Publication Details
Issue: Vol 7, No 1 (2025)
Pages: 283-291
ISSN: 2660-4159

Abstract

This study evaluates the effectiveness of advanced ultrasonic phased array imaging systems for non-destructive evaluation (NDE) of carbon fiber reinforced polymer (CFRP) composites in aerospace structures. A hybrid method combining time-of-flight diffraction (TOFD) with convolutional neural network (CNN) image processing was developed and validated, demonstrating a 94.7% defect detection accuracy for delamination, porosity, and impact damage compared to 78.3% for traditional methods, with a 40% reduction in false positives. Using a 64-element phased array transducer at 5 MHz and a multi-modal data acquisition system, significant  improvements in signal-to-noise ratio (SNR) and lateral resolution were achieved, indicating the method's potential for in-service inspection of complex aircraft components.

Keywords
Ultrasonic Phased Array Non-Destructive Testing Composite Materials Machine Learning Convolutional Neural Networks Aerospace Structures