[1]
Lift-Off-Robust Eddy-Current Pulsed Thermography Using an LCC Resonant Drive
Yahui Jia, Ling Wen, Chen Li, Zhongchi Zhang, Jingyuan Yang, Gui-Yun Tian, and Qiuji Yi
Under review IEEE Transactions on Instrumentation and Measurement
Abstract: Eddy-current pulsed thermography provides full-field inspection of conductive structures, but probe lift-off weakens coil-specimen coupling and the resulting thermal response before image acquisition. This work treats lift-off compensation as an excitation-hardware problem and introduces an LCC resonant network operated at fixed frequency to preserve specimen-side heating as lift-off increases. Steel-specimen experiments over 0–20 mm lift-off demonstrate substantially improved thermal-response retention and defect signal-to-noise ratio compared with the fixed-drive baseline.
[2]
Embedded Inductive Sensing and Fourier Neural-Operator Global Curvature Estimation in Heliostat Mirrors
Xiaolong Lu, Ruonan Li, Jingyuan Yang, Zongwen Wang, Yunhua Tan, Guiyun Tian, Wai Lok Woo, and Qiuji Yi
Under review IEEE Transactions on Industrial Informatics
Abstract: This work presents an embedded inductive sensing and operator-learning framework for heliostat mirror geometry monitoring from sparse support-point measurements. Inductance-to-digital-converter eddy-current sensors provide digitally addressable gap measurements, which are mapped to a fixed grid and processed by a curvature-conditioned Fourier neural operator. The framework combines global spectral modelling, local feature transformation and curvature-aware conditioning to estimate mirror-scale geometry, with validation against structured-light measurements.
[3]
Thermography-Tomographic Imaging: Integration of Pulse Compression Thermography and Virtual Wave for Physics-Based Learning
Dong Wang, Marco Ricci, Stefano Laureti, Xiaolong Lu, Rocco Zito, Stefano Sfarra, Peter Burgholzer, Guiyun Tian, Wai Lok Woo, and Qiuji Yi
Published IEEE Transactions on Industrial Informatics, 22(8), 7313–7324, 2026
Abstract: The study integrates pulse-compression thermography with the virtual-wave transformation to address depth-dependent blurring in infrared thermographic inspection. Thermal responses are analysed before and after virtual-wave transformation, showing improved signal-to-noise behaviour and stronger linearity for defect characterisation. A deep spatio-temporal fusion network is then trained to reproduce the virtual-wave transformation, enabling efficient and physics-guided tomographic reconstruction.