【作者】 魏莉；

【导师】 游凤荷；

【作者基本信息】 武汉理工大学 ， 检测技术与自动化装置， 2002， 硕士

【摘要】 金属表面硬化层深的检测及控制是目前机械工业部门急需解决的一个问题，它属于材质检测的范畴。用涡流法对它进行检测的最大难度在于检测信号的处理及分析，仅仅利用涡流阻抗的信息传统分析方法是不够的。为此，本文采用了脉冲涡流来进行测量，利用脉冲信号可以进行频谱分析的特点，对涡流信号进行频谱分析，提取信号特征，更有效的利用信号中包含的信息。由此得到的特征值便是对信号进一步分析和判断的依据。 人工神经网络是由许多以一定权值相互连接的计算单元组成的系统，它可以生成复杂的判决边界，所以适用于信号分类。本文选用基于误差反向传播(BP)算法的神经网络对信号分类，并给出分类结果，表明了神经网络在涡流材质检测信号处理中应用的巨大潜力。更多还原

【Abstract】 Nondestructive testing and quality control of metal hardened-depth is a important problem to be solved in machinery industry. It belongs to material property testing. The most difficult task in the testing is the processing and analyzing of the signal. Just making use of the traditional analyzing method of eddy current impedance information is not enough. In this article, pulsed eddy current is used and then spectrum analysis is performed to extract characteristic of the signal. So we can make use of the information more effectively.Artificial neural network is a system composed of many computational elements that are connected through an appropriate set of weights for generating complex decision surface. So it can be used in pattern classification. In this article we use a BP neural network to classify the eddy current signal and its result is also presented, which indicate that artificial neural network has vast potential in eddy current signal processing.更多还原