Abstract:

This paper describes the ultrasonic method to detect the inner flaws of three-dimensional braided composite materials. After the wavelet packet transform of three-dimensional braided composite material, a scan echo of ultrasonic, the material’s inner defection signal is expressed as energy characteristics after it is extracted and decomposed. Input the feature values of the defections into BP neural network which realizes classification and automated identification for micro-cracks and pores of three-dimensional braided composite materials. The experimental results proved that the method is feasible for dectecting three-dimensional braided composite materials’ inner flaws.