基于脉冲涡流的铁磁性材料屈服强度检测方法  

Yield Strength Testing Method of Ferromagnetic Materials Based on Pulsed Eddy Current

作　　者：李开宇[1] 高雯娟 王平[1] 张艳艳 杭成 LI Kaiyu;GAO Wenjuan;WANG Ping;ZHANG Yanyan;HANG Cheng(College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, 211106)

机构地区：[1]南京航空航天大学自动化学院

出　　处：《中国机械工程》2019年第18期2143-2149,共7页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(61527803);科技部重大科学仪器设备开发专项(2016YFB1100205,2016YFF0103702);南京航空航天大学研究生创新基地(实验室)开放基金资助项目(kfjj20170325)

摘　　要：钢铁工业中的铁磁性材料屈服强度的检测依赖拉伸检测,增加了检测成本,为此提出了一种多特征融合的铁磁性材料屈服强度脉冲涡流检测方法。提取脉冲涡流响应信号的时域特征、频域特征,然后建立各个信号特征与材料屈服强度的神经网络模型,最后用神经网络模型对材料的屈服强度进行估计。该方法是一种无损检测方法,检测误差不超过5%。Yield strength detection of ferromagnetic materials in iron and steel industries relied on loss detection, which greatly increased the cost of testing. So, a multi-feature fusion method for yield strength pulsed eddy current testing of ferromagnetic materials was proposed herein. Firstly, the time domain characteristics and frequency domain characteristics of pulsed eddy current response signals were extracted. And then, a neural network model for establishing the yield strength of each signal characteristic and material was established. Finally, the neural network model was used to predict the yield strength of the materials. As a non-destructive testing method, the method may achieve prediction errors less than 5%.

关 键 词：铁磁性材料 涡流检测 屈服强度 神经网络 特征融合 

分 类 号：TG115.28[金属学及工艺—物理冶金]