Upscaling the Data-driven Prognostic Methodologies Towards a Condition-based Structural Health Management of Composite Structures

摘要

We present a conceptual methodology built in the framework of the ongoing EU H2020 Real-time Condition-based Maintenance for Adaptive Aircraft Maintenance Planning project, that leverages structural health monitoring data from different sensing technologies that goes a step beyond damage detection and diagnosis towards the probabilistic remaining useful life estimation in the presence of adverse conditions during flight. The methodology relies in several parallel activities from damage detection and localization to damage identification and severity assessment, sensitive-to-damage feature extraction processes, training methodologies, data fusion and remaining useful life predictions. Various sensing technologies i.e. static and dynamic strain sensing with FBGs, guided waves and acoustic emission are employed. An extensive hierarchical test campaign on test articles of increased complexity based on the building block approach is discussed with details as to the types of damage that are going to be targeted. Single and multi-stringer composite stiffened panels are subjected to realistic loading conditions. Emphasis on impact damage and skin/stringer fatigue disbonding/delamination is given. Last but not least, sophisticated mathematical algorithms are proposed e.g., multi-state degradation models such as Non-Homogeneous Hidden Semi Markov Model in order to deal with the data-driven RUL prediction with uncertainty quantification.