Fangxing Lyu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Fangxing Lyu has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Fei Li
School of Electronic Engineering, Xian Shiyou University, Xian, China
3 shared publications Chunyang Han
Xi’an Jiaotong University, State Key Laboratory of Electrical Insulation and Power Equipment, Xi’an
2 shared publicationsHui Ding
Xi’an Jiaotong University, State Key Laboratory of Electrical Insulation and Power Equipment, Xi’an
2 shared publicationsXin Fang
School of Computer Science Xi'an Shiyou University Xi'an China
2 shared publicationsXu Niu
Xi’an Jiaotong University, State Key Laboratory of Electrical Insulation and Power Equipment, Xi’an
2 shared publicationsA. Xin Fang
Xi'an Key Laboratory of Intelligent Equipment Development for Oil, Gas and Renewable Energy School of Computer Science, Xi'an Shiyou University Xi'an China
1 shared publication
Latest Publications
- 12 July 2026
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 38 (14)
GMS ‐ RT ‐ DETR : A Geometry‐Aware and Lightweight Model for Real‐Time Insulator Defect Detection in UAV Power Line Inspection
To address the critical challenges of complex background interference, subtle micro‐crack detection, and the geometric modeling of slender targets with extreme aspect ratios in UAV power line inspection, this paper proposes GMS‐RT‐DETR (Geometric & Multi‐Scale RT‐DETR), a geometry‐aware and lightweight real‐time detection model. To reduce computational complexity for edge deployment without sacrificing feature representation, the StarNet lightweight architecture is adopted to replace the original ResNet‐18 backbone. By leveraging star operations (element‐wise multiplication), it achieves high‐dimensional nonlinear feature mapping within a compact dimensional space. Furthermore, to address the scale variance between macroscopic insulator contours and microscopic local defects, a Spatial Pyramid Pooling‐Fast (SPPF) module is embedded at the terminal stage of the backbone, utilizing serial multi‐kernel pooling to expand the effective receptive field. Within the subsequent hybrid encoder, a Split‐Directional Aware Attention Module (DAAM) is designed to capture geometric features of slender targets and suppress background noise. This module employs a channel decoupling strategy and asymmetric convolutions to accurately capture linear structures and directional cracks. To further improve bounding box regression for targets with extreme aspect ratios, a Direction‐Aware Intersection over Union (DA‐IoU) loss function is proposed, introducing an aspect‐ratio‐guided dynamic weighting mechanism to adaptively penalize localization errors along the shorter side. Experimental results demonstrate that GMS‐RT‐DETR achieves 94.3% mAP50 and 95.1% precision on the insulator defect dataset, representing absolute improvements of 4.2% and 4.0% over the baseline, respectively. In terms of model efficiency, the parameter count is reduced by 40.7% (from 18.9 to 11.2 M), the computational cost is lowered to 32.5 GFLOPs, and the inference speed reaches 79.4 FPS. These metrics satisfy the real‐time and edge‐deployment requirements of autonomous UAV power line inspection.
- 12 July 2026
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 38 (14)
GMS ‐ RT ‐ DETR : A Geometry‐Aware and Lightweight Model for Real‐Time Insulator Defect Detection in UAV Power Line Inspection
To address the critical challenges of complex background interference, subtle micro‐crack detection, and the geometric modeling of slender targets with extreme aspect ratios in UAV power line inspection, this paper proposes GMS‐RT‐DETR (Geometric & Multi‐Scale RT‐DETR), a geometry‐aware and lightweight real‐time detection model. To reduce computational complexity for edge deployment without sacrificing feature representation, the StarNet lightweight architecture is adopted to replace the original ResNet‐18 backbone. By leveraging star operations (element‐wise multiplication), it achieves high‐dimensional nonlinear feature mapping within a compact dimensional space. Furthermore, to address the scale variance between macroscopic insulator contours and microscopic local defects, a Spatial Pyramid Pooling‐Fast (SPPF) module is embedded at the terminal stage of the backbone, utilizing serial multi‐kernel pooling to expand the effective receptive field. Within the subsequent hybrid encoder, a Split‐Directional Aware Attention Module (DAAM) is designed to capture geometric features of slender targets and suppress background noise. This module employs a channel decoupling strategy and asymmetric convolutions to accurately capture linear structures and directional cracks. To further improve bounding box regression for targets with extreme aspect ratios, a Direction‐Aware Intersection over Union (DA‐IoU) loss function is proposed, introducing an aspect‐ratio‐guided dynamic weighting mechanism to adaptively penalize localization errors along the shorter side. Experimental results demonstrate that GMS‐RT‐DETR achieves 94.3% mAP50 and 95.1% precision on the insulator defect dataset, representing absolute improvements of 4.2% and 4.0% over the baseline, respectively. In terms of model efficiency, the parameter count is reduced by 40.7% (from 18.9 to 11.2 M), the computational cost is lowered to 32.5 GFLOPs, and the inference speed reaches 79.4 FPS. These metrics satisfy the real‐time and edge‐deployment requirements of autonomous UAV power line inspection.
- 25 March 2026
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 38 (7) , e70682
A Lightweight and Efficient Insulator Defect Detection Model for Unmanned Aerial Vehicle Inspection
Insulator defect detection is crucial for ensuring the safety of power grids, but it faces challenges such as complex environmental interference, small‐scale target recognition, and edge device deployment. This paper proposes an improved lightweight model based on YOLOv11. The model takes Faster‐Net as the backbone and uses partial convolution to reduce computational costs. The neck network is constructed through hierarchical feature fusion blocks to enhance multiscale feature fusion. Introduce a Parallel Patch‐aware Attention (PPA) module into the detection head to enable the model to adaptively focus on the key defect areas in the image. Meanwhile, a Lightweight Adaptive Extraction (LAE) module is integrated to optimize the feature calculation process. Experiments show that this model achieves a mAP50 accuracy rate of 93.2% and a precision of 93.7% on public datasets, with the parameter count and computational load being only 2.3 M and 4.6 GFLOPs respectively. The ablation experiment verified the effectiveness of the module, and cross‐dataset tests confirmed the excellent generalization performance. The model was successfully deployed on the Jetson Xavier NX platform and the M350 Unmanned Aerial Vehicle (UAV), achieving real‐time end‐to‐end detection. This research provides a complete solution for intelligent power inspection in complex environments that is highly accurate, lightweight and easy to deploy.
- 25 March 2026
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 38 (7) , e70682
A Lightweight and Efficient Insulator Defect Detection Model for Unmanned Aerial Vehicle Inspection
Insulator defect detection is crucial for ensuring the safety of power grids, but it faces challenges such as complex environmental interference, small‐scale target recognition, and edge device deployment. This paper proposes an improved lightweight model based on YOLOv11. The model takes Faster‐Net as the backbone and uses partial convolution to reduce computational costs. The neck network is constructed through hierarchical feature fusion blocks to enhance multiscale feature fusion. Introduce a Parallel Patch‐aware Attention (PPA) module into the detection head to enable the model to adaptively focus on the key defect areas in the image. Meanwhile, a Lightweight Adaptive Extraction (LAE) module is integrated to optimize the feature calculation process. Experiments show that this model achieves a mAP50 accuracy rate of 93.2% and a precision of 93.7% on public datasets, with the parameter count and computational load being only 2.3 M and 4.6 GFLOPs respectively. The ablation experiment verified the effectiveness of the module, and cross‐dataset tests confirmed the excellent generalization performance. The model was successfully deployed on the Jetson Xavier NX platform and the M350 Unmanned Aerial Vehicle (UAV), achieving real‐time end‐to‐end detection. This research provides a complete solution for intelligent power inspection in complex environments that is highly accurate, lightweight and easy to deploy.
- Open Access
- 1 January 2026
- Published by Wiley in International Transactions on Electrical Energy Systems
- Vol. 2026 (1) , 5526613
An Investigation of Dust Deposition Patterns and Electric Field Distortion on Insulators Based on Finite Element Simulation
In dust‐prone regions such as Northwest China, dust accumulation on insulators distorts the electric field, which is a key cause of flashovers and performance degradation. However, existing studies have mainly focused on the chemical properties of contaminants or equivalent salt deposit density. As a result, the physical mechanism by which the geometric morphology of dust deposits independently affects the electric field distribution remains poorly understood. To address this gap, this study employs finite element simulation to systematically investigate the contact angle θ at the dust–insulator interface as a key geometric parameter governing electric field distortion. An electrostatic field model of the XP‐70 disc‐type suspension insulator was established to simulate and analyze electric field distributions under different dust accumulation morphologies. Findings indicate that when θ approaches 90°, severe electric field distortion occurs in the contact region; conversely, when θ approaches 0° or 180°, the electric field distribution becomes more uniform with significantly reduced distortion. This study provides a quantitative analysis of the correlation between contact angle and electric field distortion in dust accumulation patterns, offering a new perspective on the distortion mechanism caused by dust contamination. It offers a potential theoretical basis for antiflashover design and structural optimization of insulators in severe dust environments, holding significant value for ensuring power grid operational safety.
- 1 January 2026
- Published by Wiley in International Transactions on Electrical Energy Systems
- Vol. 2026 (1) , 5526613
Open AccessAn Investigation of Dust Deposition Patterns and Electric Field Distortion on Insulators Based on Finite Element Simulation
In dust‐prone regions such as Northwest China, dust accumulation on insulators distorts the electric field, which is a key cause of flashovers and performance degradation. However, existing studies have mainly focused on the chemical properties of contaminants or equivalent salt deposit density. As a result, the physical mechanism by which the geometric morphology of dust deposits independently affects the electric field distribution remains poorly understood. To address this gap, this study employs finite element simulation to systematically investigate the contact angle θ at the dust–insulator interface as a key geometric parameter governing electric field distortion. An electrostatic field model of the XP‐70 disc‐type suspension insulator was established to simulate and analyze electric field distributions under different dust accumulation morphologies. Findings indicate that when θ approaches 90°, severe electric field distortion occurs in the contact region; conversely, when θ approaches 0° or 180°, the electric field distribution becomes more uniform with significantly reduced distortion. This study provides a quantitative analysis of the correlation between contact angle and electric field distortion in dust accumulation patterns, offering a new perspective on the distortion mechanism caused by dust contamination. It offers a potential theoretical basis for antiflashover design and structural optimization of insulators in severe dust environments, holding significant value for ensuring power grid operational safety.
- 11 September 2025
- Published by Elsevier in Geoenergy Science and Engineering
- Vol. 257
Research on wellbore trajectory control of Rotary Steerable System using back-propagation neural network-fuzzy method
- 11 September 2025
- Published by Elsevier in Geoenergy Science and Engineering
- Vol. 257
Research on wellbore trajectory control of Rotary Steerable System using back-propagation neural network-fuzzy method
- Open Access
- 7 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 129650-129659
A High-Frequency Measurement Method of Downhole Vibration Signal Based on Compressed Sensing Technology and Its Application in Drilling Tool Failure Analysis
The high-frequency information of downhole vibration can record more specific details about the dynamic response of bottom hole assembly (BHA), which is conducive to analyzing downhole abnormal vibration and diagnosing drilling tool failure. However, high frequency measurements would generate a large amount of data, which is a huge challenge for data storage. To cope with this challenge, this work proposes applying compressive sensing theory to high-frequency measurement of downhole vibration signals and failure analysis of drilling tools. First, three ADCs clocked at different sub-Nyquist frequencies are used to collect and store downhole vibration signals. Then, the stored sparse vibration data is reconstructed into complete vibration signal by an improved OMP signal reconstruction algorithm. The performance of the proposed method has been investigated via a series of experiments by applying simulation and real vibration signals, which show that the proposed method can reconstruct the downhole vibration signal well and significantly reduce the pressure of acquisition and data storage. This method provides a new way to obtain the high-frequency measurement data of downhole vibration, and provides a data basis for further mastering the comprehensive law of downhole vibration, diagnosing abnormal downhole vibration, and drilling tool failure.
- 7 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 129650-129659
Open AccessA High-Frequency Measurement Method of Downhole Vibration Signal Based on Compressed Sensing Technology and Its Application in Drilling Tool Failure Analysis
The high-frequency information of downhole vibration can record more specific details about the dynamic response of bottom hole assembly (BHA), which is conducive to analyzing downhole abnormal vibration and diagnosing drilling tool failure. However, high frequency measurements would generate a large amount of data, which is a huge challenge for data storage. To cope with this challenge, this work proposes applying compressive sensing theory to high-frequency measurement of downhole vibration signals and failure analysis of drilling tools. First, three ADCs clocked at different sub-Nyquist frequencies are used to collect and store downhole vibration signals. Then, the stored sparse vibration data is reconstructed into complete vibration signal by an improved OMP signal reconstruction algorithm. The performance of the proposed method has been investigated via a series of experiments by applying simulation and real vibration signals, which show that the proposed method can reconstruct the downhole vibration signal well and significantly reduce the pressure of acquisition and data storage. This method provides a new way to obtain the high-frequency measurement data of downhole vibration, and provides a data basis for further mastering the comprehensive law of downhole vibration, diagnosing abnormal downhole vibration, and drilling tool failure.
- Open Access
- 16 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 55350-55359
An Enhanced Photonic-Assisted Sampling Approach for Spectrum-Sparse Signal by Compressed Sensing
Spectrum-sparse signals are vital for wideband radar and wireless communication applications. A high-speed analog-to-digital converter (ADC) with the capacities of tens of gigahertz sampling rates is often required to acquire these signals. In this work, an enhanced photonic-assisted sampling approach with the combination of the photonic-assisted time-interleaved ADC and compressed sensing techniques is presented, which enables the measured signal to be reconstructed through very few samples by utilizing the sparsity of the spectrum-sparse signal. An ultrahigh spectral resolution Fourier dictionary was introduced to suppress the spectrum leakage and obtain the actual sparse expression of the spectrum-sparse signals. Moreover, a layered tracking orthogonal matching pursuit signal recovery algorithm was employed to reduce computational complexity and enhance processing speed. The performance of the proposed approach has been investigated via simulations and laboratory experiments by varying the applied spectrum-sparse signals over 100 times. The experimental results demonstrate that the proposed method can capture the blind-frequency spectrum-sparse signal at an equivalent sampling rate of 1 GS/s by utilizing four parallel ADCs with a sampling rate of 50 MS/s. It is proven that the proposed approach achieves $\sim 5$ times higher equivalent sampling rate than that of the conventional PTIADC at the same sampling rate. This work provides a valuable method for acquiring spectrum sparse signals in practical applications.
- 16 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 55350-55359
Open AccessAn Enhanced Photonic-Assisted Sampling Approach for Spectrum-Sparse Signal by Compressed Sensing
Spectrum-sparse signals are vital for wideband radar and wireless communication applications. A high-speed analog-to-digital converter (ADC) with the capacities of tens of gigahertz sampling rates is often required to acquire these signals. In this work, an enhanced photonic-assisted sampling approach with the combination of the photonic-assisted time-interleaved ADC and compressed sensing techniques is presented, which enables the measured signal to be reconstructed through very few samples by utilizing the sparsity of the spectrum-sparse signal. An ultrahigh spectral resolution Fourier dictionary was introduced to suppress the spectrum leakage and obtain the actual sparse expression of the spectrum-sparse signals. Moreover, a layered tracking orthogonal matching pursuit signal recovery algorithm was employed to reduce computational complexity and enhance processing speed. The performance of the proposed approach has been investigated via simulations and laboratory experiments by varying the applied spectrum-sparse signals over 100 times. The experimental results demonstrate that the proposed method can capture the blind-frequency spectrum-sparse signal at an equivalent sampling rate of 1 GS/s by utilizing four parallel ADCs with a sampling rate of 50 MS/s. It is proven that the proposed approach achieves $\sim 5$ times higher equivalent sampling rate than that of the conventional PTIADC at the same sampling rate. This work provides a valuable method for acquiring spectrum sparse signals in practical applications.
- Open Access
- 26 November 2019
- Published by SPIE-Intl Soc Optical Eng in Optical Engineering
- Vol. 58 (11) , 1-7
Sub-Nyquist sampling method and its application in high-frequency electric-field measurement
The flagship monthly journal of SPIE, Optical Engineering (OE) publishes peer-reviewed papers reporting on research and development in all areas of optics, photonics, and imaging science and engineering.
- 26 November 2019
- Published by SPIE-Intl Soc Optical Eng in Optical Engineering
- Vol. 58 (11) , 1-7
Open AccessSub-Nyquist sampling method and its application in high-frequency electric-field measurement
The flagship monthly journal of SPIE, Optical Engineering (OE) publishes peer-reviewed papers reporting on research and development in all areas of optics, photonics, and imaging science and engineering.
- 15 June 2018
- Published by SPIE-Intl Soc Optical Eng in Optical Engineering
- Vol. 57 (06) , 066111-7
Wavelength–time domain transformation-based demodulator for wavelength-encoded optical sensor
As for the issue of the rapid wavelength interrogation of optical fiber wavelength-encoded sensor, a demodulation method combining spectrum-to-time domain information mapping and photonics time stretch technology is proposed. Using group velocity dispersion effect of dispersive elements, the spectrum of optical pulse will be mapped to a temporal waveform whose intensity mimics the spectrum. The temporal waveform is further stretched in timescale so that it is slow enough to be captured by electronic analog-to-digital converter. The performance of the proposed demodulation method is experimentally demonstrated to demodulate a wavelength-encoded sensor.
- 15 June 2018
- Published by SPIE-Intl Soc Optical Eng in Optical Engineering
- Vol. 57 (06) , 066111-7
Wavelength–time domain transformation-based demodulator for wavelength-encoded optical sensor
As for the issue of the rapid wavelength interrogation of optical fiber wavelength-encoded sensor, a demodulation method combining spectrum-to-time domain information mapping and photonics time stretch technology is proposed. Using group velocity dispersion effect of dispersive elements, the spectrum of optical pulse will be mapped to a temporal waveform whose intensity mimics the spectrum. The temporal waveform is further stretched in timescale so that it is slow enough to be captured by electronic analog-to-digital converter. The performance of the proposed demodulation method is experimentally demonstrated to demodulate a wavelength-encoded sensor.