半月刊

ISSN 1000-1026

CN 32-1180/TP

+高级检索 English
基于集成深度神经网络的配电网联络关系辨识技术
作者:
作者单位:

1.东南大学电气工程学院,江苏省南京市 210096;2.国网江苏省电力有限公司泰州供电分公司,江苏省泰州市 225300;3.国网江苏省电力有限公司电力科学研究院,江苏省南京市 211103

作者简介:

蒋玮(1982—),男,博士,副教授,主要研究方向:大数据驱动的配电网规划评估、储能技术在电力系统中的应用。E-mail:jiangwei@seu.edu.cn
汤海波(1995—),男,通信作者,硕士研究生,主要研究方向:人工智能算法在坚强智能电网参数辨识中的应用。E-mail:15651921890@163.com
祁晖(1981—),男,高级工程师,主要研究方向:基于大数据平台的配电网规划信息化。E-mail:qihui@js.sgcc.com.cn

通讯作者:

基金项目:

国家自然科学基金资助项目(51877041);国网江苏省电力有限公司科技项目(J2018088)。


Distribution Network Connectivity Recognition Based on Ensemble Deep Neural Network
Author:
Affiliation:

1.School of Electrical Engineering, Southeast University, Nanjing 210096, China;2.Taizhou Power Supply Company, State Grid Jiangsu Electric Power Co., Ltd., Taizhou 225300, China;3.State Grid Jiangsu Electric Power Co., Ltd. Research Institute, Nanjing 211103, China

Fund Project:

This work is supported by National Natural Science Foundation of China (No. 51877041) and State Grid Jiangsu Electric Power Co., Ltd. (No. J2018088).

  • 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:

    随着城市配电网络规模不断扩大,配电网实时拓扑难以获取已成为观测其运行状态的主要瓶颈。为了解决传统拓扑辨识方法噪声敏感性高、在线运行难等问题,提出了一套基于集成深度神经网络的配电变压器(简称配变)联络关系辨识方案。首先,依据配电网测量的横纵连续性,对历史数据进行二维小波阈值去噪,降低噪声对辨识结果的影响。为提高深度学习算法的精度上限,采用搜索、生成与评价的策略对数据进行特征提取与选择。然后,以选择的特征为输入,构造交叉熵深度神经网络,通过网格搜索优化深度神经网络的超参数。采用集成学习的策略训练同质深度神经网络,保证模型的在线拓扑辨识能力。最后,通过在TensorFlow上进行的实验验证了提出的集成深度神经网络模型在配变联络关系辨识中的精确度与鲁棒性。

    Abstract:

    With the development of the urban distribution network, the difficulty in obtaining the real-time distribution network topology has become an important bottleneck for observing its operational status. To tackle high noise sensitivity and difficult online operation in traditional topology recognition methods, this paper proposes a distribution transformers (DTs) connectivity recognition method based on ensemble deep neural network. Firstly, based on the horizontal and vertical continuity of distribution network measurements, the two-dimension wavelet threshold de-noising method is implemented on historical measurements to reduce the impact of noise on the recognition results. To improve the upper limit of deep learning algorithm, this work proposes the search, generation and evaluation strategy for feature extraction and feature selection. Secondly, we use the extracted features as inputs for the deep neural network with cross-entropy function. The grid search algorithm is used to optimize the hyperparameters of deep neural network. An ensemble learning strategy is utilized to train homogeneous deep neural networks to ensure the ability of online topology recognition. Finally, the experiments implemented on TensorFlow certificate the accuracy and robustness of proposed ensemble deep neural network models in distribution transformers connectivity recognition.

    表 1 MLP模型的初始特征Table 1 Initial features of MLP model
    图1 双电源环形配电网等效拓扑Fig.1 Equivalent topology of ring distribution network with dual power
    图2 10 kV环状配电网简化拓扑模型Fig.2 Simplified topology model of 10 kV ring distribution network
    图3 不同算法在不同样本规模下的平均计算时间Fig.3 Average calculation time for different algorithms with different sample sizes
    图4 不同学习策略在测试样本扩大时的计算时间Fig.4 Calculation time with different learning strategies when test sample is expanded
    表 3 不同算法的配变互联关系辨识Table 3 Identification of interconnection relationship of distribution transformers with different algorithms
    表 4 不同学习策略的深度神经网络模型精确度对比Table 4 Accuracy comparison of deep neural network models with different learning strategies
    表 2 去噪前后精确度对比Table 2 Accuracy comparison before andafter denoising
    参考文献
    相似文献
    引证文献
引用本文

蒋玮,汤海波,祁晖,等.基于集成深度神经网络的配电网联络关系辨识技术[J].电力系统自动化,2020,44(1):101-108. DOI:10.7500/AEPS20190411010.
JIANG Wei,TANG Haibo,QI Hui,et al.Distribution Network Connectivity Recognition Based on Ensemble Deep Neural Network[J].Automation of Electric Power Systems,2020,44(1):101-108. DOI:10.7500/AEPS20190411010.

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:2019-04-11
  • 最后修改日期:2019-07-11
  • 录用日期:
  • 在线发布日期: 2020-01-04
  • 出版日期: