Predictive maintenance is a journey, not a sudden revolution. It must begin with certain aircraft, emphasize certain components and proceed in careful, logical steps.
For instance, Etihad Airways has established a project to introduce predictive maintenance on selected types of its aircraft, notes Paul Kear, senior vice president Technical. “We have already introduced Airbus Skywise predictive maintenance on our A320s and A321s,” Kear says. “For some specific systems, we have also developed in-house solutions that use system data to support condition-based maintenance.”
The carrier has retrofitted 15 of its single-aisle aircraft with technology called FOMAX, for Flight Operations and MAintenance eXchanger. This enables the aircraft to transmit information for up to 24,000 parameters to support predictive maintenance.
The predictive technology monitors several systems and components on these aircraft, identifying both chronic problems and system-related faults. Kear says Etihad prioritizes monitoring those systems that can significantly impact technical dispatch reliability and that provide sufficient data to support reliable monitoring and alerting. “These include air-conditioning systems, landing-gear proximity sensors, anti-ice valves and hydraulic systems.”
For predictions, Etihad uses Airbus’s Skywise and its internal data analytics solution, called I-Health, a real-time prognostic tool that creates self-learning processes that can be applied to trend calculations.
For predictive-maintenance algorithms to produce maintenance alerts, the algorithms need a continuous data feed, Kear says. “This data is basically the values of a particular aircraft parameter obtained at certain intervals throughout the flight.” Etihad thus uses a number of data sources, including the aircraft condition monitoring system, or ACMS, FOMAX and flight data, that are all linked to one common data frame package in which self-learning and trending commands are processed.
On challenges involved in moving toward predictive maintenance, Kear notes that FOMAX is a new unit that must be installed in the avionics compartment of the Airbus jets. “It requires specific wiring and antenna installations that take about two days, with the first installation being especially challenging.”
Furthermore, predictive maintenance alerts must be analyzed by Etihad’s maintenance teams to determine the most appropriate actions based on information available. “Our teams study information obtained from a range of sources, which is very time-consuming,” Kear explains. Moreover, a number of algorithms still require adjustment to reach maturity and avoid no-fault-found component removals.
Finally, environmental impacts also pose challenges. “Issues such as maintenance deviations or modifications, hot temperatures and high levels of humidity mean our engineers are operating mostly at the limit of certain components’ specifications,” Kear says.
Nevertheless, Etihad has made substantial progress. It has already created virtual failure corrections that eliminate deviations from environmental impacts. It is also using past failures in a self-learning algorithm that predicts behavior of some components.
“We have also had success with wing anti-ice valve fault predictive alerts, and gear proximity sensor fault detection,” Kear says. “A solution developed in-house for air-conditioning systems has also streamlined data collection, analysis and maintenance planning.”
Etihad is now focusing on translating scheduled maintenance tasks into predictive tasks. “We continue our efforts to mature algorithms and explore our internal data analytics engine for in-house capability development,” Kear says.