Smoother flights when AI runs the show
Photo Paul Vincent via Unsplash
Korean Air’s fleet of 139 passenger planes carries millions of passengers across the globe every year. Ensuring each plane is safe and well-maintained is a priority. When Korean Air addressed line maintenance issues with its fleet, its team of 2,000-plus maintenance employees has historically had to pore over troves of maintenance records to find crucial data on everything from how to fix an important plane part to a plane’s maintenance history.
“Maintenance issues represent a substantial cost to an airline—to the tune of 28 percent of the total operating cost,” said Rob Ranieri, Vice President and Partner, Global Industry Offering Leader, Travel and Transportation at IBM.
To remedy this tedious and costly practice, Korean Air recently enlisted IBM Watson® Explorer to compile and analyze data from various sources, including technician notes, material cost data, and in-flight incident history. The massive amounts of data generated by an airplane itself is beyond the ability of a ground crew to absorb and act on, but for AI, it’s easy. AI offered actionable insights to help maintenance staff act faster and more thoroughly. In doing so, Korean Air shortened its maintenance defect history analysis lead times by 90 percent.
More than half of all maintenance issues are unscheduled or unplanned, which places an added urgency on airline employees and maintenance crews to address them. Often these delays—like refueling, airplane maintenance and crew rescheduling—are within the airline’s control. In April 2018, nearly 5 percent of air travel delays were caused by the air carrier, per the Air Travel Consumer Report by the U.S. Department of Transportation’s Office of Aviation Enforcement and Proceedings (OAEP).
In the short time frame when an airplane is on the ground, technicians often have to rush to access hard copies of maintenance manuals and make necessary adjustments to the problem areas. “There are terabytes of data being generated by all maintenance actions. The question is: How do I get this data into the hands of the mechanic or to the supervisors who are scheduling maintenance actions?” said Ranieri. Airlines are now optimizing airplane maintenance activity by using technology like IoT and AI to centralize and analyze all critical data points of an airplane—like maintenance history and in-flight history—to help reduce overall schedule disruptions and ease customer service issues.
For the team of engineers and their supervisors who are tasked with improving the efficiency of airplane maintenance, having the right information at their disposal is imperative.
When paper work accounts for 40% of engineering work
In the case of Japan Airlines (JAL), the carrier’s engineering procedures and processes were outdated. Airplane manuals and engineering records were paper-based. Technicians and engineers used radio for communication. To check faulty airplane parts, technicians would have to print out manuals from a PC in the office and bring it with them to the hangar. All told, the red tape and the paperwork accounted for 40 percent of the engineering work.
JAL collaborated with IBM and Apple to develop IBM MobileFirst for iOS apps for airplane mechanics and supervisors to help airplane mechanics safely maintain and deploy airplanes. Armed with iPads and iPhones, airplane engineers have access to the tools, processes, and data they need to be more efficient and spend more hands-on time with airplanes—enabling them to concentrate on the most important tasks.
Reducing delays and service issues
Finnair’s mission is to offer the smoothest and fastest connections in the Northern Hemisphere. In 2017, Finnair flew nearly 12 million passengers. With the goal of flying 20 million passengers by 2030, its entire fleet of airplanes has to run with few to no delays or service issues.
In 2016, the carrier partnered with Apple and IBM to adopt multiple enterprise iOS apps to accelerate Finnair’s overall digital transformation. As line maintenance operations become more digitized, carriers can ultimately help extend asset life by spotting warning signs sooner and by incorporating more sophisticated and predictive data. It will ultimately improve airplane performance and customer satisfaction, said Ranieri.
“In the next five years, as the predictive capability of artificial intelligence becomes more sophisticated and the vast amount of data is much more automated, we should see a significant reduction of unplanned maintenance from 60 percent to down to 10 percent or below.”