本文關(guān)鍵詞： 負荷管理 模式分類 異常數(shù)據(jù)識別 非參數(shù)回歸分析　出處：《電力系統(tǒng)自動化》2017年18期 　論文類型：期刊論文

【摘要】：工業(yè)負荷數(shù)據(jù)記錄了用戶的用電模式以及電量需求水平等重要信息,但是會因為干擾而導致記錄數(shù)據(jù)中摻雜有異常值。針對上述問題,文中提出了利用非參數(shù)回歸理論對工業(yè)用戶負荷異常值展開辨析和更正。首先,考慮負荷數(shù)據(jù)時序相關(guān)特性,采用統(tǒng)計模糊矩陣分類法,對工業(yè)用戶負荷進行用電模式分類,將負荷數(shù)據(jù)分為基礎用電模式數(shù)據(jù)集和特殊用電模式數(shù)據(jù)集。然后,利用基礎用電模式數(shù)據(jù)集,考慮各時刻的負荷數(shù)值分布情況,通過非參數(shù)回歸分析方法提取中心負荷向量,進而構(gòu)造異常數(shù)據(jù)域,對負荷異常值進行識別。最后,在常規(guī)加權(quán)均值法的基礎上,引入負荷水平映射關(guān)系,完成對負荷異常值的修正。算例采用實際工業(yè)負荷數(shù)據(jù)進行測試,結(jié)果表明了所提方法的準確性。
[Abstract]:Industrial load data record the important information such as power consumption mode and power demand level, but it will lead to the abnormal value in the recorded data because of interference. In this paper, the nonparametric regression theory is proposed to discriminate and correct the abnormal value of industrial user load. Firstly, considering the time-related characteristics of load data, the statistical fuzzy matrix classification method is adopted. The load data are divided into basic power mode data set and special power mode data set. Then, the basic power consumption mode data set is used. Considering the load numerical distribution at each time, the center load vector is extracted by non-parametric regression analysis, and then the abnormal data domain is constructed to identify the load outliers. Based on the conventional weighted mean method, the load level mapping relationship is introduced to complete the correction of the abnormal load value. The actual industrial load data are used to test the proposed method. The results show the accuracy of the proposed method.
【作者單位】： 陜西省智能電網(wǎng)重點實驗室(西安交通大學電氣工程學院);國網(wǎng)銅川供電公司;
【基金】：陜西省重點研發(fā)計劃重點產(chǎn)業(yè)創(chuàng)新鏈資助項目(2017ZDCXL-GY-02-03)~~
【分類號】：TM714
【正文快照】： 上網(wǎng)日期:2017-07-28。陜西省重點研發(fā)計劃重點產(chǎn)業(yè)創(chuàng)新鏈資助項目(2017ZDCXL-GY-02-03)。0引言隨著中國電力網(wǎng)絡智能化水平的提高和電力市場改革的深入,電力負荷數(shù)據(jù)逐漸成為應用最為廣泛的數(shù)據(jù)之一,其在電力系統(tǒng)運行、電力消費分析和用電需求管控等方面,都起到了不可替代的，

本文編號：1473395