Detection of near-surface defects in rails combining Green's function retrieval of ultrasonic diffuse fields and sign coherence factor imaging

Abstract

A method combining Green's function retrieval theory and sign coherence factor (SCF) imaging is presented to detect near-surface defects in rails. The defects are close to the ultrasonic phased array and near-surface acoustic information of defects is obscured by the non-linear effects of the initial wave signal in directly acquired responses. To overcome this problem, cross-correlations of the diffuse field signals captured by the array transducer are performed to reconstruct the Green's function. SCF imaging is used to further improve the spatial resolution and signal-to-noise ratio (SNR) of near-surface defects in rails. Experiments are conducted on two rails containing two and four defects, respectively. The results show that these defects can be clearly identified when using the reconstructed Green's function. However, the images of near-surface defects are masked and cannot be distinguished when using directly captured signals and total focus imaging. The proposed method reduces the background noise and allows for effective imaging of near-surface defects in rails.