基于特征向量动态轨迹分析的脉冲涡流检测技术研究

项目名称： 基于特征向量动态轨迹分析的脉冲涡流检测技术研究

项目编号： No.51267016

项目类型： 地区科学基金项目

立项/批准年度： 2013

项目学科： 电工技术

项目作者： 付跃文

作者单位： 南昌航空大学

项目金额： 49万元

中文摘要： 脉冲涡流检测技术近年在飞机多层结构腐蚀检测等方面取得了进展，但在提离效应抑制、检测灵敏度及系统优化方法等方面，与工业界检测需求仍有相当差距。本项目提出从信号特征向量的动态轨迹方面进行研究。内容包括：从检测信号中提取频域或时域特征量组成特征向量，随后提取检测传感器移动过程中特征向量点在向量空间中形成的动态轨迹，通过分析提离等干扰信号和不同缺陷信号在动态轨迹方面的区别，抑制在实际应用中很强的提离等干扰信号及对缺陷进行检测、分类、定量；同时依据特征向量的动态轨迹，针对具有特定厚度、电导率等物理参数及对检测重点区域的要求，对检测频率和传感器等系统配置及参数进行优化选择；利用特性均匀传感器阵列来模拟单探头移动情形；比较传统方法与本方法在性能方面的区别。项目将特征向量与动态轨迹相结合，目的在于提高脉冲涡流检测灵敏度及抗干扰能力，为系统优化提供有效方法，为该技术满足航空航天工业的高要求建立基础。

中文关键词： 脉冲涡流；特征向量；动态轨迹；信号处理；快速傅里叶变换

英文摘要： Advancements have been made in pulsed eddy current (PEC) testing technology in recent years in such as corrosion detection of multi-layered metal structure in aircrafts. However, improvements in lift-off reduction,detection sensitivity and system optimization are still demanded by industry. In this proposal, we propose to address these issues by analylizing dynamic locus of feasure vector, which includes several aspects to be explored. Extract features in time domain or frequency domain and form a vector from them, then dynamic locus of the vector during the probe moves is computed. Interferences such as lift-off effect, usually intense enough in dection, are reduced and defects are detected, classified and quantified by analyzing difference in dynamical locus between interference and different defect signals. System is optimized in sensor, frequency and detection parameters with respect to thickness and conductivity based on dynamical locus. Sensor array with uniform characteristic will be made to simulate movement of single probe. Compare the performance of the proposed method with the traditional PEC method. The proposed project combines feature vector and dynamical locus to analyze signal, aiming at improving detection sensibility and anti-interference performance, providing effective method for system optim

英文关键词： pulsed eddy current；feature vector；dynamic trajectory；signal processing；fast Fourier transform