基金项目:
国家重点研发计划项目(2017YFB0903000);
国家电网公司科技项目(52020116000G);
Project Supported by National Key Research and Development Program of China (2017YFB0903000);
Science and Technology Project of SGCC(52020116000G);
It is significant for operation and maintenance of distribution network to forecast risk level more accurately. In allusion to the problems of many factor and strong redundancy related to fault, a method of fault feature selection and fault outage risk level prediction for distribution network considering meteorological factors was proposed. Eighteen fault features of distribution network were summarized after data preprocessing and the basis of risk level was determined after considering failure frequency, proportion of outage duration and power supply. An improved G-ReliefF algorithm was proposed to calculate fault feature weights and eliminate redundancy. Then, an Adaboost based C4.5 decision tree algorithm was used to forecast the risk level of distribution network fault and find relationship between fault outage risk level and meteorological factors. Results of analyzing calculation example showed that the proposed method is effective. It provides an efficient basis for pre-control of distribution network risk.
KEY WORDS :distribution network;weather factor;feature selection;correlation;Adaboost;risk prediction;
[1]
徐特威,鲁宗相,乔颖,等.基于典型故障与环境场景关联识别的城市配电网运行风险预警方法[J].,2017,41(8):2577-2584.XuTewei,LuZongxiang,QiaoYing,et al.A risk warning method for urban distribution network based on associated recognition of typical fault and environment scenario[J].,2017,41(8):2577-2584(in Chinese).
[2]
李蕊,李跃,苏剑,等.配电网重要电力用户停电损失及应急策略[J].,2011,35(10):170-176.LiRui,LiYue,SuJian,et al.Power supply interruption cost of important power consumers in distribution network and its emergency management[J].,2011,35(10):170-176(in Chinese).
[3]
马瑞,周谢,彭舟,等.考虑气温因素的负荷特性统计指标关联特征数据挖掘[J].,2015,35(1):43-51.MaRui,ZhouXie,PengZhou,et al.Data mining on correlation feature of load characteristics statistical indexes considering temperature[J].,2015,35(1):43-51(in Chinese).
[4]
刘科研,盛万兴,张东霞,等.智能配电网大数据应用需求和场景分析研究[J].,2010,35(2):287-293.LiuKeyan,ShengWanxing,ZhangDongxia,et al.Big data application requirements and scenario analysis in smart distribution network[J].,2010,35(2):287-293(in Chinese).
[5]
KankanalaP,DasS,PahwaA.AdaBoost+: an ensemble learning approach for estimating weather-related outages in distribution systems[J].,2014,29(1):359-367.
[6]
ZhuD,ChengD,Broadwater RP,et al.Storm modeling for prediction of power distribution system outages[J].,2006,77(8):973-979.
[7]
LiuH,Davidson RA,Rosowsky DV,et al.Negative binomial regression of electric power outages in hurricanes[J].,2005,11(4):258-267.
[8]
LiuH,Davidson RA,Apanasovich TV.Spatial generalized linear mixed models of electric power outages due to hurricanes and ice storms[J].,2007,93(6):897-912.
[9]
Radmer DT,Kuntz PA,Christie RD,et al.Predicting vegetation-related failure rates for overhead distribution feeders[J].,2002,17(4):1170-1175.
[10]
LiH,Treinish LA,Hosking J R M.A statistical model for risk management of electric outage forecasts[J].IBM Journal of Research & Development,2010,54(3):8:1-8:11.
[11]
ZhouY,PahwaA,Yang SS.Modeling weather-related failures of overhead distribution lines[J].,2006,21(4):1683-1690.
[12]
孙小军,林圣,冯玎,等.考虑负荷特性的牵引变压器短期风险评估[J].,2016,40(9):2817-2823.SunXiaojun,LinSheng,FengDing,et al.Short-time risk evaluation of traction transformer based on loading characteristics[J].,2016,40(9):2817-2823(in Chinese).
[13]
费胜巍,孙宇.融合粗糙集与灰色理论的电力变压器故障预测[J].,2008,28(16):154-160.FeiShengwei,SunYu.Fault prediction of power transformer by combination of rough sets and grey theory[J].,2008,28(16):154-160(in Chinese).
[14]
程宝清,韩凤琴,桂中华.基于小波的灰色预测理论在水电机组故障预测中的应用[J].,2005,29(13):40-44.ChengBaoqing,HanFengqin,GuiZhonghua.Application of wavelet transform based grey theory to fault forecasting of hydroelectric generating sets[J].,2005,29(13):40-44(in Chinese).
[15]
吴文可,文福拴,薛禹胜,等.基于马尔可夫链的电力系统连锁故障预测[J].,2013,37(5):29-37.WuWenke,WenFushuan,XueYusheng,et al.Prediction of chain failure of power system based on markov chain[J].,2013,37(5):29-37(in Chinese).
[16]
李再华,白晓民,周子冠,等.基于特征挖掘的电网故障诊断方法[J].,2010,30(10):16-22.LiZaihua,BaiXiaomin,ZhouZiguan,et al.Method of power grid fault diagnosis based on feature mining[J]., 2010,30(10):16-22(in Chinese).
[17]
刁赢龙,盛万兴,刘科研,等.大规模配电网负荷数据在线清洗与修复方法研究[J].,2015,39(11):3134-3140.DiaoYinglong,ShengWanxing,LiuKeyan,et al.Research on online cleaning and repair methods of large-scale distribution network load data[J].,2015,39(11):3134-3140(in Chinese).
[18]
牛东晓,谷志红,邢棉,等.基于数据挖掘的SVM短期负荷预测方法研究[J].,2006,26(18):6-12.NiuDongxiao,GuZhihong,XingMian,et al.Study on forecasting approach to short-term load of SVM based on data mining[J].,2006,26(18):6-12(in Chinese).
[19]
胡丽娟,刁赢龙,刘科研,等.基于大数据技术的配电网运行可靠性分析[J].,2017,41(1):265-271.HuLijuan,DiaoYinglong,LiuKeyan,et al.Operational reliability analysis of distribution network based on big data technology[J].,2017,41(1):265-271(in Chinese).
[20]
张文俊. 配电网故障停电风险评估指标体系及评估方法研究[D].,2014.
[21]
周湶,廖婧舒,廖瑞金,等.含分布式电源的配电网停电风险快速评估[J].,2014,38(4):882-887.ZhouQuan,LiaoJingshu,LiaoRuijin,et al.Rapid assessment of power system blackout risk with distributed generation[J].,2014,38(4):882-887(in Chinese).
[22]
蒋玉娇,王晓丹,王文军,等.一种基于PCA和ReliefF的特征选择方法[J].,2010,46(26):170-172.JiangYujiao,WangXiaodan,WangWenjun,et al.New feature selection approach by PCA and ReliefF[J].,2010,46(26):170-172(in Chinese).
[23]
王雁凌,吴梦凯,周子青,等.基于改进灰色关联度的电力负荷影响因素量化分析模型[J].,2017,41(6):1772-1778.WangYanling,WuMengkai,ZhouZiqing,et al.Quantitative analysis model of power load influencing factors based on improved grey relational degree[J].,2017,41(6):1772-1778(in Chinese).
[24]
吴俊利,张步涵,王魁.基于Adaboost的BP神经网络改进算法在短期风速预测中的应用[J].,2012,36(9):221-225.WuJunli,ZhangBuhan,WangKui.Application of adaboost-based bp neural network for short-term wind speed forecast[J].,2012,36(9):221-225(in Chinese).
[25]
栗然,刘宇,黎静华,等.基于改进决策树算法的日特征负荷预测研究[J].,2005,25(23):36-41.LiRan,LiuYu,LiJinghua,et al.Study on the daily characteristic load forecasting based on the optimized algorithm of decision tree[J].,2005,25(23):36-41(in Chinese).