Frequency and Time Domain Analysis of Guided Waves for Damage Detection

Abstract

Guided waves have been shown to be a very useful approach for damage detection. Their interaction with inhomogeneities, holes, cracks, delaminations or other defects present in the structure carries important information and can reveal the type and extent of the damage itself. Furthermore, depending on the nature and size of the damage to be monitored, an appropriate mode and wavelength can be used. Explicit algorithms are used to numerically simulate these interactions giving accurate predictions of what is obtained experimentally. However, the selection of the proper mode and wavelength to be used is strongly dependent on the nature and dimension of the defect and it often requires quite long computational times, as well as a tuning process that can identify the appropriate characteristic of the ultrasonic wave for each defect. Structures of a certain complexity and extent may contain defects of various nature and size and therefore a significant number of analyses may be required: depending on the specific problem, it is necessary to use more than one excitation pulse at difference frequencies and the time-domain analysis needs to be carried out more than once with a further increase of computational costs. Conversely, guided waves in the frequency domain can be used for the same class of problems and may offer some advantages in terms of computational time saving, especially in the case of a multi- damage problem for which a multi-frequency excitation is required. In this paper the numerical solution in terms of frequency response functions (FRF) of the guided waves problem is demonstrated to be an alternative and efficient approach to the time domain analysis. Moreover, the time domain solution can be reconstructed from the FRF data for any input frequency and waveform. The time domain solution reconstructed from the FRF data are compared to solution directly obtained in the time domain. Subsequently a damaged structure will be considered to demonstrate again the efficiency of the FRF approaches. Finally, the advantages and limitations of time and frequency approaches are discussed.