostics & Risk Management (PRM)—now available for the A330— will be tested with several airlines on the A320 for a launch in mid-2017. PRM for the A350 could be released in early 2017 and for the A380 toward the end of 2017 or early 2018.
PRM algorithms use data from aircraft health management systems to detect degrading trends that foreshadow component failure. The tool focuses on systems with the biggest impact on reliability. For A330s, these are fuel, hydraulic, air conditioning, landing gear, oxygen, electric and pneumatic systems. Yet not all components in each system are monitored. PRM focuses on leaks in hydraulic, oxygen and pneumatic systems, pressure-regulating valves in pneumatics, and extension and retraction times for landing gear.
The A320 generates less data, so its PRM will exclude data from fuel and air conditioning systems, while PRM for the A350 and A380 will be much like that of the A330. So far, PRM downloads data at the end of each flight, not during flight.
Airbus has demonstrated that PRM could avert about 15% of unscheduled removals on the A330. During its use for Delta Air Lines from spring of 2015 to spring 2016, PRM flagged 11 pneumatic pressure-regulating valves for early removal, all of which turned out to be faulty. This zero rate of false alerts is important in securing airline trust in the system and thus promoting its use.
Along with extending PRM to more models, Airbus is also working with EasyJet, IBM and data miners at Palantir Technologies to see if PRM can be improved by adding much more data from the aircraft to the aircraft condition monitoring system (ACMS) data. The current PRM chiefly uses advice from design engineers on which triggers to track and use for alerts. The new project will exploit data from digital aircraft recorders as well.
At present, PRM does not track engine data, which is left to engine OEMs.
Having secured Delta as a PRM user for its A330s, Airbus is now talking to the operators of more than 1,200 other A330s flying. Pereira-Garcia says the biggest challenge is convincing an airline to change its business processes to move from reactive to predictive maintenance. Significant changes must be made on the line, in hangars and in the supply chain to fully exploit PRM, and the airline must evaluate gains at all these stages to recognize potential benefits. Further, engineers, “must trust the system and know there will not be a lot of false alerts,” she says.
In 2015, GE Aviation moved its analytics platform onto Predix, a software platform for the Industrial Internet. Fleet operations director Matt Ridgway says the move has increased detection of engine problems, shortened lead times and improved accuracy of alerts.
GE now analyzes data on 35,000 GE and CFM engines and more than 100 million flights annually. All GE, CFM and Engine Alliance engines are monitored. Troubleshooting generally begins with a review of recent flight data to identify diagnostic signatures that signal failure modes and then sending specific maintenance recommendations.
In mid 2016, GE introduced several new software applications. Flight Phase Analyzer uses continuous engine data to analyze taxi, takeoff, climb, cruise and ground idle, comparing each to metrics for global fleets. It also looks at variation in city-pair flight durations. GE’s Water Wash Optimizer calculates gains in exhaust-gas temperature and time on wing achieved by washes and helps customize washes to save fuel.
GE also collects takeoff, climb and cruise snapshots from engines, more than 2,800 records per engine annually. Hundreds of parameters are sampled, including exhaust-gas temperature, core speed, fan speed and fan vibrations. Ridgway says the best way to get data is in near real-time by the Aircraft Communications Addressing and Reporting System (ACARS), but GE also uses wireless ground links, Secure File Transfer Protocol (SFTP) and email.
Engine monitoring requires many GE groups to create new detection modes, improve existing algorithms and perform day-to-day monitoring. GE Centers in Cincinnati, Shanghai and Dubai do the monitoring, while on-site teams help airlines understand maintenance recommendations.
Ridgway estimates engine monitoring now help operators to avoid more than 4,500 service interruptions annually. Analytics also revealed a GE90 component that was limiting time on wing for a Middle Eastern carrier. De-rating during climb and optimizing water washes reduced stress on the part and reduced fuel burn, saving the airline $7 million a year.
He expects gains to increase in the future, lengthening time on wing and saving money with smarter work scopes during shop visits. GE’s advantages are expertise in both data science and engine physics, plus a huge fleet being monitored. Of course, eventually GE plans to support much more than engines.
UTC Aerospace Systems (UTAS), the largest aerospace systems company, makes cutting-edge sensors and has a growing team of analytics experts. UTAS now works with over 50 major airlines on analytics for aircraft maintenance.
Ajay Agrawal, vice president-aftermarket, says UTAS is uniquely positioned for monitoring aircraft systems from end to end. For example, UTAS can monitor its air-conditioning, electric power and generation, electric braking, fire-protection, and nitrogen-generation systems. It monitors data on many aircraft, including the A320 family, A380s and Boeing 777s and 787s. In all, UTAS monitors components on more than 1,200 aircraft, either via ACARS inflight or data downloaded on the ground, depending on airline and aircraft capabilities.
UTAS relies on design, aftermarket and field support engineers to develop predictive analytics, and it works with United Technologies Research Center scientists to meet the biggest challenges with leading-edge mathematics. Potential gains include reducing unscheduled maintenance, reducing no-fault-founds and extending time on wing, thus reducing total maintenance costs.
Agrawal says UTAS wants to be the preferred provider of intelligent components enabled for prognostics and aircraft health management. Its Pulse Health Monitoring System combines and analyzes many onboard sensor feeds, and its Aircraft Data Management tool can add other aircraft data, displaying results on cockpit or tarmac-based tablets or transmitting data to ground analysts.
UTAS has been developing intra-aircraft wireless communication among sensors, actuators and processing nodes. A new spectrum band and standards for this communication will eventually ease deployment of more sensors and better prognostics. Agrawal expects UTAS monitoring and prognostic capabilities to increase with newer aircraft.
He argues that UTAS is best suited to provide intelligence on its components. Its Aircraft Interface Device, certified on nearly every large aircraft, can collect, transfer and analyze data on mixed fleets.
Operations and Shop Repairs
In 2017, Lufthansa Technik (LHT) will launch Condition Analytics, combining condition monitoring and predictive maintenance. The MRO has been monitoring the condition of its own Boeing aircraft—based on fault messages and ACMS reports—for several years and has been working on predictive maintenance based on full-flight data from direct-access recorders and quick-access recorders. Program manager Holger Appel says the two approaches were combined in late spring 2016 to “provide the user with all needed information in one tool.”
LHT currently monitors components such as flight-control surfaces to save fuel, radar altimeters to avoid inflight failures and igniter plugs so removals can be based on condition rather than fixed intervals.
Condition Analytics will start in early 2017 with the A320 family, 747-400s and 747-8s, then expand to cover other Airbus and Boeing models. LHT aims to include all mainline jets by the end of 2017.
Condition monitoring will continue to use inflight fault messages and ACMS reports, and predictive maintenance will use the full-flight data as currently collected by airlines. But Appel says LHT can customize the tools to suit the needs and constraints of each airline. He says early results are “very promising” for reducing costs and increasing reliability and safety.
Appel says LHT’s operation and maintenance activities enable it to provide a more complete maintenance tool. For example, by predicting component removals and then detecting no-fault- founds, LHT can refine its predictive tools. He believes the best tools must exploit all data, engineering knowledge, maintenance data, operational data and shop results, all analyzed by data scientists. “Our big advantage is that we have everything needed,” he notes.
AFI-KLM E&M recently launched Prognos to exploit aircraft data to improve maintenance. Innovation Director James Kornberg says the MRO has just developed a much better tool for monitoring all the engines it supports, including GE90s; CFM56-7s; -5Bs and -5Cs; CF6s; GP7200s and later, Trent XWBs. The new tool was adapted from a widely used application. “We use an industry platform which has been adapted and customized to support several alert and analytic models,” Kornberg explains.
The MRO provider is also developing analytics for non-engine components on A380s, 747s and 787s. It will offer a service that Kornberg says will complement support offered by OEMs, based on AFI-KLM E&M’s experience with cancellations and delays.
The new service will focus on monitoring systems that cause the most cancellations and delays and that have the most data for monitoring. Kornberg says this focus will differ by aircraft.
Some data will be taken inflight from ACARS, while other data will be downloaded at gates. AFI-KLM E&M can either store and analyze data at its own site or do the work on customer premises if that is preferred.
He would not estimate potential gains from either engine monitoring or more general component analytics. But Kornberg insists that the business case is good, based on AFI-KLM E&M’s own experience in attacking the problem.
Kornberg acknowledges there is plenty of competition on analytics now, which is good for airlines, and argues an airline MRO enjoys some unique advantages, for example in operations and in shop repairs down to the sub-assembly level.
Digital Extra See our buyer’s guide to MRO Software & E-Enablement Tools at: AviationWeek.com/MROIT