引用本文: | 王 罡,刘敬文,李国鹏,等.基于多源异构数据融合的综合管廊电力舱系统保护[J].电力系统保护与控制,2021,49(7):103-109.[点击复制] |
WANG Gang,LIU Jingwen,LI Guopeng,et al.System protection of a pipe corridor power cabin based on multi-source heterogeneous data fusion[J].Power System Protection and Control,2021,49(7):103-109[点击复制] |
|
本文已被:浏览 4572次 下载 961次 |
码上扫一扫! |
基于多源异构数据融合的综合管廊电力舱系统保护 |
王罡1,刘敬文2,李国鹏2,李志雷2,沈学良2,龚志丹3,黄文林3,赵辉4,谷志成4 |
|
(1.国网河北省电力有限公司,河北 石家庄 050022;2.国网河北省电力有限公司雄安新区供电公司,河北 雄安
071800;3.厦门亿立吉奥信息科技有限公司,福建 厦门 361008;4.上海谷元电气科技有限公司,上海 201708) |
|
摘要: |
针对综合管廊电力舱运维数据的利用和分析不足、未实现基于大数据的智能运维的问题,提出了一种新的综合管廊电力舱状态分析判断方法,旨在保护综合管廊电力舱系统,提高安全性,提升运维水平。首先,将来自管廊电力舱的多个分布式数据源采用中间件技术进行数据层融合。其次,将具有相似特性的数据分配到同一子空间,并提出一种新的既能保持数据集全局结构又能保持数据集局部结构的局部全局投影方法提取特征。进而,对各个子空间提取的特征进行融合,利用支持向量数据描述方法构建分类模型。最后,在综合管廊电力舱的运维数据上进行测试以证明所提出方法的有效性以及优越性。 |
关键词: 管廊电力舱 多源异构 数据融合 状态分析判断 特征提取 |
DOI:DOI: 10.19783/j.cnki.pspc.200904 |
投稿时间:2020-07-29修订日期:2020-09-25 |
基金项目:国家自然科学基金项目资助(61703161);国家电网公司科技项目资助(KJGW2019-04-II)“城市电网综合管廊电力舱智能化运检关键技术研究”;上海市自然科学基金项目资助(19ZR1473200) |
|
System protection of a pipe corridor power cabin based on multi-source heterogeneous data fusion |
WANG Gang1,LIU Jingwen2,LI Guopeng2,LI Zhilei2,SHEN Xueliang2,GONG Zhidan3,HUANG Wenlin3,ZHAO Hui4,GU Zhicheng4 |
(1. State Grid Hebei Electric Power Co., Ltd., Shijiazhuang 050022, China; 2. Xiong’an New District Power Supply Company,
State Grid Hebei Electric Power Co., Ltd., Xiong’an 071800, China; 3. Xiamen Great Power Geo Information Technology
Co., Ltd., Xiamen 361008, China; 4. Shanghai Guyuan Electric Technology Co., Ltd., Shanghai 201708, China) |
Abstract: |
There is insufficient use and analysis of the operational and maintenance data of the pipe corridor power cabin. Also there is a failure to achieve intelligent operation and maintenance based on big data. Thus a new method for analyzing and judging the status of a pipe corridor power cabin is proposed to protect the integrated pipeline gallery power cabin system. This is intended to improve safety and enhance the level of operation and maintenance. First, the data from multiple distributed data sources is integrated using middleware technology for data layer fusion. Secondly, data with similar characteristics are allocated to the same subspace. Then, in each subspace, a novel Locality Global Projections (LGP) method that can maintain both the global structure of the dataset and the local structure of the dataset is proposed to extract features. The extracted features of each subspace are fused, and the Support Vector Data Description (SVDD) method is used to build a classification model. Finally, through testing on the operational and maintenance data of the power cabin of the integrated pipe corridor, the effectiveness and superiority over other methods are proved.
This work is supported by the National Natural Science Foundation of China (No. 61703161), the Science and Technology Project of State Grid Corporation of China (No. KJGW2019-04-II), and the Natural Science Foundation of Shanghai (No. 19ZR1473200). |
Key words: pipe corridor power cabin multi-source heterogeneous data fusion state analysis and judgment feature extraction |