This article is published in Aviation Week & Space Technology and is free to read until Dec 01, 2024. If you want to read more articles from this publication, please click the link to subscribe.
Digitalization has been one of the major trends in aviation maintenance over the past two decades, and at its heart lies the burgeoning capabilities of health monitoring systems. Each new generation of aircraft and engines features more sensors than the last. This, alongside new machine learning software analysis, enables analysis of more data points for increasingly accurate preventative and predictive maintenance alerts.
Rolls-Royce, for example, is transitioning its engines and health monitoring systems to a new platform under its IntelligentEngine concept. The system promises an interconnected infrastructure of fleetwide engine data feeds, maintenance technology such as inspection robots and digital twins—the bit-based replicas of in-service engines.
“This allows for an increase in the number of parameters monitored by utilizing the continuous data available, as well as traditional fixed event and fault messages,” says Nick Ward, vice president of digital systems at Rolls-Royce. “These new data points combined with situation data allow us to create new diagnostics to prevent potential in-service events.”
Health monitoring is also useful for maintenance and inventory planning. This potentially brings great benefits amid the supply chain crisis, as it allows operators and MRO providers to order parts well in advance and get ahead of long lead times, rather than having a heavy check held up by an unexpected fault found with no spare to replace it.
AIRCRAFT MONITORING
While engine OEMs pioneered health monitoring systems, aircraft manufacturers are catching up and using their own technology to create optimized maintenance programs. For instance, rather than sticking to a rigid maintenance planning document, airlines can use service history data to create a customized workscope that better fits their needs and allows for more efficient shop visits and better-timed heavy check intervals.
Embraer’s Aircraft Health Analysis and Diagnosis (AHEAD) platform provides real-time monitoring of critical aircraft systems including the auxiliary power unit, fuel system, bleed, hydraulics, avionics and flight controls across both the E1 and E2 lines.
“Leveraging data analytics and predictive algorithms, the platform provides proactive maintenance, enabling operators to reduce unscheduled downtime,” says Carlos Naufel, president and CEO of Embraer Services & Support.
The AHEAD platform can also detect and alert operators of adverse exceedance events, ensuring timely maintenance interventions. In addition, Embraer is evaluating initiatives such as drone-based inspections to enhance the precision and efficiency of maintenance processes.
“Flight data analysis for exceedance events accelerates the identification of potential issues, significantly reducing aircraft downtime,” Naufel says. “In parallel, ATA reliability trends offer actionable insights for scheduling proactive maintenance, including servicing, before unplanned disruptions occur. By analyzing key system data, operators can better anticipate maintenance needs, optimize aircraft availability and improve overall fleet reliability.”
Airbus says half of its aircraft incorporate health monitoring systems. The newest models report 10 times as many parameters thanks to systems like Flight Operations and Maintenance Exchanger and RMAX, an interface device for downloading huge volumes of operational data from an aircraft.
These systems are more advanced and provide more information than a traditional aircraft condition monitoring system (ACMS), which cannot provide the continuous data needed to inform predictive maintenance.
“We need much more data than ACMS because these ACMS reports have not been designed for predictive purposes,” notes an Airbus spokesperson. “To feed efficient predictive maintenance models, we need time series—the continuous recording of parameters at very high frequency—to ensure proper and accurate predictions.”
All new aircraft rolling off Airbus lines are delivered with these capabilities, while the technology needed to enable time series reporting is available as a retrofit for older models.
Airbus presents the data and analysis to operators via its Skywise health monitoring platform, which has 10,000 users. Skywise’s main function is to allow an airline’s maintenance control center to ensure the aircraft’s dispatch by anticipating and preparing necessary maintenance actions while the aircraft is still flying. For example, based on continuous monitoring data, an airline might opt to divert a flight to ensure it lands at a location with suitable maintenance resources to avoid an aircraft-on-ground situation. Furthermore, airline technical departments can use the tool for engineering and troubleshooting analysis.
Airbus also feeds the data back into its own product development. “We have reduced by 50% the time it takes to identify a major in-service problem and provide a fix to our customer,” an Airbus spokesperson says. “Originally, such a fix was only addressed with a modification of the aircraft, requiring the creation of documentation like a service bulletin, the availability of physical kits and also an approval from the authority.”
ENGINE MONITORING
As aircraft manufacturers step up the capabilities of their health monitoring systems, engine-makers are looking to integrate their own dedicated engine health monitoring technology.
“The new high-frequency parameters combined with environmental or situational conditions outside the engine have allowed us to introduce new analytics,” Rolls-Royce’s Ward says. “We can configure the parameters being recorded to enhance our analytics development. This technique still uses the event-based data and fault message information and combines it all to increase the effectiveness. The speed at which we can detect and alert events is increasing with the new platform and processing we are introducing. Overall, this is supporting extended time on wing and reduced disruption for our customers.”
The state of the art for Rolls-Royce is its engine vibration and health monitoring unit (EVHMU), which is being introduced on its Pearl business jet engine before being rolled out to the commercial fleet.
Enabled by a connection to the Internet of Things, the EVHMU can provide instant access to around 10,000 engine performance and health parameters with unprecedented levels of data quality, Rolls claims. Working in conjunction with aircraft health monitoring, the EVHMU system looks at pressure, temperatures and vibrations, but can also monitor the condition of line replaceable units. This enables engineers to remove a part before it develops a fault. The system also provides bidirectional communications, allowing for remote reconfiguration of engine-monitoring features from the ground.
Alongside the hardware installed in engines, software is also advancing. CFM International announced in April that it was using machine learning to enhance the analysis of the data provided by its engines. The system now being used for CFM Leap 1A and 1B engines models data from multiple engine sensors at takeoff, climb and cruise via probabilistic diagnostic and prognostic machine learning tools. These tools then provide targeted alerts based on known engine operating signatures, with CFM touting big detection rate and accuracy improvements.
“We have achieved 60% earlier lead time in identifying preventative maintenance recommendations [and] a 45% increase in detection rates, and cut the number of false alerts in half over the past decade,” says David Harper, fleet support director at CFM parent GE Aerospace.
Rolls-Royce also reports fewer false alerts and more accurate predictions through using machine learning and artificial intelligence (AI) to combine and analyze a greater number of sensor inputs. “We have been able to create new advanced analytics that would not have been possible previously to prevent in-service disruption,” Ward says. “With the new platform we can deliver these new diagnostics quicker and with a well-defined success ratio—the obvious limitation being on older engines that do not allow for such granularity in data.”
Ward expects more maintenance actions to be generated by sensor data than by physical inspections, thereby promoting efficiencies and time savings in MRO. “The industry has always strived toward this,” he notes. “Even with the potential for spurious sensor alerts, physical inspections also attract a risk and cost factor.”
MRO VALUE
Like the OEMs, many large MRO providers offer engine health and trend monitoring to optimize maintenance and overhauls for their customers. MTU Maintenance conducts health trend monitoring for about 2,000 engines with WebETM 3.0, its engine trend monitoring tool accessible via the myMTU platform.
“The system observes various aircraft engine parameters, such as exhaust gas temperature, fuel flow, shaft speeds, oil parameters and bleed settings, and detects abnormalities,” says Wladimir Bickel, senior manager for MRO digital transformation at MTU Maintenance Hannover. “Any trend deviations are alerted to our expert engineering team, who will assess the data and make qualified recommendations for a course of action to our customers.”
One of MTU’s main goals is optimizing engine shop visits and holistic engine maintenance management. “The combination of our health monitoring with our fleet management software Cortex, for instance, can generate optimal shop visit scenarios with additional factors such as parts market dynamics, cost structures, utilization and operational conditions, among others,” Bickel says.
Like the OEMs, MTU reports the quality of its health monitoring is constantly improving thanks to new hardware feeding more data points and new software taking advantage of machine learning and AI.
“Modern software nowadays allows for the analysis of large data sets and with better performance than we have known it in the past,” Bickel says. “We are seeing progress in the aviation industry with respect to sensors, the acquisition of data, additional hardware or means for recording data, the infrastructure for data transmission and the ground-based hardware that relays the collected information.
“The key to improving [engine trend monitoring] methods and engine health diagnoses is to look at a multitude of parameters and their interactions rather than a single or a limited number of them,” he continues. “With more and improved data, we can also apply and derive more complex correlations between different parameters, which leads to better diagnosis.”
Of course, engine manufacturers have the advantage of being the primary sources and collectors of health monitoring information. However, Bickel sees such services as complementary rather than in competition.
“Our advantage is that we offer our health monitoring services to a wide range of engine programs so that customers with broader engine portfolios can benefit from a one-stop solution,” he says. “In some cases, customers also use multiple platforms to gain insight into the health of their engines or engine fleet.”
FUTURE DATA
While engine trend monitoring customers can expect modeling and predictive tools to become increasingly accurate as more data is collected and machine learning algorithms improve, they may also see health monitoring services extend to other parts of the aircraft.
Airbus is seeking to mature new technologies that can record deformation and stress. The next generation of aircraft may incorporate structural sensors to warn of impending damage or weakness in wing and fuselage structures. Airframers already use a variety of such sensors in testing, and research has proven the viability of embedding strain gauges into materials such as carbon-fiber.
Meanwhile, Bickel at MTU still sees room for more data acquisition within the newest in-service generation of aircraft and engines. “Using sensors does not mean the data is by default also recorded,” he says. “However, there is an increasing understanding that data acquisition and recording is important to gain a better picture of engine health. That in turn allows for a more targeted interpretation in case of unusual behavior, which is why MTU uses all applicable data that it can get for engine health analyses.”