AUTOMATED ULTRASONIC NDE OF COMPOSITE SPECIMENS BY CLUSTER AND FRACTAL ANALYSIS

Abstract Non destructive evaluation (NDE) of the damaged states in composite materials is far more complicated than in case of conventional homogenous and isotropic materials. As composites may contain several types of defects (existing either individually or in combination), analysis of commonly used time-domain features, such as peak amplitude, or time of flight, alone seems to be insufficient for identification of defects. In the presenl investigation, automated ultrasonic C-scan procedure, developed in-house, is undenaken on artificial inclusion implanted and impacted glass epoxy composite laminates. Features of ultrasonic waveform are extracted from time and frequency domain. Cluster analysis is employed to obtain C-scan images of specimens by grouping the data-sets in a systematic manner. Fractal analysis is performed on these C-scan images to characlerize them in a quantitative way. Fractal parameters of these images are evaluated from their respective fractal graphs. The results provide information about the utility of frequency domain features in NDE of composites and fractal theory is observed as a viable supplement, if not an alternative, lo human inspection in evaluation of the C-scan images.