Aircraft structural health monitoring by fusing sensor data

Abstract

Aircraft structural maintenance is often done following a fixed-time interval or failure-based approach. However, this can be costly and time-consuming. A smarter way can be to move towards condition-based maintenance in which maintenance is done based on the actual condition of a structure: if there are indications that the structure is damaged, maintenance is performed. In my work, I develop AI-based diagnostic methods that can be used for such purpose. For this, I fuse sensor data obtained using acoustic emission sensors and distributed fiber optic strain sensors. Each sensing technique has its own capacities and monitors different damage. By fusing their data, a holistic image of the damage in a structure can be obtained on all four structural health monitoring levels: 1)detection, 2) localization, 3) identification, and 4) severity.