摘要
介绍基于神经网络的钢材表面裂纹定量检测原理及方法 ,设计了基于LabVIEW的钢杆裂纹缺陷漏磁定量检测系统 ,并对钢杆表面横向裂纹缺陷进行模拟实验 ,运用神经网络技术初步建立数学模型 ,通过评判Vpp得到反映裂纹状况的定量检测结果 ,从而验证了该方法的可行性 。
The quantitative testing principle for the crack magnetic flux leakage signal on steels based on neural network is presented on this paper.quantitative testing system for crack defect on steels surface based on LabVIEW is designed,and simulation experiments are done,mathematical model is built by the use of the neural network technology,the quantitative testing result to reflect on condition of crack can be acquired by quantitative analysis the signal peak_peak value.experiment results verified that this method is possible,and relative results are achieved.
出处
《自动化技术与应用》
2003年第4期70-73,共4页
Techniques of Automation and Applications
