|本期目录/Table of Contents|

[1]危 欢,栗 娟*.基于联邦分割学习的输电线路异物检测算法[J].武汉工程大学学报,2024,46(04):439-445.[doi:10.19843/j.cnki.CN42-1779/TQ.202312023]
 WEI Huan,LI Juan*.Transmission line foreign object detection algorithm based onfederated split learning[J].Journal of Wuhan Institute of Technology,2024,46(04):439-445.[doi:10.19843/j.cnki.CN42-1779/TQ.202312023]
点击复制

基于联邦分割学习的输电线路异物检测算法(/HTML)
分享到:

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
46
期数:
2024年04期
页码:
439-445
栏目:
机电与信息工程
出版日期:
2024-08-28

文章信息/Info

Title:
Transmission line foreign object detection algorithm based on
federated split learning
文章编号:
1674 - 2869(2024)04 - 0439 - 07
作者:
危 欢1栗 娟*2
1. 武汉工程大学计算机科学与工程学院,湖北 武汉 430205;
2. 智能机器人湖北省重点实验室(武汉工程大学),湖北 武汉 430205
Author(s):
WEI Huan1LI Juan*2
1. School of Computer Science and Engineering,Wuhan Institute of Technology,Wuhan 430205,China;
2. Hubei Key Laboratory of Intelligent Robot(Wuhan Institute of Technology),Wuhan 430205,China
关键词:
边缘计算联邦学习线路检测分割学习
Keywords:
edge computing federated learning grid detection split learning
分类号:
TP393
DOI:
10.19843/j.cnki.CN42-1779/TQ.202312023
文献标志码:
A
摘要:
异物入侵是导致输电线路故障的主要原因之一,但现有输电线路异物检测方法未能充分利用终端设备的计算能力,造成资源浪费与隐私数据泄露等问题。针对上述问题,提出了一种基于联邦分割学习的检测算法(FSLDA)。该算法结合联邦学习和分割学习,提高输电线路异物检测效率和终端设备数据安全性。FSLDA通过构建可分割的小型神经网络,将计算负载分散至不同设备,减轻终端设备的运算压力,同时确保训练数据的隐私安全得到有效保障。实验结果表明:与经典联邦学习相比,FSLDA在保持预测精度的同时,终端设备的训练时间和能源消耗分别减少了10%和20%。由此可知,FSLDA在提升输电线路异物检测效率和可靠性方面具备有效性,并有助于优化系统总体性能和保障终端数据隐私安全。
Abstract:
Foreign object intrusion is one of the primary causes of power transmission line failures. However, existing research on power transmission line foreign object detection has not fully utilized the computational capabilities of terminal devices, leading to issues such as resource waste and privacy breaches. In response to these problems, in this paper we proposed a federated split learning detection algorithm (FSLDA). This model integrates federated learning and split learning to enhance the efficiency and data security of foreign object detection systems. The FSLDA, by developing a divisible small-scale neural network, distributes the computational workload across different devices, thereby reducing the computing pressure on devices and ensuring the privacy security of training data is effectively guaranteed. Experimental results demonstrate that, compared to classic federated learning, FSLDA reduces the training time and the energy consumption by 10% and 20%, respectively, while maintaining the prediction accuracy. Thus, FSLDA is effective in enhancing the efficiency and reliability of power transmission line foreign object detection, contributing to the optimization of overall system performance and the safeguarding of data privacy.

参考文献/References:

[1] MADABHUSHI S, DEWRI R. A survey of anomaly detection methods for power grids[J]. International Journal of Information Security,2023,22(6): 1799-1832.

[2] LI J L, GU C H, XIANG Y, et al. Edge-cloud computing systems for smart grid: state-of-the-art,architecture,and applications[J]. Journal of Modern Power Systems and Clean Energy,2022,10(4): 805-817.
[3] ZHOU Z Y, ZHANG C T, XU C, et al. Energy-efficient industrial internet of UAVs for power line inspection in smart grid[J]. IEEE Transactions on Industrial Informatics,2018,14(6): 2705-2714.
[4] VOULODIMOS A, DOULAMIS N, DOULAMIS A,et al. Deep learning for computer vision: a brief review[J]. Computational Intelligence and Neuroscience,2018,2018: 7068349.
[5] LIANG F T, ZHOU Y, CHEN X, et al. Review of target detection technology based on deep learning[C]//Proceedings of the 5th International Conference on Control Engineering and Artificial Intelligence. New York: Association for Computing Machinery,2021: 132-135.
[6] YU C H, LIU Y K, ZHANG W R, et al. Foreign objects identification of transmission line based on improved YOLOv7[J]. IEEE Access,2023,11: 51997-52008.
[7] LIU Z Y,WU G P, HE W S, et al. Key target and defect detection of high-voltage power transmission lines with deep learning[J]. International Journal of Electrical Power & Energy Systems,2022,142(Part A): 108277.
[8] 刘凌志,栗娟,秦志威. 基于化学反应优化算法的边缘计算任务卸载策略[J]. 武汉工程大学学报,2023,45(4):435-441.
[9] CHEN S L,WEN H,WU J S,et al. Internet of things based smart grids supported by intelligent edge computing[J]. IEEE Access,2019,7: 74089-74102.
[10] XU D L, LI T, LI Y, et al. Edge intelligence: empowering intelligence to the edge of network[J]. Proceedings of the IEEE,2021,109(11):1778-1837.
[11] SAMIKWA E,MAIO A D,BRAUN T. Ares: adaptive resource-aware split learning for internet of things [J]. Computer Networks,2022,218: 109380.
[12] 康宇洋,刘为凯. 批量归一化的自适应联邦学习算法[J]. 武汉工程大学学报,2023,45(5):549-555.
[13] VEPAKOMMA P, GUPTA O, SWEDISH T, et al. Split learning for health: distributed deep learning without sharing raw patient data[OL].(2018-12-3)[2023-12-18]. https://doi.org/10.48550/arXiv.1812. 00564.
[14] MCMAHAN H B, MOORE E, RAMAGE D,et al. Communication-efficient learning of deep networks from decentralized data[OL].(2023-1-16)[2023-12-18]. https://doi.org/10.48550/arXiv.1602.05629.
[15] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[OL]. (2015-04-10) [2023-12-18]. https://doi.org/10.48550/arXiv.1409.1556.
[16] CHEN J B,XUE J F,WANG Y,et al. Privacy-preserving and traceable federated learning for data sharing in industrial IoT applications[J]. Expert Systems with Applications,2023,213: 119036.
[17] 王昊天,郑栋毅,刘芳,等. 面向多元时序数据的个性化联邦异常检测方法[J]. 计算机工程与应用,2022,58(11):60-65.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期:2023-12-18
基金项目:国家自然科学基金(62102292);湖北省智能机器人重点实验室开放基金(HBIRL 202204);武汉市知识创新专项曙光项目(2023010201020440)
作者简介:危 欢,硕士研究生。Email:22107010104@stu.wit.edu.cn
*通信作者:栗 娟,博士,副教授。 Email:juanli2018@wit.edu.cn
引文格式:危欢,栗娟. 基于联邦分割学习的输电线路异物检测算法[J]. 武汉工程大学学报,2024,46(4):439-445.
更新日期/Last Update: 2024-08-31