Sparse sensing damage detection procedure for composite plates

Structural health monitoring (SHM) procedures for aerospace structures require a global interaction among actuation, sensing, data feature extraction and data feature classication for damage pattern recognition purposes. The objective of this thesis is to present an integrate SHM procedure for damage pattern recognition of composite plates through the changes in the structural dynamics response when a reduced number of piezo-electric sensors is considered. In this context, the term sparse is used to address the proposed procedure which considers one device only for actuation and sensing purposes, respectively. Both numerical and experimental test cases are considered in order to verify and then validate the developed routine for a wide range of damage shapes and severities. In particular, the reduction of Young modulus and thickness of an aluminum isotropic plate is rstly considered in the initial numerical assessment. Then, the delamination in carbon-bre-reinforced-plastic (CFRP) laminate plates subjected to low-velocity impacts (LVI) is investigated as the experimental test case. The methodology here presented focuses on vibration-based damage detection of basic aerospace structural components in order to establish an eective and accurate baseline condition aiming to be the starting point for complex shape and large-scale damaged specimen identication. The time-domain responses of the dierent damaged congurations are analysed by means of two approaches, namely the wavelet transform (WT) and the auto-regressive (AR) models in order to extract suitable damage sensitive features. A proper choice of the extracted features driven by statistical techniques has to be performed in order to enhance the ability of the damage pattern classication. The achieved results make this proposed approach a promising solution for online health monitoring system development for smart aerospace structures when a reduced number of sensors is available. The performed work is organized in six main chapters. In chapter 1 the main denitions, concepts, and results of vibration-based SHM techniques are introduced. A special emphasis is dedicated to wavelet analysis and auto-regressive models techniques. Lastly, the low-velocity impact-induced damage and delamination in composite laminate plates are brie y reported. In chapter 2 an essential overview of the WT and AR models theories associated with data analysis for SHM purposes are reported in order to highlight the potential of such approaches compared to other mathematical tools for damage sensitive feature extraction, such as Fourier Transform (FT). Chapter 3 provides some elements of statistical analysis for pattern recognition purposes. In chapter 4 the proposed SHM routine is described. In chapter 5 a numerical assessment of the SHM strategy is presented. Chapter 6 is devoted to the description of the developed experimental test campaign on CFRP laminate plates subjected to low-velocity impact. Finally, the concluding remarks of the present work are summarized.