Recursive application of cross-correlation for extraction of sparse impulse trains from ultrasonic nondestructive evaluation data

Abstract

The goal of Quantitative Nondestructive Evaluation is to determine numerical values for a specimen’s properties. In many cases, especially when the specimen is composed of thin layers, the material properties manifest themselves in a sparse train of impulses. This is the case for many state of the art hybrid materials. Signal processing techniques can be used to extract the desired impulse train from the ultrasonic data. In this paper, a cancellation method is used to recursively subtract a reference wavelet from the acoustic response of a test specimen in order to determine the desired impulse train of the specimen. The method was applied to acoustic data for a thin aluminum sheet. The resulting impulse train displayed enhanced fidelity as compared to the original wave form. Wave travel times and transmission coefficients thus obtained agree with expected values to within 6%.