基于自組織競(jìng)爭(zhēng)網(wǎng)絡(luò)與RPROP算法的線損計(jì)算研究
張艷，徐衛(wèi)鋒
(國(guó)網(wǎng)上海市電力公司市南供電公司，上海 200233)
摘 要：為更好地發(fā)現(xiàn)高效的降損措施，并為科學(xué)地制定線損目標(biāo)提供依據(jù)，提出了一種基于自組織競(jìng)爭(zhēng)神經(jīng)網(wǎng)絡(luò)的 RPROP 神經(jīng)網(wǎng)絡(luò)的線損計(jì)算方法。RPROP 神經(jīng)網(wǎng)絡(luò)確保了網(wǎng)絡(luò)在有限的訓(xùn)練次數(shù)下能夠收斂，利用自組織競(jìng)爭(zhēng)神經(jīng)網(wǎng)絡(luò)對(duì)信息數(shù)據(jù)進(jìn)行有效分類(lèi)，提高了 RPROP 神經(jīng)網(wǎng)絡(luò)的輸出精度。通過(guò)在 MATLAB 平臺(tái)進(jìn)行仿真實(shí)驗(yàn)，并與線性回歸算法、標(biāo)準(zhǔn) BP 神經(jīng)網(wǎng)絡(luò)算法，以及未分類(lèi)的 RPROP 算法進(jìn)行比較，驗(yàn)證了該方法的有效性。
關(guān)鍵詞： 線性回歸算法；BP 神經(jīng)網(wǎng)絡(luò)；RPROP 神經(jīng)網(wǎng)絡(luò)；自組織競(jìng)爭(zhēng)神經(jīng)網(wǎng)絡(luò)；線損
中圖分類(lèi)號(hào)：TM744 文獻(xiàn)標(biāo)識(shí)碼：A 文章編號(hào)：1007-3175(2022)07-0031-04
The Research on Line Loss Calculation of RPROP Algorithm Based on
Self-Organizing Competitive Network
ZHANG Yan, XU Wei-feng
（State Grid Shanghai Shinan Electric Power Supply Company, Shanghai 200233, China）
Abstract: This paper proposed a line loss calculation based on the self-organizing competitive network of the RPROP neural network to find efficient loss reduction measures and provide the basis for scientifically formulating line loss targets.The RPROP neural network ensured that the network could converge under a limited number of training times. Moreover, it utilized a self-organizing competitive neural network to effectively classify informative data, which improved the output accuracy of the RPROP neural network.By doing simulation experiments on the MATLAB platform and comparing with linear regression algorithm, standard BP neural network algorithm, unclassified RPROP algorithm,it verified the effectiveness of the proposed method.
Key words: linear regression algorithm; BP neural network; RPROP neural network; self-organizing competitive neural network; line loss
參考文獻(xiàn)
[1] 李俊楠，閆利，張世林，等. 綜合線路率及線損波動(dòng)分析[J]. 電力系統(tǒng)裝備，2020(15) ：115-116.
[2] 張銀，張祥華，伏圣群，等. 中壓配電網(wǎng)極限線損計(jì)算方法研究[J] . 廣西科技大學(xué)學(xué)報(bào)， 2017，28(2) ：67-73.
[3] 陳哲. 基于 BP 神經(jīng)網(wǎng)絡(luò)的配網(wǎng)設(shè)備故障預(yù)測(cè)[D] .廣州：廣東工業(yè)大學(xué)，2017.
[4] 倪洋. 基于 BP 神經(jīng)網(wǎng)絡(luò)的配網(wǎng)線損計(jì)算分析[D] .大連：大連理工大學(xué)，2018.
[5] 馬銳．人工神經(jīng)網(wǎng)絡(luò)原理[M] . 北京：機(jī)械工業(yè)出版社，2010.
[6] 陳明.MATLAB 神經(jīng)網(wǎng)絡(luò)原理與實(shí)例精解[M]. 北京：清華大學(xué)出版社，2013.
[7] RIEDMILLER M, BRAUN H.A direct adaptive method for faster back propagation learning:The RPROP algorithm[C]//IEEE International Conference on Neural Networks，1993.
[8] 朱凱，王正林. 精通 MATLAB 神經(jīng)網(wǎng)絡(luò)[M]. 北京：電子工業(yè)出版社，2010.
[9] 張德豐.MATLAB 神經(jīng)網(wǎng)絡(luò)應(yīng)用設(shè)計(jì)[M].2 版. 北京：機(jī)械工業(yè)出版社，2012.
