Built on cognitive data discovery technology, Korean Air’s maintenance defect analysis system offers a superhuman ability to rapidly synthesize vast amounts of historical data and find subtle clues on how to solve complex mechanical problems. The IBM Korea Lab team provided consulting and training services, while a project manager from IBM® Global Technology Services® – Infrastructure Services managed the rollout of the hardware systems required to support the software. Ingesting and analyzing structured and unstructured data from the airline’s ERP environment, the solution detects patterns in how and why equipment fails and recommends courses of action to help technicians implement the most effective fix. The system uses natural language processing (NLP) and advanced content analytics to transform raw unstructured data, mainly in the form of technician notes, into insights focused on the problem at hand. For instance, Korean Air can compare the performance of different flights, types of aircraft and maintenance crews to isolate mechanical, environmental and human variables that may affect equipment, such as the increased need for tire replacements in the summertime, given higher temperatures and passenger volumes, and potential solutions for reducing wear and avoiding failures.
Korean Air shortened its maintenance lead times by 90 percent by dramatically simplifying how aircraft technicians search and analyze data to determine root causes and solutions. The airline also reduced the number of flight delays and cancellations caused by maintenance issues, which made a positive impact on customer satisfaction and operating efficiency. In addition, the cognitive solution alleviates much of the burden on technicians to know exactly what to look for as they seek answers to complex problems, and reduces the time they spend searching, exporting and analyzing the data.
Aircraft technicians are trained to understand complex systems and notice patterns of equipment failure, but human capacity is limited, especially when it comes to using big data to understand rare or novel situations. Korean Air is supplementing human intuition with a cognitive computing solution that can ingest larger, more diverse data sets, detect more subtle connections and reach conclusions faster. This changes the equation for airlines, allowing them to spend less time on maintenance and more time in the air.
In the past, aircraft technicians relied heavily on experience to solve maintenance challenges, having to know what to look for and how to analyze it for diagnostic purposes. Now, the cognitive search and discovery solution does the work for them, sifting through years of unstructured historical records to find connections that illuminate root causes and point to the right solution.
The solution creates a system of cognitive discovery and decision making by extracting insights that can be applied to the puzzle at hand, helping aircraft technicians understand the problem faster, make better decisions about how to fix it and keep flights running smoothly.