Damage identification for beam-like structures based on proper orthogonal modes of guided wavefields

Abstract

Damage in structures can cause local anomalies in a guided wavefield due to reflection of guided waves in neighborhoods of damage. These local anomalies can be used for baseline-free damage identification if the structures are geometrically smooth and made of materials that have no stiffness and mass discontinuities. Recently, guided wavefield-based methods have been studied for damage identification by extracting and localizing local anomalies in guided wavefields in the time- and frequency–wavenumber domains. Meanwhile, proper orthogonal modes (POMs) obtained by the proper orthogonal decomposition have been studied for vibration-based damage identification. In this paper, the effectiveness of POMs of guided wavefields for damage identification in beam-like structures is studied. Since local anomalies in the POMs can be covered by global trends of the POMs, the continuous wavelet transform is used to suppress the global trends and intensify the local anomalies. The fundamental mechanism of how the continuous wavelet transform with Gaussian wavelet functions of a proper order can suppress the global trends of POMs and intensify local anomalies of POMs is explained. Significant POMs used for damage identification are determined by an adaptive truncation technique. The proper orders of the Gaussian wavelet functions, i.e., their number of vanishing moments, are determined based on the modal assurance criterion and a statistical criterion. The continuous wavelet transform of the significant POMs with Gaussian wavelet functions of the proper orders is used to yield an accumulative damage index. Numerical and experimental investigations of the proposed method are conducted on damaged beam-like structures. Their results verified that the proposed method is accurate and noise-robust for identifying the location and extent of damage in beam-like structures.