Vibration-based damage identification in beam structures through wavelet analysis

Nowadays the topic of damage identification in structures is of primary interest in the field of civil, aerospace, mechanical engineering research. Indeed, due to the increasing use of advanced structural systems (e.g. airplanes, space shuttles, wind turbines, etc.) and to the aging of old structures (e.g. buildings, towers, bridges, etc.), the safety and reliability of structures have to be guaranteed to avoid catastrophic events and loss of human lives. The present thesis is focused on vibration-based damage identification in beam structures through wavelet analysis. The dissertation is arranged in six chapters. Chapter 1 introduces the topic of the thesis through a broad presentation of the state of the art of damage identification methods for structural health monitoring and control, with particular attention to vibration-based structural damage identification methods. In Chapter 2, the time-frequency technique, named wavelet analysis, is firstly theoretically presented and its application, available in the literature particularly for beam-like structures, as a damage detection tool both in time and in space domains is discussed. In Chapter 3, the mechanical models of homogeneous and fiber-reinforced cracked beams are presented. The models are used to simulate the real static and dynamic responses of beam structures for successive damage detection through wavelet analysis. The last three chapters of the thesis are devoted to the original findings of the present research. Chapter 4 focuses on the issue of border distortions in damage detection by continuous wavelet transform. To tackle the problem, a new polynomial padding method is proposed and compared with the most effective padding methods commonly used in the literature. In Chapter 5 the effect of spatial sampling in damage detection of cracked beams by continuous wavelet transform is thoroughly investigated through a parametric study. From the outcomes, some general indications on the optimal number of sampling intervals for an effective damage detection are obtained. Finally in Chapter 6, a new health structural monitoring method based on time-spatial wavelet analysis is presented to control the static and dynamic, elastic-plastic behaviour of a cracked fiber-reinforced beam. The capability of the method is discussed particularly with respect to scale of the analysing wavelet, the noise level and the spatial sampling interval, considering a small crack.