Energy evaluation and passive damage detection for structural health monitoring in aerospace structures using machine learning models

Abstract

Structural Health Monitoring (SHM) in aerospace engineering is more and more based on the use of Artificial Intelligence. In this manuscript machine learning algorithms were trained to identify and to characterize the structural effects of impacts on a typical aerospace aluminum panel. A significant experimental campaign was conducted to create suitable impact datasets (the vibrational behavior of the reinforced plate, acquired by piezo sensors). Shallow neural networks, properly trained, were applied to determine critical events affecting the operational conditions. The focus of the manuscript was double: on the severity of the event (a regression problem regarding impact energy) and on the detection of preexisting damage to monitored areas (a classification problem regarding the identification of damaged zones). The scope of this work was to demonstrate the validity of the machine learning approach as an SHM tool for impact effect characterization in a realistic aerospace structure (i.e., energy prediction with a percentage error never more than 10% and identification of previous damaged zones with an accuracy of more than 95%) and to demonstrate its computational efficiency despite the test complexity, provided that the selection of features is guided by a meaningful physical and mechanical interpretation of the underlying phenomena.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata

Altmetrics