Abstract: Ultrasonic phased array inspection system is used to test carbon fiber composite experiment panels which contain three kinds of defects such as crack, inclusion and delamination. The collected signals of defects are analyzed by wavelet packet transform. The measure of separability based on distance is proposed as the criterion of evaluation of the wavelet bases to determine the optimum wavelet base. Then the signals of three types of defects are decomposed and reconstructed by wavelet packet transform using sym8 wavelet. And the energy features of all types of defects are extracted by the “frequency-energy” method. The results indicate that this method has a good effect on identifying the categories of carbon fiber composite defects, which provides a basis for automatic identification of defects.

Key words: wavelet packet, ultrasonic phased array, carbon fiber composite, measure of separability

摘要： 利用超声相控阵检测系统对含有裂纹、夹杂、分层三种缺陷的碳纤维复合材料实验板进行检测。对采集到的缺陷信号，在计算机中利用小波包变换进行分析。提出基于距离可分性测度的小波基评估准则，以确定最优小波基。对三种缺陷信号用sym8小波进行小波包的分解和重构，用“频率-能量”的方法提取各类缺陷信号的能量特征；结果表明，该方法对碳纤维复合材料的缺陷类型的区分具有较好的效果，从而为缺陷自动识别奠定了基础。

关键词: 小波包, 超声相控阵, 碳纤维复合材料, 可分性测度