摘要
采用人工神经网络完成涡流无损检测中缺损的定量识别 .首先 ,用数值方法分析了阻抗增量的幅值、相位与缺损尺度的关系 ,得出结论 :用相位值来表征缺损深度效果更好 ,精度更高 .然后 ,应用小波边缘检测方法确定的信号特征值作为网络的输入 ,结果表明 :计算量大为减小 。
Artificial neural network(ANN) is applied to eddy current nondestructive testing( ECNT) for detecting defects. After establishing the correlation between the amplitude or phase with the defect size, phase value was found to correlate better with the depth of defect. Wavelet edge detecting method can be used finding free edges. The results were used as inputs to ANN for analysis.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2000年第6期6-10,共5页
Journal of Xi'an Jiaotong University
基金
教育部博士点基金资助项目 !(980 6 982 1 )
