Methodologies for the Damage Detection Based on Fiber-Optic Sensors. Applications to the Fuselage Panel and Lower Wing Panel

Abstract

Damages in aircrafts start as local cracks or delaminations, which, even though they do not change strongly the overall strain field, form initiation points for an eventual failure. Fiber-optic sensors act primarily as strain sensors, so unless damage happens to be close to the sensor location, the changes in the strain data may be very slight and damage may go undetected. Three main independent algorithms were developed for damage detection from strain measurements: (1) a strain field pattern recognition technique, based on principal component analysis, from which damage indices were defined. The validity and accuracy of the approach was proven by experimental tests, as well as by simulations. (2) the extension of the Modal Strain Energy Damage Index algorithm, as a suitable method to identify skin–stringer debonds and optionally larger impact-induced damages in a CFRP composite structure with stiffeners, by using the strain data obtained during vibration tests from wisely distributed fiber Bragg grating (FBG) sensors; and (3) a strain difference-based algorithm which is threefold consisting of the following parallel methodologies: (i) strain difference limit method, (ii) Mahalanobis distance-based method, and (iii) statistical hypothesis testing-based method. These developments were put together on two realistic full-size aircraft structures, to validate the methodologies. In addition to the discretely located FBGs, distributed fiber-optic sensing was also used in the final testing, accurately measuring the strain signatures of damages caused by applied impacts.