Classification of ultrasonic echo signals to detect embedded defects in carbon fibre reinforced plastic laminates

Autor: Anish Poudel, Shanglei Li, Raghuveer Kanneganti, Lalit Gupta, Tsuchin Philip Chu
Rok vydání: 2015
Předmět:
Zdroj: International Journal of Microstructure and Materials Properties. 10:216
ISSN: 1741-8429
1741-8410
DOI: 10.1504/ijmmp.2015.072917
Popis: Ultrasonic non-destructive evaluation (NDE) technique is a commonly applied tool for the inspection of composite structures. But, it is very difficult to discern ultrasonic echo signals arising from different flaws in composites, thus making the damage characterisation process very difficult. In this paper, a parametric classification approach based on nearest mean is developed to automatically characterise defective ultrasonic echo signals. The study employed three different composite laminate test standards with the following simulated defects: impact damage, foreign object inclusion, and porosity. To test the classification strategy, 100 time-domain 5 MHz ultrasonic echo signals were acquired for each type of defect and from non-defect areas on composite laminates. The signals were start-point aligned using cross-correlation to improve the estimation of the means and to improve the classification accuracy during testing. Through extensive testing it is demonstrated that classification accuracies exceeding 98% are possible using this classification approach.
Databáze: OpenAIRE
načítá se...