AI ushers in a new era for Japan Airlines: Predicting Engine Blade Failures
Japan Airlines (JAL) has embarked to develop inspection tools utilizing artificial intelligence (AI) and image recognition technology to examine turbine blades. Successful implementation could lead to enhanced failure prediction capabilities.
Aircraft jet engines feature "turbine blades," delicate feather-shaped components that play a crucial role. These engines comprise hundreds of thin blades aligned in rows. They harness the energy of high-temperature, high-pressure gases resulting from burnt jet fuel to rotate the engine's shaft, generating propulsion. This section of the engine experiences the highest stress.
In contemporary blade inspections, mechanics commonly employ industrial endoscopes to peer inside. However, evaluating the risk of failure for each of the numerous blades is complex, necessitating years of experience, intuition, and expertise. Addressing this, efforts are underway to distribute the mechanics' workload evenly and facilitate the transfer of skills to prepare for future labor shortages.
25 May 2023 - JAL and CRESCO are fostering collaboration through the "JAL Innovation Lab," JAL's hub for open innovation. By leveraging CRESCO's AI-based image recognition technology and machine learning expertise from the field of ophthalmology, the initiative aims to refine endoscope inspections, predict blade failures, automatically recognize damage, and facilitate skill transfer.
Achieving this objective demands the accumulation of extensive real-world machine data. Initially, 16 European Airbus A350-900 aircraft, renowned for their advancement, will undergo inspection. Endoscopes will be used to observe the engine interiors, capturing image data of the blades. These inspection records will be chronologically stored as data (JAL Engineering), supplemented by engine data collected during regular operations. Patterns of scratches indicative of potential failures and malfunctions will be learned.
The A350 serves as JAL's flagship aircraft, with plans to introduce up to 56 planes (including 25 options) for international flights, totaling 1,000 aircraft. As these aircraft are slated for long-term service, the development of inspection tools promises sustained effectiveness. Moreover, newer aircraft can systematically track deterioration trends (JAL Engineering).
Practical application of these inspection tools is anticipated post-2025. Once failure prediction technology is established, it could enable "predictive maintenance," proactively maintaining blades. This development has the potential to significantly enhance air safety.
