半月刊

ISSN 1000-1026

CN 32-1180/TP

+高级检索 English
基于生成对抗和双重语义感知的配电网量测数据缺失重构
作者:
作者单位:

1.华北电力大学控制与计算机工程学院;2.华北电力大学电气与电子工程学院

作者简介:

通讯作者:

基金项目:

国家自然科学基金资助项目(51277069); 国家电网公司科技项目(52094018001C)。


Missing Reconstruction of Measurement Data Based on Generative Adversarial Network and Double Semantic Perception in Distribution Network
Author:
Affiliation:

1.School of Control and Computer Engineering, North China Electric Power University, Beijing 102206, China;2.School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
    摘要:

    传统的数据缺失重构技术大多依赖数理统计方法和先验知识结合机理分析构建数学模型,但是配电网量测数据具有高维、时变、非线性特征,复杂度高、表征难度大,难以保证高精度重构。文中提出一种利用无监督生成对抗训练方式自主提取数据特征并结合双重语义感知重构约束实现数据缺失重构的方法。其中,基于二维卷积的重构模型和量测数据二维灰度图像化训练增强了模型泛化能力和稳定性。该方法无需先验知识的分布假设与显式物理建模,在保证数据特征提取最大化的同时,有效提高了重构数据的精确性。最后,利用实测数据验证了该方法在重构缺失数据上的有效性。

    Abstract:

    Traditional data missing reconstruction technology mostly relies on mathematical statistics method and prior knowledge combined with mechanism analysis to construct mathematical models. However, measurement data in distribution network has high dimensional, time-varying, non-linear characteristics, high complexity, difficult characterization, and it is difficult to ensure high-precision reconstruction. In this paper, an unsupervised generation antagonism training method is proposed to extract data features independently and reconstruct missing data with dual semantic perception constraints. Among them, the reconstructed model based on two-dimensional convolution and the two-dimensional gray image training of measurement data enhance the generalization ability and stability of the model. This method does not need prior knowledge distribution hypothesis and explicit physical modeling, and can effectively improve the accuracy of reconstructed data while guaranteeing maximum feature extraction. Finally, the validity of this method in reconstructing missing data is verified by the measured data. National Natural Science Foundation of China (No. 51277069) and State Grid Corporation of China (No. 52094018001C).

    表 3 Table 3
    表 1 缺失数据重构误差分析Table 1 Error analysis of missing data reconstruction
    图1 基于GAN和双重语义感知的缺失重构网络结构Fig.1 Missing reconstructing network structure based on GAN and double semantic perception
    图2 基于GAN的特征提取模型架构图Fig.2 Architecture diagram of GAN based features extraction model
    图3 基于双重语义感知的重构过程图Fig.3 Reconstruction process map based on double semantic perception
    图4 缺失数据重构过程Fig.4 Reconstruction process of missing data
    图 二维变换Fig. two-dimensional transformation
    图 缺失重构总损失Fig. Missing data reconstruction loss
    表 2 Table 2
    图 训练损失值变化图Fig. Change Chart of Training Loss Value
    图 掩码遮盖效果图Fig. Mask Cover Chart
    参考文献
    相似文献
    引证文献
引用本文

杨玉莲,齐林海,王红,等.基于生成对抗和双重语义感知的配电网量测数据缺失重构[J/OL].电力系统自动化,http://doi.org/10.7500/AEPS20190605007.
YANG Yulian,QI Linhai,WANG Hong,et al.Missing Reconstruction of Measurement Data Based on Generative Adversarial Network and Double Semantic Perception in Distribution Network[J/OL].Automation of Electric Power Systems,http://doi.org/10.7500/AEPS20190605007.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2019-06-05
  • 最后修改日期:2020-02-14
  • 录用日期:2019-10-15
  • 在线发布日期:
  • 出版日期: