基于小波包分析的超声铜-钢感应熔敷焊焊接界面缺陷特征参量提取

Extraction of defect parameters in UT of deposition weld interface between copper and steel based on wavelet packet analysis

通过信号截取和幅值归一化对铜一钢感应熔敷焊焊接界面超声检测信号进行了预处理研究。采用Db8小波基函数对预处理后的典型缺陷波形进行了3层小波包分解,提取了典型缺陷的时域特征和频域特征;利用欧式距离公式对时域、频域及时频域3种不同的特征向量下缺陷类型的可分性进行了比较。结果表明:选择时频域特征向量作为缺陷性质判定的依据,能确保缺陷分类的准确性与可靠性。

The UT signal was preprocessed through the signal intercept and amplitude normalization for the weld interface in dissimilar metal welding between copper and steel.In order to extract the time domain and frequency domain characteristics of the typical defects,the typical defect waveforms were decomposed by preprocessing three scale with Db8 wavelet function,then the separabilities of the defect type under different conditions of time domain,frequency domain and time- frequency domain were compared.It turned out that the accuracy and reliability of defect classification can be ensured by judging the quality of defect based on the time- frequency domain.