基于社群特征的配電網(wǎng)異常用電行為分析
董津辰，雷景生
（上海電力學(xué)院 計算機科學(xué)與技術(shù)學(xué)院，上海 200090）
摘 要：針對目前配電網(wǎng)異常用電行為精度欠佳、效率低下、人力資源耗費量大等問題，在海量用電數(shù)據(jù)中利用數(shù)據(jù)挖掘技術(shù)實現(xiàn)異常用電數(shù)據(jù)的精確查找與定位。通過引入社群習(xí)慣的行業(yè)季節(jié)用電水平等異常分類指標，對可能存在非技術(shù)性損耗（NTL）的配網(wǎng)用戶進行分析和檢測，利用改進粒子群LM 神經(jīng)網(wǎng)絡(luò)算法建立了有效的異常用電行為的自動識別模型。實驗結(jié)果表明：該模型能夠有效地提取用電特征，實現(xiàn)對異常用戶的檢測，具有較強的識別能力和較高的實用性。
關(guān)鍵詞：異常用電；非技術(shù)性損耗；社群特征；改進粒子群算法
中圖分類號：TM744 文獻標識碼：A 文章編號：1007-3175(2019)01-0014-06
Abnormal Power Consumption Behavioural Analysis of Power Distribution Network Based on Association Characteristic
DONG Jin-chen, LEI Jing-sheng
（College of Computer Science and Technology, Shanghai University of Electric Power, Shanghai 200090, China）
Abstract: In order to solve the problem of poor accuracy, low efficiency, and high consumption of human resources in abnormal power consumption of power distribution network, this paper used data mining technology to accurately locate abnormal power consumption data in magnanimity power utilization data. The network users who might have non-technical loss (NTL) were analyzed and detected by using the industry's seasonal power consumption level of the community's habits and other abnormal classification indicators. The improved particle swarm LM neural network optimization algorithm was utilized to establishe an effective automatic recognition model for abnormal power consumption. The experimental results show that this model can effectively extract the electricity characteristics and realize the detection of abnormal users with strong recognition ability and high practicability.
Key words: abnormal power consumption; non-technical loss; community feature; improved particle swarm optimization
參考文獻
[1] 宋亞奇，周國亮，朱永利. 智能電網(wǎng)大數(shù)據(jù)處理技術(shù)現(xiàn)狀與挑戰(zhàn)[J]. 電網(wǎng)技術(shù)，2013，37(4)：927-935.
[2] LEAL A G, BOLDT M. A big data analytics design patterns to select customers for electricity theft inspection[C]//IEEE PES Transmission & Distribution Conference & Exposition-Latin America，2016.
[3] CABRAL J E, GONTIJO E M, PINTO J O P, et al. Fraud detection in electrical energy consumers using rough sets[C]//IEEE International Conference on Systems, Man and Cybernetics，2004.
[4] FOURIE J W, CALMEYER J E. A statistical method to minimize electrical energy losses in a local electricity distribution network[C]//IEEE Africon Conference in Africa，2004.
[5] BILBAO J, TORRES E, EGUFA P, et al. Determination of energy losses[C]//16th International Conference & Exhibition on Electricity Distribution，2001.
[6] MONEDERO I, BISCARRI F, LEON C, et al. Detection of frauds and other non-technical losses in a power utility using Pearson coefficient, Bayesian networks and decision trees[J]. International Journal of Electrical Power & Energy Systems，2012，34(1)：90-98.
[7] FILHO J R, GONTIJO E M, DELAIBA A C, et al. Fraud identification in electricity company customers using decision tree[C]//IEEE International Conference on Systems, Man and Cybernetics，2004.
[8] NIZAR A H, DONG Z Y, ZHAO J H, et al. A data mining based NTL analysis method[C]//IEEE Power Engineering Society General Meeting，2007.
[9] NAGI J, MOHAMMAD A M, YAP K S, et al. Non-Technical Loss Analysis for Detection of Electricity Theft Using Support Vector Machines[C]//IEEE 2nd International Power & Energy Conference，2008.
[10] NAGI J, YAP K S, TIONG S K, et al. Improving SVM-Based Nontechnical Loss Detection in Power Utility Using the Fuzzy Inference System[J]. IEEE Transactions on Power Delivery，2011，26(2)：1284-1285.
[11] 薛安榮，姚林，鞠時光，等. 離群點挖掘方法綜述[J]. 計算機科學(xué)，2008，35(11)：13-18.
[12] 劉濤，楊勁鋒，闕華坤，等. 自適應(yīng)的竊漏電診斷方法研究及應(yīng)用[J]. 電氣自動化，2014，36(2)：60-62.
[13] 張長勝，歐陽丹彤，岳娜，等. 一種基于遺傳算法和LM 算法的混合學(xué)習(xí)算法[J]. 吉林大學(xué)學(xué)報( 理學(xué)版)，2008，46(4)：675-680.
[14] 馬廷洪，姜磊. 基于混合粒子群算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的機床熱誤差建模[J]. 中國工程機械學(xué)報，2018，16(3)：221-224.
[15] 田野，張程，毛昕儒，等. 運用PCA改進BP神經(jīng)網(wǎng)絡(luò)的用電異常行為檢測[J]. 重慶理工大學(xué)學(xué)報( 自然科學(xué)版)，2017，31(8)：125-133.
