基于相空间模糊熵算法的金属缺陷形状辨识

发布时间：2018-07-08 08:45

本文选题：涡流检测 + 复杂度　；参考：《传感技术学报》2017年05期

【摘要】：提出了一种多线圈涡流无损检测方法,通过相空间模糊熵算法分析涡流信号复杂度,进而实现对金属微小缺陷形状的辨识。为了从足够的测量信息中获取有效的缺陷特征,设计了多线圈传感器模型。通过仿真实验选取适合的传感器参数和激励模式。采用相空间模糊熵算法,研究不同大小、深度、形状的缺陷对涡流信号复杂度的影响。为了准确提取涡流信号的内在规律,获得对缺陷敏感的信号分析结果,对涡流信号进行相空间重构,并在重构的相空间中计算信号的模糊熵。分析结果表明:随着缺陷体积的增加,模糊熵增大,涡流信号的复杂度增加。根据不同形状缺陷的模糊熵均值分布图,可以实现对孔、洞、裂缝3种缺陷较精确的区分。
[Abstract]:A multi-coil eddy current nondestructive testing method is proposed in this paper. The complexity of eddy current signal is analyzed by phase space fuzzy entropy algorithm, and the shape identification of metal microdefects is realized. In order to obtain effective defect features from enough measurement information, a multi-coil sensor model is designed. The suitable sensor parameters and excitation modes are selected through simulation experiments. A phase space fuzzy entropy algorithm is used to study the effects of different sizes, depths and shapes on the complexity of eddy current signals. In order to accurately extract the inherent rule of eddy current signal and obtain the result of signal analysis which is sensitive to defect, the phase space reconstruction of eddy current signal is carried out, and the fuzzy entropy of signal is calculated in the reconstructed phase space. The results show that with the increase of defect volume, the fuzzy entropy increases and the complexity of eddy current signal increases. According to the distribution map of fuzzy entropy mean value of different shape defects, it is possible to distinguish accurately three kinds of defects: hole, hole and crack.
【作者单位】：天津工业大学电子与信息工程学院;天津工业大学理学院;天津大学电气与自动化工程学院;
【基金】：国家科技支撑计划重点项目(2013BAF06B00) 国家自然科学基金项目(61373104,61402330,61405143,61601324) 高等学校博士点专项科研基金项目(20131201120002) 天津市应用基础与前沿技术研究计划项目(15JCQNJC01500)
【分类号】：TG115.28

【相似文献】