Validation of hybrid SAFE-FE guided wave scattering predictions in rail

Abstract

Guided waves are well suited for structural interrogation of continuously welded rail track, and are being exploited in a commercial monitoring system. This system relies on a pitch-catch mode of operation to reliably detect complete breaks in the rail. Efforts to extend this system to include a pulse-echo capability (for detecting damage before complete breaks occur) require numerical modelling to predict how propagating waves interact with damage. To this end, an efficient hybrid SAFE-FE numerical method, capable of predicting reflections from arbitrary discontinuities, has previously been implemented. In this paper, we attempt to validate the SAFE-FE method by comparison with both numerical and experimental results. A numerical comparison between the SAFE-FE method and a time domain analysis using the commercial finite element package ABAQUS Explicit is performed for two example problems. The ABAQUS Explicit time domain results are processed so that a comparison of power reflection coefficients is performed. In both instances, good agreement between the hybrid method and ABAQUS Explicit were observed. The experimental comparison involved the measurement of reflections from passive reflectors bonded to a rail in the field. These measurements were used to estimate the power transmission and reflection coefficients of the reflectors. A reasonable correlation was achieved for the power reflection coefficients, but the hybrid method was shown to somewhat over-predict the power transmission coefficients.