Digital-Twins of composite aerostructures towards Structural Health Monitoring

Abstract

The present work is primarily dedicated to the development and validation of a Digital Twin representative of an aeronautical structure. A finite element model is thus created to simulate the structural response of a composite single-stringer panel under compressive loading. Quasi-static tests are conducted and health monitoring techniques, i.e. displacement and strain measurements via digital image correlation and fiber Bragg grating sensors respectively, are utilized. The mechanical response of the model presents very good agreement with the experimental evidence. The validated model is then exploited, augmented by exogenous details, i.e. external loading, to train a surrogate mathematical model. The resultant surrogate maps efficiently the Digital Twin strains with respect to the load. This advantage is envisaged by the Digital-Twin concept where realtime data flowing from the physical twin would be used to detect the presence of skin-to-stringer disbonds. The proposed methodology is tested for the case of an artificially disbonded panel, subjected to block loading compression-compression fatigue, utilizing static strains periodically acquired during several quasi-static test intervals.