摘要
采用小波多分辨分析和快速傅里叶变换,对焊接裂纹金属磁记忆信号进行了处理,找出了焊接裂纹存在的磁记忆信号的判据特征,并建立了该特征的阈值。现场检测结果表明:采用小波多分辨分析和快速傅里叶变换得出的特征,可以准确判断出裂纹等缺陷是否存在。
Based on multiscale wavelet analysis and fast Fourier transform, metal magnetic memory signals were processed for welding cracks. A criterion feature to identify the existence of welding cracks was found out. Corresponding threshold value of the feature was established. Field inspection indicated that welding cracks could be accurately identified with the criterion feature.
出处
《材料工程》
EI
CAS
CSCD
北大核心
2008年第1期34-37,共4页
Journal of Materials Engineering
基金
国家自然科学基金资助项目(50475113)
高等学校博士学科点专项基金资助项目(20030056002)
关键词
金属磁记忆
小波分析
多尺度
特征提取
metal magnetic memory
wavelet analysis
multiscale
feature extraction
