Tohoku University · Engineering
Professor Jiuhao Ge's research lab specializes in electromagnetic non-destructive evaluation (NDE) techniques, focusing on advanced eddy current testing and alternating current field measurement (ACFM) for the detection and characterization of surface and subsurface defects in metallic structures. The lab develops innovative signal processing and imaging methods—such as rotating eddy current testing, pulsed eddy current, and wavelet packet analysis—to enhance defect detection sensitivity, improve signal-to-noise ratio, and enable real-time classification of defects like cracks, pitting corrosion, and slits. A key research direction involves understanding and optimizing the velocity-insensitive behavior of ACFM for high-speed rail inspection, alongside developing adaptive signal processing algorithms for complex defect profiling and quantification.
Figures are computed from collected data and may differ slightly.
In industry, natural cracks usually exist in complex forms such as adjacent, colony, irregular, and crossing. Mere magnetic field is inadequate to accurately evaluate the complex cracks using eddy current testing. In this study, an improved eddy current imaging method using the signal of rotating eddy current testing (RECT) was proposed to reconstruct the surface profile of complex cracks. A transformation method was adopted to transform the signals of RECT into those under uniform eddy current
This study investigated the applicability of detecting pitting corrosions on stainless steel weld overlay cladding using uniform eddy current testing, which is insensitive to surface undulation, and rotating eddy current testing, which can be considered a superposition of uniform eddy current testing in orthogonal directions. The performance of the two techniques in detecting artificial pitting corrosions on and next to weld beads was compared. The signal distributions revealed that only rotatin
The detectability of an alternating current field measurement (ACFM) for non-surface defects is improved by a pulsed alternating current field measurement (PACFM) technique. In this paper, the optimal time-domain feature of the PACFM technique for non-surface defect detection is investigated by numerical and experimental methods. In the simulation, the numerical model of the PACFM technique is built by the finite element software COMSOL. The mechanism for non-surface defect detection and the tim
In this paper, a feature termed as the postpeak value is proposed for the pulsed eddy current technique (PECT). Moreover, a method using the postpeak value is proposed to classify surface and reverse defects. A PECT system is built for verification purposes. Experiment results prove that the postpeak feature value has better performance than that of the traditional peak value in the case of reverse defect detection. In contrast, the peak value is better than the postpeak value in the case of sur
The alternating current field measurement (ACFM) technique has been validated for its velocity-insensitive feature, which is a merit in the high-velocity detection of rails. However, little has been done to explain its velocity-insensitive mechanism. Moreover, as the detecting velocity varies during detection, the cutoff frequency in the key module lock-in amplifier of the ACFM device should be adaptable, and no method of determining the cutoff frequency is currently available. In the present st
In this Letter, a wavelet packet energy‐based algorithm is proposed to classify surface and non‐surface slits through the pulsed alternating current field measurement technique. Wavelet packet decomposition is used to obtain the coefficient of each frequency band. The slit is classified as non‐surface slit or surface slit according to the normalised energy of the first band. The quantification of the remaining ligament or depth of slits is performed using the entropy of the wavelet packet energy
Open papers in the app to read, cite, and organize with AI.
