Abstract:Intelligent perception is the cornerstone of digitalisation of power grid. Due to the closest connection with users, the information perception ability of distribution network directly affects the reliability of power supply. Distribution network has a wide range of geographical distribution and equipment types. The application of intelligent perception is with high complexity and dynamics under the limitation of both technology and management. Therefore, the study first sorts out the current application statu… Show more
“…In addition, F1 score of TRM has a large improvement in all classes except for switch. Some reasons can be found: (1) The number of training samples of switch is relatively small; (2) switch and bus are similar from the perspective of appearance as well as object relationship. If more samples of this class can be added, the accuracy will be further improved.…”
Section: Comparison With Related Workmentioning
confidence: 99%
“…Construction of intelligent substations has become increasingly important in recent years. One typical scenario in these substations is the application of inspection robots, which is hoped to significantly improve the efficiency of operation and maintenance work [1,2]. For automatic inspection task, detectable faults from pictures include visual defects and abnormal temperature rise.…”
Power equipment detection and localization is the key component of automatic inspection tasks in substations. To solve the challenges such as complex environment and lack of training data, utilization of context information is considered here. For substation scenes, object relation modelling (RM) is proven to be useful, but an end‐to‐end efficient framework which is suitable for non‐Convolutional Neural Networks (CNN) is still missing. Therefore, an extended Transformer network with an elaborately designed RM module is proposed. As the foundation, transformer network is better than CNN in terms of context dependency construction. On top of that, an RM module is plugged to adjust the decoded feature embeddings based on their appearance, position and class information. The module is based on a graph attention neural network which uses similarity as weights of nodes. The experiments show that the proposed method has a 16.2% improvement in accuracy compared to pipeline, and even 6.4% higher than the most recent models, largely promoting the construction of intelligent substations.
“…In addition, F1 score of TRM has a large improvement in all classes except for switch. Some reasons can be found: (1) The number of training samples of switch is relatively small; (2) switch and bus are similar from the perspective of appearance as well as object relationship. If more samples of this class can be added, the accuracy will be further improved.…”
Section: Comparison With Related Workmentioning
confidence: 99%
“…Construction of intelligent substations has become increasingly important in recent years. One typical scenario in these substations is the application of inspection robots, which is hoped to significantly improve the efficiency of operation and maintenance work [1,2]. For automatic inspection task, detectable faults from pictures include visual defects and abnormal temperature rise.…”
Power equipment detection and localization is the key component of automatic inspection tasks in substations. To solve the challenges such as complex environment and lack of training data, utilization of context information is considered here. For substation scenes, object relation modelling (RM) is proven to be useful, but an end‐to‐end efficient framework which is suitable for non‐Convolutional Neural Networks (CNN) is still missing. Therefore, an extended Transformer network with an elaborately designed RM module is proposed. As the foundation, transformer network is better than CNN in terms of context dependency construction. On top of that, an RM module is plugged to adjust the decoded feature embeddings based on their appearance, position and class information. The module is based on a graph attention neural network which uses similarity as weights of nodes. The experiments show that the proposed method has a 16.2% improvement in accuracy compared to pipeline, and even 6.4% higher than the most recent models, largely promoting the construction of intelligent substations.
“…The visual system is the key part for the robots to detect the power components such as insulators and drop fuses in order to recognise and locate them. Meanwhile, the visual system with real-time detection enables the live working robots in the distribution network to keep the stability of the operation and maintenance of the power grid system [17]. In the new century, deep learning has made amazing breakthroughs in the field of detection [18][19][20].…”
In this study, an autonomous robot navigation system is designed for live working on distribution line. The developed system features a real‐time detection and motion planning system, incorporating a manipulator capable of grasping power components. In order to accurately identify targets, the authors propose an object detection method based on the Larger Scale ‘You Only Look Once’ Version 4 (LS‐YOLOv4) algorithm for detecting the insulators and drop fuses. The LS‐YOLOv4 extracts features of power components by Convolutional Neural Network (CNN), and then performs feature fusion. Then the authors develop a motion planning method based on the Node Control Optimal Rapidly Exploring Random Trees (NC‐RRT*), which can drive the robot to realise the autonomous robot motion planning and obstacle avoidance. On the grasping function, the authors present a reliable Lightweight‐based Convolutional Neural Network (L‐CNN) grasping point detection method. Finally, the authors evaluate fully autonomous robotic system in both simulated and real‐world experiments. The experimental results demonstrate that the proposed system can effectively identify the target and complete the grasping task in an efficient way. Notably, the proposed motion planning method can take into account both planning efficiency and accuracy to manipulation tasks.
“…Consequently, it is urgent to study the formation mechanism and release method of interfacial residual stress between metal and epoxy insulation materials commonly existing in power equipments, and understand the influencing factors of interfacial stress and strain at different stages of the curing process. In addition, the digital process of power equipments is accelerating [9]. Through the coupling calculation of multiple physical fields on the interface structure, it provides a theoretical basis for the study of digital twinning technology of power equipments.…”
There are many metal and epoxy resin interfacial structures in power equipments, which can generate interfacial stress concentrations and may induce accidents due to epoxy curing. In this paper, through simplifying interface structure into a coaxial cylindrical model, the formation mechanism of interfacial stress concentration of this interface structure was studied during epoxy resin curing process. The strain and temperature at the interface were measured by the strain and temperature measurement system. Then a model for calculating the interface stress distribution was established, including the heat conduction module, curing dynamics module and curing deformation module. The validity of the model was verified by comparing the calculated results with the measured results. In addition, the air gap defects at the interface, appearing after the curing process, were simulated. The results show that, after the curing process, the first principal stress at the centre of the interface gradually increases from the centre to the edge, reaching 41.06 and 40.20 MPa at the upper and lower edges respectively. The existence of air gap leads to partial discharge at low voltage. The calculation model can be extended to the stress prediction between the metal and epoxy resin in other power equipments.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.