All Nippon Airways Uses MATLAB to Predict Aircraft Component Failures

“Thanks to MATLAB Compiler, we have been able to summarize the entire analysis process into a single executable file. We monitor flight data on a daily basis, and, as a result, we have successfully predicted several journal bearing failures about one month prior to their occurrence.”

Key Outcomes

  • Extracted insights for predictive maintenance by combining operational data and domain knowledge in MATLAB
  • Component failures predicted one month in advance with machine learning models trained using MATLAB
  • Several cabin air compressor journal bearing degradations identified before failure from real-time sensor data

All Nippon Airways Co., Ltd. (ANA) started as an air transport business in 1952. Today, it is a full-service carrier headquartered in Tokyo operating flights within Japan and internationally to Asia, Europe, and the United States.

ANA’s Maintenance Center keeps the fleet flying on schedule. Airlines usually rely on two types of maintenance: preventive and corrective. Corrective maintenance is performed in response to a component failure, which can result in downtime, delays, or cancellations.

Over the past decade, ANA has launched a predictive maintenance strategy to help reduce aircraft downtime and increase operational efficiency. This strategy leverages ANA’s operational data and domain knowledge to help anticipate component failures before they occur.

ANA uses MATLAB® for visualization and hypothesis testing. The team also uses MATLAB to gain insights from sensor data, such as identifying the root cause component and early signs of failure. The insights are used to train machine learning models, which are deployed to a data pipeline using MATLAB Compiler™ to evaluate real-time sensor data from safety-critical systems. The new predictive maintenance strategy has enabled ANA to find several cabin air compressor journal bearing degradations before failure.