Autonomous impact damage detection and isolation prediction for aerospace structures

This paper presents a practical yet innovative impact damage identification and prognosis approach for aerospace structures that uses an optimized suite of reliable COTS sensors coupled with advanced damage detection and modeling algorithms. The presented methodology utilizes a monitoring approach based on acceleration measurements that are analyzed using advanced signal processing and dispersive wave theory models that capture frequency and orientation dependent wave propagation effects. The acceleration measurements and associated processing modules are used to provide immediate detection and isolation estimates, while an energy amplitude feature allows for assessments of damage severity after the impact. By embedding wave theory model results with the adaptive signal processing algorithms, a more accurate understanding of the time-frequency behavior of the dispersive waves produced at impact is gained. Damage localization is performed based on the comparison between the predicted and measured wave group velocities, with a genetic algorithm used to optimize the parameters of a triangulation procedure. This combination of model and feature-based algorithms allows the system to make use of limited, but readily available accelerometer data. This procedure also minimizes the learning and modeling difficulties associated with other techniques that are based solely on models or measurements. A few selected demonstrations are presented that illustrate the impact location prediction capabilities in realistic carbon fiber reinforced polymer (CFRP) composite panels