Duanjiao Li
Duanjiao Li has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Bin Sun
Hunan University
1 shared publication - Chao Yuan
College of Electrical and Information Engineering, Hunan University, Changsha 410082, China
1 shared publication - Dezhi Chen
State Key Laboratory of Advanced Electromagnetic Technology, Huazhong University of Science and Technology, Wuhan, China
1 shared publication - Feng Zhang
CSG Guangdong Power Grid Corporation, Guangzhou, China
1 shared publication - Jianming Liu
Guangdong Power GRID Co. Ltd., Guangzhou, China
1 shared publication - Linglong Cai
Guangdong Electric Power Science Institute, Guangdong Province 1 , 510080 Guangzhou, Guangdong
1 shared publication
Latest Publications
- Open Access
- 1 March 2025
- Published by AIP Publishing in AIP Advances
- Vol. 15 (3)
Research on electromagnetic vibration and noise of transformer under short circuit conditions
For an open-type 50 kV A three-dimensional laminated core transformer, this paper establishes a research framework for inter-turn short-circuit faults from the perspective of the power system. First, the B–H curve and magnetostriction curve of the silicon steel sheets were measured using a single-sheet silicon steel magnetostriction measurement system, providing data support for electromagnetic, vibration, and noise analysis. Based on finite element analysis, electromagnetic and vibration characteristics of the three-dimensional laminated core transformer were simulated and analyzed, establishing a steady-state model and verifying its validity. Second, an equivalent circuit model and a finite element model for the transformer winding under inter-turn short-circuit conditions were established. Simulations were conducted to analyze current, leakage magnetic field, and electrodynamic forces. Through simulation, the effects of different short-circuit timings on the transformer’s status were examined, revealing the size and distribution of axial and radial leakage magnetic fields, as well as axial and radial electrodynamic forces under different numbers of shorted turns. Finite element simulation software was used to calculate the magnetic field, vibration, and noise in the three-dimensional laminated core transformer following an inter-turn short circuit. This study delves into the magnetic field, vibration, and noise during normal operation and under varying numbers of shorted turns, analyzing the impact of inter-turn short circuits on the performance and stability of the transformer.
- 1 March 2025
- Published by AIP Publishing in AIP Advances
- Vol. 15 (3)
Open AccessResearch on electromagnetic vibration and noise of transformer under short circuit conditions
For an open-type 50 kV A three-dimensional laminated core transformer, this paper establishes a research framework for inter-turn short-circuit faults from the perspective of the power system. First, the B–H curve and magnetostriction curve of the silicon steel sheets were measured using a single-sheet silicon steel magnetostriction measurement system, providing data support for electromagnetic, vibration, and noise analysis. Based on finite element analysis, electromagnetic and vibration characteristics of the three-dimensional laminated core transformer were simulated and analyzed, establishing a steady-state model and verifying its validity. Second, an equivalent circuit model and a finite element model for the transformer winding under inter-turn short-circuit conditions were established. Simulations were conducted to analyze current, leakage magnetic field, and electrodynamic forces. Through simulation, the effects of different short-circuit timings on the transformer’s status were examined, revealing the size and distribution of axial and radial leakage magnetic fields, as well as axial and radial electrodynamic forces under different numbers of shorted turns. Finite element simulation software was used to calculate the magnetic field, vibration, and noise in the three-dimensional laminated core transformer following an inter-turn short circuit. This study delves into the magnetic field, vibration, and noise during normal operation and under varying numbers of shorted turns, analyzing the impact of inter-turn short circuits on the performance and stability of the transformer.
- 26 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-14
A Real-World Large-Scale Infrared Image Dataset and Multitask Learning Framework for Power Line Surveillance
Nowadays, power line surveillance widely utilizes infrared images collected by drones for instance segmentation to determine whether the power equipment is abnormal, such as overheating. With the development of deep learning, state-of-the-art instance segmentation methods can be effectively applied to infrared images of electrical scenarios. However, the low spatial resolution and limited data scale of power line infrared images pose challenges for these methods. In this paper, we make the following contributions: 1) We present a real-world large-scale dataset, CSG, which includes 11,715 power line infrared images with a spatial resolution of 640×512, and a total of 45,145 instances. The CSG dataset covers 7 foreground objects coupled with different backgrounds, and has various scenarios from 295 geo-sites, providing a benchmark for evaluating the performance of instance segmentation methods; 2) We design a multi-task learning framework, PowerNet, where super-resolution and instance segmentation are unified into a collaborative network. Super-resolution allows the preservation of high-resolution feature representations, while instance segmentation incorporates them through a knowledge transfer pipeline. This enhances the discriminative ability of objects, particularly those that are small and complex, leading to more accurate predictions under the condition of low-resolution images. The pipeline measures task affinity to transfer knowledge and is effectively supervised by multi-task loss. Extensive experiments reveal that the proposed framework achieves remarkable performance on the CSG dataset.
- 26 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-14
A Real-World Large-Scale Infrared Image Dataset and Multitask Learning Framework for Power Line Surveillance
Nowadays, power line surveillance widely utilizes infrared images collected by drones for instance segmentation to determine whether the power equipment is abnormal, such as overheating. With the development of deep learning, state-of-the-art instance segmentation methods can be effectively applied to infrared images of electrical scenarios. However, the low spatial resolution and limited data scale of power line infrared images pose challenges for these methods. In this paper, we make the following contributions: 1) We present a real-world large-scale dataset, CSG, which includes 11,715 power line infrared images with a spatial resolution of 640×512, and a total of 45,145 instances. The CSG dataset covers 7 foreground objects coupled with different backgrounds, and has various scenarios from 295 geo-sites, providing a benchmark for evaluating the performance of instance segmentation methods; 2) We design a multi-task learning framework, PowerNet, where super-resolution and instance segmentation are unified into a collaborative network. Super-resolution allows the preservation of high-resolution feature representations, while instance segmentation incorporates them through a knowledge transfer pipeline. This enhances the discriminative ability of objects, particularly those that are small and complex, leading to more accurate predictions under the condition of low-resolution images. The pipeline measures task affinity to transfer knowledge and is effectively supervised by multi-task loss. Extensive experiments reveal that the proposed framework achieves remarkable performance on the CSG dataset.
- Open Access
- 22 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 166-171
Development and Research of Image Recognition System for Substation Equipment based on Convolutional Neural Network
With the rapid development of national modernization, the safe and stable operation of substation electrical equipment plays a very important role in social stability and economic development. In this paper, the image recognition system of substation equipment is developed based on convolutional neural network to improve the accuracy and intelligence level of thermal fault detection of substation high-voltage electrical equipment. To solve the problem of low efficiency of traditional manual inspection, we adopted intelligent inspection robot equipped with infrared equipment and combined with deep learning technology for image analysis. An improved Faster RCNN infrared detection model is constructed to meet the challenges of poor infrared image quality and device diversity. Using VGG16 network as the basis of feature extraction, our model performs well in feature extraction of electrical equipment. After optimization, the overall recognition accuracy of the model in equipment category determination and location reaches 83.1%. However, due to the recognition errors caused by the similar appearance of some equipment parts, we further introduced the non-maximum suppression strategy to improve Faster RCNN, which effectively improved the recognition accuracy to 92.8%. This research not only provides technical support for the safety, stability and intelligent operation of substation, but also lays a solid foundation for the development of automatic fault diagnosis technology of electrical equipment.
- 22 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 166-171
Open AccessDevelopment and Research of Image Recognition System for Substation Equipment based on Convolutional Neural Network
With the rapid development of national modernization, the safe and stable operation of substation electrical equipment plays a very important role in social stability and economic development. In this paper, the image recognition system of substation equipment is developed based on convolutional neural network to improve the accuracy and intelligence level of thermal fault detection of substation high-voltage electrical equipment. To solve the problem of low efficiency of traditional manual inspection, we adopted intelligent inspection robot equipped with infrared equipment and combined with deep learning technology for image analysis. An improved Faster RCNN infrared detection model is constructed to meet the challenges of poor infrared image quality and device diversity. Using VGG16 network as the basis of feature extraction, our model performs well in feature extraction of electrical equipment. After optimization, the overall recognition accuracy of the model in equipment category determination and location reaches 83.1%. However, due to the recognition errors caused by the similar appearance of some equipment parts, we further introduced the non-maximum suppression strategy to improve Faster RCNN, which effectively improved the recognition accuracy to 92.8%. This research not only provides technical support for the safety, stability and intelligent operation of substation, but also lays a solid foundation for the development of automatic fault diagnosis technology of electrical equipment.