Skin–stringer debonding detection using distributed dispersion index features

Abstract

Statistical approaches have been extensively used to detect structural damage. This article proposes a preliminary investigation on the use of a non-model-based damage identification method, implementing “dispersion feature” indicators, extracted from the differential strain signal. In detail, the method takes advantage of the elaboration of the full-spectrum acquisition, irrespective of the healthy state of the structure under test. Large variations of some selected statistical features, representative of the acquired signals, are taken as indicators of the presence of anomalous structural condition. At this stage, these features have been arbitrarily selected. In synthesis, a cumulative dispersion index is defined as the combination of the selected feature variations, contemporarily overcoming prescribed thresholds, providing a scalar output that can be used as an index to identify and possibly quantify the extension of a damage location. Experimental results confirm the envisaged potentiality of the proposed technique. In this study, damage in the form of a stringer debonding is referred to, obtained as a consequence of point impacts on a wing box stiffened panel. A distributed fiber optic strain sensor network is used to detect the input data.