- Case study
Highly refined monitoring fuels growth for a flight engine maker
Capturing and analyzing critical data mid-air is changing the game
An aircraft manufacturer.
To make sense of the millions of records its 38,000 aircraft engines produced as they flew across the globe every day as a way to prevent events and keep passengers in the sky.
Genpact and its engineers helped an aircraft engine manufacturer make sense of huge amounts of flight data to solve maintenance and safety issues before they happen. It’s keeping passengers safe as they fly, and saving airlines money too.
Challenge
Managing millions of records from 38,000 aircraft engines flying across the globe —and deriving useful information from them — was the biggest challenge. Our client was struggling to cope in real time with the huge amounts of data that different aircraft sensors transmitted. External conditions made it even more difficult to keep the scattered data organized during different flight phases and then remotely monitor and diagnose aircraft engines for possible failures. That led to major downtime, causing delays and flight cancellations — and significant losses for our client.
Solution
Genpact jumped on board, and helped the company revolutionize its maintenance systems from nose to tail. Our unique approach landed that perfect mix of domain knowledge, digital technology, design thinking, and smart analytics. All this helped the company better harness its sensors and system. Here’s how it played out:
Condition-based remote monitoring and diagnostics
Genpact receives 4 million records a month from a fleet of more than 40,000 engines flying across the globe. We get these records through Aircraft Communications Addressing and Reporting System (ACARS) data transmission models from more than 30 sensors fitted around different components of each engine. Armed with that data, a team of more than 30 engineers in our remote monitoring and diagnostics (RM&D) centers in India and the US analyzes over 190,000 operational alarms per year. That gives customers greater confidence in their asset reliability and performance 24/7/365.
Using a combination of off-the-shelf and custom analytic tools, the team diagnoses problems ranging from fuel consumption to engine blade damage. Drawing on the experience of hundreds of thousands of resolved alarms, the team has developed physics-based algorithms that provide early warning of more than 120 different failure mechanisms.
We combine our deep understanding of the physical characteristics of aircraft with the intelligent application of statistical methods. In that way, we continually improve relevant algorithms to increase the likelihood of finding problems while reducing false alarms. This depth of expertise, as well as operational excellence in executing downstream processes, has resulted in a 50% reduction of false alerts.
Sense and acquire
More than 30 sensors positioned strategically on the engine record key technical parameters several times during each flight phase — take-off, climb, cruise, and descent and landing. These sensors read engine characteristics such as temperatures, pressures, speeds, flows, and vibration. An engine condition monitoring system (ECM) acquires this information in the form of snapshots that indicate unusual engine condition reports and give flight summaries.
Transfer
A critical aspect of the ECM system is the transfer of data from aircraft to an analytics engine. ACARS digital data-link systems are the primary means of communication. They transmit Aircraft Condition Monitoring System (ACMS) reports via a satellite link while the aircraft is in flight (Figure 1).
Figure 1: ACARS digital data-link system
Analyze
The analytics engine then validates the data and applies corrections to normalize the information. The snapshot data is always followed closely, so teams can detect subtle changes in condition from one flight to another. Automated algorithms based on neural networks do this, and information brought together from multiple sensors provides extremely sensitive detection capabilities.
Act
The data signature will typically highlight a change in an engine characteristic. Skilled Genpact engineers then turn this symptom into a diagnosis — and, usually, a prognosis. They assess the most likely physical cause of a data signature, how an operator can confirm this, and how quickly to act on the information.
Impact
Our client experienced the following benefits as a result of this work: