Research on anomaly detection method of power grid control system based on CVE-ReS-GRU algorithm

Zhao Ruifeng; Zhong Wei; Lu Jiangang; Guo Wenxin; Lin Yueting; Dai Yue

doi:10.11959/j.issn.1000-0801.2026040

您当前的位置：

首页 >

文章列表页 >

Research on anomaly detection method of power grid control system based on CVE-ReS-GRU algorithm

Research and Development | 更新时间：2026-03-02

- Research on anomaly detection method of power grid control system based on CVE-ReS-GRU algorithm
- Telecommunications Science Vol. 42, Issue 1, Pages: 174-185(2026)
- 作者机构：
  
  1.广东电网有限责任公司电力调度控制中心，广东广州 510062
  2.国电南瑞南京控制系统有限公司，江苏南京 211106
- 作者简介：
- 基金信息：
  
  Science and Technology Project of China Southern Power Grid Company(036000KC23110012)
- DOI：10.11959/j.issn.1000-0801.2026040
  CLC： TM73
- Received：09 September 2025，
  
  Revised：2025-11-03，
  
  Accepted：18 November 2025，
  
  Published：20 January 2026
- 稿件说明：
移动端阅览
赵瑞锋,仲卫,卢建刚等.基于CVAE-ReS-GRU算法的电网调控系统异常检测方法研究[J].电信科学,2026,42(01):174-185.

Zhao Ruifeng,Zhong Wei,Lu Jiangang,et al.Research on anomaly detection method of power grid control system based on CVE-ReS-GRU algorithm[J].Telecommunications Science,2026,42(01):174-185.
赵瑞锋,仲卫,卢建刚等.基于CVAE-ReS-GRU算法的电网调控系统异常检测方法研究[J].电信科学,2026,42(01):174-185. DOI： 10.11959/j.issn.1000-0801.2026040.

Zhao Ruifeng,Zhong Wei,Lu Jiangang,et al.Research on anomaly detection method of power grid control system based on CVE-ReS-GRU algorithm[J].Telecommunications Science,2026,42(01):174-185. DOI： 10.11959/j.issn.1000-0801.2026040.

摘要

针对现有电网调控系统异常检测方法在复杂时序数据处理及特征提取方面的局限，提出了一种基于条件变分自编码器-残差结构-门控循环单元（conditional variational autoencoder-residual structure-gated recurrent unit，CVAE-ReS-GRU）的电网调控系统异常检测方法。首先，基于电网调控系统的层级化运行架构，对系统内多源异构数据进行分类与结构化整理。其次，采用支持向量机（support vector machine，SVM）对缺失数据进行智能填补，并通过归一化消除原始数据的量纲差异。然后，引入时间监督机制以增强VAE对电网时序特征的建模能力；将残差结构嵌入GRU网络，以提升对复杂电力序列的拟合精度；以改进后的GRU网络代替CVAE中传统的BP神经网络结构，从而加速特征提取。最后，设计动态阈值判定机制，自适应调整阈值评估重构误差，实现对异常数据的精准识别。仿真实验表明，所提方法在异常数据识别准确率、检测效率等指标上均展现出显著优势，为提升电网调度智能化水平、保障电力系统安全稳定运行提供了新的技术路径与理论支撑。

Abstract

In view of the limitations of existing anomaly detection methods for power grid regulation systems in handling complex time-series data and feature extraction

an anomaly detection method for power grid regulation systems based on conditional variational autoencoder-residual structure-gated recurrent unit (CVAE-ReS-GRU) was proposed. Firstly

based on the hierarchical operation architecture of the power grid regulation system

systematic classification and structured organization of multi-source heterogeneous data within the system were carried out. Secondly

the support vector machine (SVM) was employed to intelligently fill in missing data

and normalization was used to eliminate the dimensional differences of the original data. Then

a time supervision mechanism was introduced to improve the variational autoencoder (VAE)

enhancing its ability to model the time-series characteristics of the power grid. The residual structure was embedded into the GRU network to improve the fitting accuracy for complex power time-series. The improved GRU network replaces the traditional BP neural network structure in CVAE

accelerating feature extraction. Finally

a dynamic threshold determination mechanism was designed to adaptively adjust the threshold for evaluating reconstruction errors

enabling accurate identification of abnormal data. Simulation results demonstrate that the proposed method exhibits significant advantages in terms of abnormal data recognition accuracy and detection efficiency compared with traditional methods

providing a new technical approach and theoretical support for enhancing the intelligent level of power grid dispatching and ensuring the safe and stable operation of the power system.

关键词

Keywords

references

Ourahou M , Ayrir W , El Hassouni B , et al . Review on smart grid control and reliability in presence of renewable energies: Challenges and prospects [J ] . Mathematics and Computers in Simulation , 2020 , 167 : 19 - 31 .

谈林涛 , 李军良 , 任昺 , 等 . 基于RB-XGBoost算法的智能电网调度控制系统健康度评价模型 [J ] . 电力自动化设备 , 2020 , 40 ( 2 ): 189 - 198 .

Tan L T , Li J L , Ren B , et al . Health evaluation model of smart grid dispatch and control system based on RB-XGBoost algorithm [J ] . Electric Power Automation Equipment , 2020 , 40 ( 2 ): 189 - 198 .

常荣 , 徐敏 . 基于改进K-Means和DNN算法的电力数据异常检测 [J ] . 南京理工大学学报 , 2023 , 47 ( 6 ): 790 - 796, 858 .

Chang R , Xu M . Power data anomaly detection based on improved K-Means and DNN algorithm [J ] . Journal of Nanjing University of Science and Technology , 2023 , 47 ( 6 ): 790 - 796, 858 .

郭泽阳 , 程拥军 , 温家铭 , 等 . 基于支持向量机的光伏电站异常数据序列提取 [J ] . 中国测试 , 2023 , 49 ( 7 ): 156 - 161 .

Guo Z Y , Cheng Y J , Wen J M , et al . Extraction of abnormal data sequence of photovoltaic power station based on support vector machine [J ] . China Measurement & Test , 2023 , 49 ( 7 ): 156 - 161 .

杨海柱 , 田馥铭 , 张鹏 , 等 . 基于CEEMD-FE和AOA-LSSVM的短期电力负荷预测 [J ] . 电力系统保护与控制 , 2022 , 50 ( 13 ): 126 - 133 .

Yang H Z , Tian F M , Zhang P , et al . Short-term load forecasting based on CEEMD-FE-AOA-LSSVM [J ] . Power System Protection and Control , 2022 , 50 ( 13 ): 126 - 133 .

陈逸枞 , 张大海 , 李宇欣 , 等 . 基于DWT-MOSMA-SVM的多目标优化短期母线负荷预测 [J ] . 电力建设 , 2023 , 44 ( 3 ): 49 - 55 .

Chen Y C , Zhang D H , Li Y X , et al . Multi-objective optimization based on DWT-MOSMA-SVM for short-term bus load forecasting [J ] . Electric Power Construction , 2023 , 44 ( 3 ): 49 - 55 .

张鑫 , 孙国繁 , 高磊 , 等 . 基于机器学习的变电站故障预测方法研究 [J ] . 电子器件 , 2024 , 47 ( 4 ): 1047 - 1052 .

Zhang X , Sun G F , Gao L , et al . The fault prediction of converting station based on big data technology and machine learning [J ] . Chinese Journal of Electron Devices , 2024 , 47 ( 4 ): 1047 - 1052 .

闫梦秋 , 杨轶俊 , 赵舫 . 基于改进OCSVM的智能变电站数据流异常检测方法研究 [J ] . 电力系统保护与控制 , 2022 , 50 ( 6 ): 100 - 106 .

Yan M Q , Yang Y J , Zhao F . A data stream anomaly detection method based on an improved OCSVM smart substation [J ] . Power System Protection and Control , 2022 , 50 ( 6 ): 100 - 106 .

马智强 , 康洁滢 , 梁飞 , 等 . 基于XGBoost的电网过流异常检测模型 [J ] . 电子测量技术 , 2024 , 47 ( 21 ): 130 - 137 .

Ma Z Q , Kang J Y , Liang F , et al . XGBoost-based power grid overcurrent anomaly detection model [J ] . Electronic Measurement Technology , 2024 , 47 ( 21 ): 130 - 137 .

田江 , 吕洋 , 赵奇 , 等 . 基于粗糙熵加权密度的电网调控系统异常检测 [J ] . 兵工自动化 , 2024 , 43 ( 2 ): 40 - 44, 57 .

Tian J , Lyu Y , Zhao Q , et al . Anomaly detection of power grid control system based on weighted density of rough entropy [J ] . Ordnance Industry Automation , 2024 , 43 ( 2 ): 40 - 44, 57 .

王瑾 , 裴亮 . 基于深度学习的电网调控系统异常检测与多阶段风险预警 [J ] . 沈阳工业大学学报 , 2021 , 43 ( 6 ): 601 - 607 .

Wang J , Pei L . Anomaly detection and multi-stage risk pre-warning technology of power grid control system based on deep learning [J ] . Journal of Shenyang University of Technology , 2021 , 43 ( 6 ): 601 - 607 .

荆志朋 , 柴林杰 , 胡诗尧 . 基于改进LSTM-VAE的配电网异常负荷检测方法研究 [J ] . 电测与仪表 , 2024 , 61 ( 9 ): 71 - 76 .

Jing Z P , Chai L J , Hu S Y . Research on abnormal load detection method for distribution network based on improved LSTM-VAE [J ] . Electrical Measurement & Instrumentation , 2024 , 61 ( 9 ): 71 - 76 .

李丹 , 罗娇娇 , 孙光帆 , 等 . 基于条件变分自编码器和贝叶斯神经网络的短期电力负荷概率预测 [J ] . 三峡大学学报(自然科学版) , 2024 , 46 ( 5 ): 68 - 75 .

Li D , Luo J J , Sun G F , et al . Probabilistic short-term load forecasting based on conditional variational autoencoder and Bayesian neural network [J ] . Journal of China Three Gorges University (Natural Sciences) , 2024 , 46 ( 5 ): 68 - 75 .

陈冰 , 唐永旺 . 基于Bi-GRU和自注意力的智能电网虚假数据注入攻击检测 [J ] . 计算机应用与软件 , 2021 , 38 ( 7 ): 339 - 344, 349 .

Chen B , Tang Y W . False data injection attacks detection in smart grid based on bi-gru and self-attention [J ] . Computer Applications and Software , 2021 , 38 ( 7 ): 339 - 344, 349 .

夏景 , 梁薇 , 吴珠瑛 . 基于移动小波树的电力监控异常数据自动识别算法研究 [J ] . 电子设计工程 , 2020 , 28 ( 18 ): 148 - 152 .

Xia J , Liang W , Wu Z Y . Research on automatic recognition algorithm of abnormal data in power monitoring based on mobile wavelet tree [J ] . Electronic Design Engineering , 2020 , 28 ( 18 ): 148 - 152 .

方华强 , 颜寒祺 , 陈波 , 等 . 基于自编码网络的移动轨迹异常检测 [J ] . 地理信息世界 , 2019 , 26 ( 5 ): 41 - 44, 52 .

Fang H Q , Yan H Q , Chen B , et al . Anomaly detection in mobile trajectory using auto-encoder network [J ] . Geomatics World , 2019 , 26 ( 5 ): 41 - 44, 52 .

胡昌斌 , 张亚 , 李迎丽 , 等 . 基于朴素贝叶斯的电网用户行为分析 [J ] . 沈阳工业大学学报 , 2020 , 42 ( 3 ): 259 - 263 .

Hu C B , Zhang Y , Li Y L , et al . Analysis of electrical grid user behavior based on Naive Bayesian [J ] . Journal of Shenyang University of Technology , 2020 , 42 ( 3 ): 259 - 263 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research on log text fault association rule mining based on GPT network

Research on anomaly detection algorithm based on sparse variational autoencoder using spike and slab prior

Research progress of KPI anomaly detection in intelligent operation and maintenance

Anomaly detection algorithm based on Gaussian mixture variational auto encoder network

The Strengthening Mechanism of Cu Bearing High Strength Steel As-Quenched and Tempered and Cu Precipitation Behavior in Steel

Related Author

LIN Yuting

WANG Xiaodong

HE Zhenqin

WANG Yuxi

YE Qingwei

LI Bing

Zhe CHEN

Huahua CHEN

Related Institution

Faculty of Electrical Engineering and Computer Science, Ningbo University

State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications

Guangdong Communications ＆ Networks Institute

Purple Mountain Laboratories

School of Information and Electrical Engineering,Zhejiang Gongshang University

⁰