###
DOI:
电力大数据:2020,23(5):-
←前一篇   |   后一篇→
本文二维码信息
大数据技术在配网单线图自动成图的应用研究
何雄坤1, 周宏志1, 聂辉1, 陈满超1, 齐志刚2
(1.贵州电网有限责任公司都匀荔波供电局;2.杭州昊美科技有限公司)
Application research of big data technology in automatic mapping of single-line diagrams of distribution network
Xiongkun He1, Hongzhi Zhou1, Hui Nie1, Manchao chen1, Zhigang qi2
(1.Guizhou Power Grid Corporation,Guizhou Qiannan;2.hangzhouhaomei Limited company,Zhejiang Hangzhou)
摘要
图/表
参考文献
相似文献
本文已被:浏览 557次   下载 13
投稿时间:2020-01-13    修订日期:2020-06-02
中文摘要: 为解决配网单线图自动成图后,需要大量人工调整才能达到实用要求问题。本文基于贵州某配网自动成图系统,引入大数据Spark框架的机器学习技术到单线图布局算法,并进行了应用研究。首先,介绍自动成图系统的技术架构,阐述机器学习技术的应用环境,提出构建基于迭代二叉树3代(ID3)算法的决策树机器学习模型,采用配电网数据为训练样本,描述决策树模型的构建过程。进一步的,以图形调整样本数据为训练集,训练布局算法库中的子设备相对母设备的布局角度算法。最后,用配电网的单线图数据实例,验证机器学习模型及布局算法。结果显示,经机器学习后,生成的单线图走线横平竖直,站房分层均衡,布局效果合理美观。证明大数据和机器学习技术在单线图自动成图布局中应用的有效性。
中文关键词: 单线图  大数据  布局算法  机器学习  决策树  
Abstract:In order to overcome the difficulty that after the single-line diagram of the distribution network is automatically drawn, a tremendous amount of manual adjustment is required to satisfy the practical standards. This article focused on an automatic mapping system platform of a distribution network in Guizhou, introduced via machine learning technology of The Big Data Spark framework to single-line graph layout algorithm, with emphasis on its application usage. First, it introduced the technical architecture of the automatic mapping system and elaborated the application environment of machine learning technology, proposed to build a decision tree machine learning model based on iterative dichotomiser 3 generation(ID3)algorithm, using the distribution network data as training samples, describing the construction process of the decision tree model. Further, the sample data adjusted by graphics is used as the training set, to train the child devices relative to the parent device in the differential layout algorithm. Finally, the single-line diagram data examples of the distribution network are used to verify the machine learning model and the difference layout algorithm. The results showed that after using the algorithm, the generated single-line diagram is horizontal and vertical, the station building is layered and balanced, thus the layout is fine and legit. This proves the effectiveness of the application of big data and machine learning techniques in the automatic layout of single-line diagrams.
文章编号:     中图分类号:    文献标志码:
基金项目:
引用文本: