GosNIIAS
The project coordinator
Summary
Analytics
Project map
Russian version
Home
English version
Write the mail
Predictive data analytics: the digital approach to maintaining airplanes
News/ > 2018/ > Predictive data analytics: the digital approach to maintaining airplanes/
Predictive data analytics: the digital approach to maintaining airplanes
17 September 2018 New data shows the top driver for investing in aircraft connectivity upgrades is the enablement of predictive aircraft maintenance.

Honeywell Aerospace’s June 2018 survey confirms an upward trend in airlines looking to the technology to predict when critical components and systems will fail and replace them before the failure would cause a delayed takeoff or arrival. 88% of the surveyed 106 commercial airline technology purchasing managers and decision-makers consider investment in connected aircraft maintenance to be a top priority.

Data analytics-driven aircraft maintenance prognostics are changing the way airlines monitor the health of their airplanes. Historically, the airline maintenance teams captured data that then took hours, days or even weeks to analyze once a plane had landed. Now, most of that process can occur while the aircraft is still airborne.

Predictive data analytics adoption at UPS

UPS Airlines’ new fleet of Boeing 747-8s feature an integrated data management unit (IDMU) that uses a PC card for recording QAR/DAR data. That unit also has an Ethernet connector and is fully user-programmable using a Windows-based application generation software. This functionality allows air carriers to customize the various elements of their aircraft condition-monitoring system applications, such as QAR/DAR output data maps, to airline maintenance technicians.

The 747’s data bus, networking and data-processing capabilities are a generational leap compared to UPS’ older 747-400s, which rely on a central maintenance computer to capture and store faults that an airline maintenance technician analyzes post-flight.

The new 747s are the air cargo carrier’s first «e-enabled airplanes,» said Tom Wagner, manager of aircraft maintenance technical support and data analytics at UPS.

«The aircraft that have central maintenance computers like our 747s, or the centralized fault display system (CFDS) like an MD-11, those were intended for maintenance crews to get on board and physically remove and then analyze data,» he explained. «We’re now tapping into that data in the air before the aircraft lands.» >>>
>>> UPS is showing that an air carrier’s transition from reactive to predictive maintenance is fueled by data analytics software. The key for an airline maintenance and data analytics team is to produce a graphical picture of large amounts of actionable information from the aircraft data by means of ACARS and non-ACARS data transmission channels.

The Louisville, Kentucky-based carrier does this by using a combination of data analytics, processing and visualization tools. UPS in 2017 started using AirFASE, a flight data monitoring software developed by Airbus and Teledyne Controls to translate raw flight data into engineering values.

As part of the back-end support functionality for the data-capturing and storage capability provided by AirFASE, Teledyne is using IBM for its cloud computing needs and has expanded support of cloud-computing analysis and storage for operators, said Edgar Salvador, director of aircraft data services at Teledyne. «We recognize that not all airlines have the resources to maintain an in-house data analysis facility,» he said.

UPS also uses an internally developed aircraft telemetry system that is connected to the aircraft condition monitoring systems on each aircraft. That telemetry system and Boeing’s airplane health management (AHM) analyzes the ACMS data and produces alerts to the air carrier’s maintenance team who can then order replacement parts for aircraft before they fail in the air. AHM also provides maintenance technicians with every service letter, fleet team digest and maintenance tip that has ever been generated within the per-flight operational history of each fault code on each Boeing aircraft model operated by the airline.

AirAsia wants to increase the percentage of data it uses

Cloud computing is also driving the use of predictive maintenance analytics for operational efficiency gains at AirAsia. The airline is in the process of upgrading its fleet of Airbus A320s with FOMAX — a new onboard data-capture/transmission module, a key component of Airbus’ Skywise predictive maintenance service.

AirAsia, like UPS, is also using an assortment of data analytics applications, including Google’s BigQuery, Data Studio and Cloud computing to acquire and analyze data from multiple sources, explained AirAsia Deputy Group CEO Aireen Omar at Google’s Cloud Next conference in San Francisco in July.

BigQuery uses Google’s cloud storage service to sort data and provide maintenance technicians with the ability to run faster searches of large data sets, Omar said. >>>
>>> «The key thing is to integrate all this data from various sources and be able to combine those data and make meaningful algorithms out of it,» said Omar. With the software, the airline was able to reduce the number of aircraft on the ground and reduce operational costs by 10%, she added.

Even more noteworthy is that AirAsia currently only uses about 20% of the captured data about the performance of its aircraft and its overall operations as a whole.

Outside of Google and Airbus, AirAsia also works extensively with GE Aviation’s flight efficiency services (FES) division for precision navigation services, flight data analytics and fuel management services.

Virtual QARs at Emirates

Emirates will add a new virtual quick access recorder (QAR) to its fleet of Boeing 777s, per an agreement it signed with UTC Aerospace Systems at the 2018 Farnborough International Airshow. Under the agreement, the fleet will be upgraded with an aircraft interface device (AID) and new wiring for power and access to flight data. The virtual QAR is an application hosted on the AID and records flight data and ACMS reports received by the AID. That data is then prepared and transmitted over available communications links based on Emirates’ communications rules.

This is improving Emirates’ collection of flight data by reducing the amount of data loss and improving timeliness of the receipt of the data, said UTC Engineering Director Mike Haukom.

«By collecting the flight data on the aircraft and then transmitting it from the aircraft to the ground, the loss of data associated with manual retrieval is reduced significantly,» he said. «By increasing the frequency of the transmission to after every flight, instead of on a once-a-week frequency, the data is available much sooner for processing.»

Before transitioning to the virtual QAR, Emirates’ process for retrieving flight data was completely manual, with technicians removing portable media devices from the aircraft, in the form of a PCMCIA cards or USB drives. That data would then be offloaded to a server on a once-per-week basis. That type of schedule lead to the discovery of data leading to the airline learning of a part failure about to occur or a hard landing happening up to a week after that data was first retrieved. Now, the flight data is instead saved on the AID as a backup, and the data files are transmitted after each flight, allowing for the maintenance team to review it in near real time. >>>
>>> Real-time decision-making?

At UPS, the transition to the full-scale adoption of a predictive maintenance process is extensive. The maintenance team does still find some limitations to the decision-making capabilities have available based on the ability to receive real-time information about the aircraft.

Randy Miller, who leads the data analytics program for the MD-11 fleet at UPS, gave some perspective on this during the UPS presentation at the 2018 Global Connected Aircraft Summit. According to Miller, one of the limitations UPS faces is in its ability to dispatch aircraft using ACARS information. He described a situation where a flight crew is preparing to taxi down the runway and they receive a fault message that reads, «Engine 1 Fire Detect.» Maintenance and operations procedures associated with the flight crew encountering that message require the aircraft to return to the gate where a maintenance crew member must then extract the centralized fault display system information to determine which fire detect system has a fault, the A-loop or B-loop.

Miller can already see that same fault information in real time remotely within his flight data monitoring software. However, because the information is transmitted over ACARS, it is not considered to be secure enough for to actually lead to the decision to dispatch the aircraft despite the crew seeing the fault notice.

«For those who are in the profession of improving security, if you can figure out a way to secure that level of ACARS information, maintenance information, I can use this on the fly and prevent a delay by allowing the crew to do the dispatch. There’s a number of variables out there by which I can use this. Right now I’m waiting on that approval from Boeing and the FAA to get that done,» said Miller.

For more information, please visit the following links:
http://interactive.aviationtoday.com/avionicsmagazine/august-september-2018/
predictive-data-analytics-the-digital-approach-to-maintaining-airplanes/
.

Search on the project
Look!
© 2018 State Research Institute of Aviation Systems. All rights reserved. Terms of Use.