基于多模型融合的电网故障抢修时长预测

doi:10.11959/j.issn.1000-0801.2020016

电信科学 ›› 2020, Vol. 36 ›› Issue (1): 144-150.doi: 10.11959/j.issn.1000-0801.2020016

• 电力信息化专栏 • 上一篇

基于多模型融合的电网故障抢修时长预测

潘坚跃¹,吴懿臻²,徐汉麟¹

¹ 国网浙江省电力有限公司杭州供电公司，浙江杭州 310009
² 浙江华云信息科技有限公司，浙江杭州 310008

修回日期:2020-01-05 出版日期:2020-01-20 发布日期:2020-02-13
作者简介:潘坚跃（1961- ），男，国网浙江省电力有限公司杭州供电公司高级工程师，长期从事数据集成、大数据应用分析方面的工作|吴懿臻（1984- ），女，浙江华云信息科技有限公司助理工程师，长期从事数据应用分析、数据挖掘方面的工作|徐汉麟（1990- ），男，国网浙江省电力有限公司杭州供电公司助理工程师，主要研究方向为信息系统及网络安全

Prediction of power grid fault repair time based on multi-model fusion

Jianyue PAN¹,Yizhen WU²,Hanlin XU¹

¹ State Grid Zhejiang Electric Power Company Hangzhou Power Supply Company,Hangzhou 310009,China
² Zhejiang Huayun Information Technology Co.,Ltd.,Hangzhou 310008,China

Revised:2020-01-05 Online:2020-01-20 Published:2020-02-13

摘要/Abstract

摘要：

电网故障种类繁多、原因复杂，故障抢修时长的预测相对较难。由于深度学习等新技术的兴起，从故障工单中挖掘有效的信息，进而准确地预测故障抢修时长的方法正在变得可行。以历史电网故障抢修工单为研究对象，提出多模型融合的预测方法，将LightGBM、XGBoost和长短期记忆网络的预测结果进行加权融合。实验结果表明，该多模型融合的预测方法可以较为准确地对故障抢修时长进行预估，为电网故障抢修的自动化和智能化提供更好的支撑。

关键词: 电网故障, 抢修时长, LSTM, LightGBM, XGBoost

Abstract:

There are many types of power grid faults,and the reasons are complicated.The prediction of fault repair time is difficult.Due to the rise of new technologies such as deep learning,it is feasible to accurately mine the faulty worksheet and accurately predict the fault repair time.Taking the historical grid fault repair worksheet as the research object,the multi-model fusion prediction method was proposed,and the prediction results of LightGBM,XGBoost and LSTM were weighted and fused.The experimental results show that the multi-model fusion prediction method can accurately estimate the fault repair time and provide better support for the automation and intelligence of grid fault repair.

Key words: grid fault, repair time, LSTM, LightGBM, XGBoost

中图分类号:

TN929.5

潘坚跃,吴懿臻,徐汉麟. 基于多模型融合的电网故障抢修时长预测[J]. 电信科学, 2020, 36(1): 144-150.

Jianyue PAN,Yizhen WU,Hanlin XU. Prediction of power grid fault repair time based on multi-model fusion[J]. Telecommunications Science, 2020, 36(1): 144-150.

图/表 10

图1

图2

图3

图4

图5

表1

图6

图7

图8

表2

参考文献 12

[1]	喻圣, 邹红波, 余凡 ,等. 基于多种智能方法结合的电网故障诊断综述[J]. 电力学报, 2019,34(1): 35-41.
	YU S , ZOU H B , YU F . Review on fault diagnosis based on a combination of intelligent methods[J]. Journal of Electric Power, 2019,34(1): 35-41.
[2]	薛涵磊, 刘晓琴 . 基于模型预测方法的电网故障预测[J]. 辽宁石油化工大学学报, 2017(2).
	XUE H L , LIU X Q . Power grid fault forecast based on model prediction method[J]. Journal of Liaoning Shihua University, 2017(2).
[3]	王鑫, 吴际, 刘超 ,等. 基于 LSTM 循环神经网络的故障时间序列预测[J]. 北京航空航天大学学报, 2018(4): 772-784.
	WANG X , WU J , LIU C ,et al. Exploring LSTM based recurrent neural network for failure time series prediction[J]. Journal of Beijing University of Aeronautics and Astronautics, 2018(4): 772-784.
[4]	陈小峰, 赵雅迪, 张利鹏 ,等. 电力潜在敏感客户预测的大数据方法应用[J]. 电信科学, 2019,35(11): 117-124.
	CHEN X F , ZHAO Y D , ZHANG L P ,et al. Application of big data method in forecasting potential sensitive customers of electric power[J]. Telecommunications Science, 2019,35(11): 117-124.
[5]	SCHAPIRE R E , . A brief introduction to boosting[C]// The Six-teenth International Joint Conference on Artificial Intelligence,IJCAI 99,July 31-August 6,1999,Stockholm,Sweden.[S.l.:s.n]. 1999:1450.
[6]	TU Z , . Probabilistic boosting-tree:learning discriminative mod-els for classification,recognition,and clustering[C]// Tenth IEEE International Conference on Computer Vision,October 17-21,2005,Beijing,China.. Piscataway:IEEE Press, 2005.
[7]	FRIEDMAN J , HASTIE T , TIBSHIRANI R . Additive logistic regression:a statistical view of boosting[J]. Annals of Statistics, 2000,28(2): 337-374.
[8]	KE G , MENG Q , FINLEY T . LightGBM:a highly efficient gradient boosting decision tree[C]// Advances in Neural Infor-mation Processing Systems,December 4-9,2017,Long Beach,CA,USA.[S.l.:s.n]. 2017: 3146-3154.
[9]	CHEN T , GUESTRIN C . XGBoost:a scalable tree boosting system[C]// ACM Sigkdd International Conference on Knowledge Discovery ＆ Data Mining,August 24-27,2016,San Francisco,USA. New York:ACM Press, 2016: 785-794.
[10]	SCHUSTER M . Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing, 2002,45(11): 2673-2681.
[11]	HOCHREITER S , SCHMIDHUBER J . Long short-term memory[J]. Neural Computation, 1997,9(8): 1735-1780.
[12]	Mean absolute error[EB]. 2019.

统计指标	未过滤	过滤时长超过1天数据	过滤时长超过6 h数据
数据总数	201 412	200 916	194 333
过滤数据量 —		496(0.25%)	7 079(3.5%)
抢修时长均值	69.65	63.07	43.57
抢修时长标准差290.84		131.09	58.47
抢修时长最大值44 718		1 439	359

算法类型	平均绝对误差MAE
LightGBM	1.21
XGBoost	1.52
LSTM	1.23
加权结果	1.15

基于多模型融合的电网故障抢修时长预测

Prediction of power grid fault repair time based on multi-model fusion

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 12

相关文章 7

Metrics

推荐阅读 0

[1]	王建斌, 王淑春, 廖尚金, 施淑媛. 基于DCNN-LSTM负荷预测算法的5G基站节能系统研究[J]. 电信科学, 2023, 39(4): 133-141.
[2]	阿克弘, 胡晓东. 基于GAN数据重构的电信用户流失预测方法[J]. 电信科学, 2023, 39(3): 135-142.
[3]	诸葛斌, 王林超, 宋杨, 邵瑜, 董黎刚, 蒋献. 基于LSTM流量预测的路由规划和切换[J]. 电信科学, 2022, 38(8): 86-100.
[4]	冯楠, 刘贺林, 周泉, 杨海俊, 付蜜能, 钟大平. 5G核心网业务模型的智能化预测研究[J]. 电信科学, 2022, 38(8): 111-119.
[5]	沈潇军,葛亚男,沈志豪,倪阳旦,吕明琪,翁正秋. 一种基于LSTM自动编码机的工业系统异常检测方法[J]. 电信科学, 2020, 36(7): 136-145.
[6]	朱琳,赵娟,王伊婷,冯俊兰,邓超. 移动通信网络投诉热点问题智能预警方法[J]. 电信科学, 2019, 35(5): 17-24.
[7]	蔡鑫,娄京生. 基于LSTM深度学习模型的中国电信官方微博用户情绪分析[J]. 电信科学, 2017, 33(12): 136-141.