A Novel FRF-Based Damage Localisation Method Using Random Vibration

Abstract

This paper presents a novel damage localization method based on the measured Frequency Response Functions (FRFs) without demanding any previous data records of the structure in its healthy state. The main innovation of this study starts with reconstruction of FRFs curvature to develop spatial shape functions. It is demonstrated that reconstructed data significantly magnifies the influence of low-frequency spectra in damage detection procedure which is considered the milestone of this approach as excitation of the higher frequencies is not easy to obtain in most practical applications. The modified curvature data in all measured frequencies and locations is interpreted as a two dimensional image and then processed by employing 2-D discrete wavelet transform to detect any abrupt variation at damage site. Level one wavelet decomposition is utilised to provide the finest detail coefficients. It is illustrated that this approach presents a more recognizable pattern at damage site in all measured frequencies. The pattern can be described by a horizontal line parallel to the frequency spectra in 2-D image. Hence, the horizontal detail coefficients are utilised to detect this pattern as they are more sensitive to perturbation with orientation parallel to horizontal axis in the image. The main contribution of this approach lies in the fact that the proposed technique is able to detect the structural damage in all measured frequencies and the effectiveness of the method is independent of the excitation location. Moreover, the results provide a better visualisation at damage site which other FRF-based damage detection methods could not obtain. Applying broadband FRF data in this approach and the fact that there is no need for data from the healthy state of the structure are other advantages accompanying this method. The robustness of the proposed damage identification method was examined with various damage conditions in both single and multiple states. Moreover, the feasibility of the method was verified in presence of practical uncertainties such as noise using extensive numerical simulations. It was demonstrated that the proposed method is particularly attractive for practical applications as it opens an opportunity for online monitoring of the structural integrity without demanding any previous data records of the structure.