Whose Plane to Take? A Novel Approach to Identify the Optimal Model for Defect Detection in Composite Aerostructure Material

Abstract Carbon-fiber reinforced polymers are a common material used to manufacture aerospace structures due to their good tensile strength and resistance to compression. During the manufacturing process, these composites are thoroughly inspected for flaws and defects to ensure structural integrity during commercial use. Recently, researchers developed DOC-DICAM, an AI-based assistance tool to speed up the identification process by training models using archived fuselages. We extend their method by identifying the optimal model to use for inference based on a collection of models each trained on different fuselages over time. A naïve approach is to use the model trained on the most recent fuselage minimizing concept drift. However, we show that this may not always give the best results suggesting the need for a recommendation step in the workflow. Due to the high volume and velocity of fuselage data, archived data is routinely moved from disk to tape storage making the comparison between new and old data impossible. Therefore, we propose a method to identify the optimal model using only the collection of models and the new data. We validate our approach by showing the rankings produced by our method are consistent to those produced using conventional methods when archived data is available. Finally, we show that our method identifies the optimal model for identifying defects in composites generated in a production environment.