Development of High-Sensitive Detection and Diagnosis System

这项研究是在国际联合研究计划中开展的，旨在开发高灵敏度的检测和诊断系统。为了满足新的检测目标，我们在涡流无损检测中开发了各种传感器和检测方法。结果表明，与传统的ECT系统相比，本研究开发的先进ECT系统的缺陷敏感性和表征能力有了很大的提高。此外，随着这种测量方法的数值逼近，我们还发展了裂纹位置、形状和深度的估计方法。详细结果见本报告列出的每篇论文。在这里，我们将向您展示在该联合研究中获得的典型改进和发明的内容。(1)用于导电板和管的磁传感器：-旋转磁通型传感器和那些使用参考板的差动型探头，-移动磁通型传感器，-使用阴影线圈及其差动型的移动磁通型传感器，-矩阵型传感器和那些差动型。(2)激发方法：-使用旋转磁场来获得磁通量矢量中的有益信息，-使用移动磁场作为不均匀磁场，-利用多频率分量克服集肤效应(MFE)。(3)数值方法(缺陷估计的逆问题)：-奇异值分解边界元逆分析方法-模糊推理边界元逆分析方法，-前馈神经网络裂纹识别方法。(4)包含缺陷信息的旋转巴克豪森信号：-旋转巴克豪森信号中混沌现象的新发现。

This research has been carried out to develop a high-sensitive detection and diagnosis system in the international joint research program. We have developed various kinds of sensors and detecting methods in the eddy current non-destructive testing (ECT), in order to meet the new goals. As a result, flaw sensitivity and characterization ability in the advanced ECT system developed in this research, were considerably improved in comparison with the conventional ones. Furthermore, coming numerical approach with this measuring method, we have also developed me estimation methods of crack position, shape, and depth. The detailed results were shown in each paper listed in this report. Here we will show you the contents of typical improvements and inventions obtained in this joint research.(1) Magnetic Sensors for Conducting Plates and Tubes:-Rotational magnetic flux type sensors and those differential type probes using a reference plate,-Moving magnetic flux type sensor,-Moving magnetic flux type sensor using shading coil and its differential types,-Matrix type sensors and those differential types.(2) Excitation Methods:-Use of rotational magnetic field to obtain beneficial information in the magnetic flux vector,-Use of moving magnetic field as an inhomogeneous magnetic field,-Use of multi-frequency components to overcome the skin effect (MFES).(3) Numerical Approach (Inverse Problem to Estimate Defects):-Inverse analysis method by BEM with singular value decomposition-Inverse analysis method by BEM with Fuzzy Inference,-Crack identification method with feed-forward neural network.(4) Rotational Barkhausen Signal Including Information of Defects :-New discovery of chaos phenomena in the rotational Barkhausen signals.