基于数据挖掘的配电网线路不停电作业能力定级评估Research on Grading Evaluation of Uninterrupted Operation Lines in Distribution Network based on data Mining
孙伟,赵江宁,徐昭麟,王华昕,麻宁杰
SUN Wei,ZHAO Jiangning,XU Zhaolin,WANG Huaxin,MA Ningjie
摘要(Abstract)：
为适应配电网由停电检修向不停电检修的转变,亟需制定一套科学合理的配电线路不停电作业综合评价指标体系,并且对其进行定量等级评价指标分析。从配电网可靠性和不停电作业难度出发。首先,基于用户不停电和检修不停电两大目标,提出不停电作业综合评估指标体系;其次,依托国网浙江供服系统中台读取的海量数据建立了不停电作业指标数据库,采用极端梯度提升算法结合序列前向搜索策略从数据库中筛选岀关键评估指标,达到数据降维的目的;最后,使用多因素Logistic回归方法构建不停电作业等级评估模型,完成对线路不停电作业能力的评估。
In order to accommodate the transition from power maintenance to non-power maintenance in the distribution network, it is urgent to develop a scientific and reasonable comprehensive evaluation index system for non-power maintenance of distribution network lines, and conduct quantitative grade evaluation index analysis. First of all, the article starts from the reliability of distribution network and the difficulty of nonpower maintenance, based on the two goals of non-power for user and non-power maintenance, proposes the comprehensive evaluation index system of non-power maintenance; Secondly, Based on the massive data read by data middle platform of National Grid in Zhejiang Province electric power company Service System, it establishes the database of non-power maintenance index and uses XGBoost combined with sequence forward search strategy to screen the key evaluation indexes from the database to chieve the purpose of data dimension reduction; Finally, it uses the multi-factor Logistic regression method to construct the evaluation model of non-power maintenance level, which completes the evaluation of the operation ability of the Uninterrupted Operation Lines. The paper takes the actual situation of the distribution network line in Zhejiang area as an example to carry out on-site evaluation.
关键词(KeyWords)：
配电网;不停电作业;极端梯度提升算法;评估指标体系
distribution network;non-power maintenance;XGBoost;evaluation index system
基金项目(Foundation): 国网浙江省电力有限公司科技项目（5211SX1900PS）
作者(Author):
孙伟,赵江宁,徐昭麟,王华昕,麻宁杰
SUN Wei,ZHAO Jiangning,XU Zhaolin,WANG Huaxin,MA Ningjie
DOI: 10.19585/j.zjdl.202109013
参考文献(References)：
- [1]王爱国，刘瑞庆.浅谈10 kV配网不停电作业技术发展与应用[J].电子测试，2017(7):103-104.
- [2]穆永铮，鲁宗相，乔颖，等.基于多算子层次分析模糊评价的电网安全与效益综合评价指标体系[J].电网技术，2015,39(1):23-28.
- [3]李晨，殷自力，王晓辉，等.基于层次分析法和熵权法的配电网调度评价[J].电力系统及其自动化学报，2019,31(7):81-87.
- [4]刘根宝，周莹珣.10 kV配电线路带电作业安全综合评价应用研究[J].通讯世界，2016(16):140-141.
- [5]辛洪杰.10 kV配电线路带电作业安全问题综合评价[J].山东工业技术，2016(6):161.
- [6]陈碧云，陈千懿，周恒旺.电力系统输配电可靠性的快速协同评估方法[J].电力系统及其自动化学报，2018,30(11):101-106.
- [7]孙明，苏毅方，邵先军，等.基于数据挖掘的配电网线路台风故障影响分级评价[J].浙江电力，2019,38(10):78-83.
- [8]窦孝祥，郝涛，王云龙，等.配网不停电作业在区域应急供电中的应用[J].农村电气化，2021(4):12-14.
- [9]董雷，李佳.考虑时变因素的配电系统可靠性分析[J].电力系统保护与控制，2009,37(12):6-10.
- [10]何迎春.基于故障树的架空输电线路可靠性评估方法研究[D].重庆：重庆大学，2012.
- [11]姚李孝，彭金宁，伍利，等.复杂配电系统的可靠性评估[J].西安理工大学学报，2004,20(1):44-48.
- [12]祁彦鹏，张焰，余建平，等.配电系统可靠性分析的最小割集-网络等值法[J].电力系统及其自动化学报，2011,23(2):98-104.
- [13]胡栋鹏，曾坚毅.数据挖掘常见分类算法比较分析[J].计算机产品与流通，2018(8):107.
- [14]CHEN T,GUESTRIN C.XGBoost:a scalable tree boosting system[C]//ACM SIGKDD International Conference on Know ledge Discovery and Data Mining.ACM,2016:785-794.
- [15]周志华.机器学习[M].北京：清华大学出版社，2016.
- [16]FRIEDMAN J,HASTIE T,TIBSHIRANI R.Special invited paper. additive logistic regression:a statistical view of boosting[J].The Annals of Statistics,2000,28(2):337-374.
- [17]颜磊，祁冰.基于Android平台的移动学习系统大数据挖掘技术研究[J].现代电子技术，2017,40(19):142-144.
- [18]李哲瑜，简宋全，李青海.基于Logistic回归的个人消费贷款预测[J].现代计算机，2017(29):9-12.
- [19]潘舒笑.创伤失血性休克时间窗预测及风险评分方法研究[D].北京：北京交通大学，2020.
- [20]成亮，孙茜.基于大数据的配电网风险预警管控[J].山西电力，2020(3):39-42.