基于相空间模糊熵算法的金属缺陷形状辨识

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.

Identification of Flaw Shape in Metal Material Based on Phase Space and Fuzzy Entropy

A metal flaw detection method based on eddy current testing is proposed. The size and shape of the flaws are identified by analyzing complexity of the measured signal with the aid of phase space and fuzzy entropy method. The multiple coil sensor models are designed in order to obtain valid flaw features from enough measurement infor-mation. The appropriate sensor parameters and excitation mode are selected through electromagnetic simulation. The fuzzy entropy is used for evaluating the complexity of measured signal with different size,depth and shape of defects. In order to accurately extract the inherent law of the eddy current signal and improve the sensitivity of the signal analysis algorithm,the phase space reconstruction is used for eddy current signal before fuzzy-entropy analysis. The results show that the fuzzy entropy,i. e. the complexity,of eddy current signal is increased with the growth of flaw size. Furthermore,different shapes of the flaws can be effectively distinguished by the profile of mean fuzzy entropy.