Analysis of ultrasonic non destructive evaluation data using singular value decomposition of the Hankel data matrix

A method is presented for processing data from ultrasonic non-destructive evaluation (NDE) tests of material specimens by performing a singular value decomposition of the Hankel data matrix. The singular vectors and singular values of the Hankel data matrix thus obtained correspond roughly to modal shapes and modal amplitudes of the impulse response data. A great deal of information about the specimen being tested can thus be conveyed with relatively little data; one approach which has been explored is to generate a color map in which the intensity of the RGB components are determined by selected singular values. This approach was applied to simulated data for both nominal and flawed specimens and the resulting color map has been compared to a gray scale based on the RMS value of the response data. The color map provides more information about the flaw than the gray scale in two ways: the intensity of the color components increases monotonically with increasing flaw severity, and the color map provides information about the flaw location.