Suzhou Electric Appliance Research Institute
期刊號(hào): CN32-1800/TM| ISSN1007-3175

Article retrieval

文章檢索

首頁(yè) >> 文章檢索 >> 往年索引

一種基于長(zhǎng)短期記憶網(wǎng)絡(luò)的線路覆冰預(yù)測(cè)模型研究

來(lái)源:電工電氣發(fā)布時(shí)間:2020-03-27 13:27 瀏覽次數(shù):1126
一種基于長(zhǎng)短期記憶網(wǎng)絡(luò)的線路覆冰預(yù)測(cè)模型研究
 
陳雨鴿1,高偉1,林鴻偉2,阮肇華2,鄭為湊2,林福2,陳錦植2
(1 福州大學(xué) 電氣工程與自動(dòng)化學(xué)院,福建 福州 350108;
2 國(guó)網(wǎng)福建省電力有限公司寧德供電公司,福建 寧德 352100)
 
    摘 要:輸電線路覆冰災(zāi)害易引發(fā)危害電網(wǎng)安全運(yùn)行的事故,對(duì)輸電線路覆冰情況進(jìn)行短期預(yù)測(cè)十分必要。提出了一種基于結(jié)合氣象因素和導(dǎo)線覆冰量的時(shí)間序列模型預(yù)測(cè)法,建立一個(gè)由5 個(gè)氣象要素和一個(gè)導(dǎo)線覆冰量數(shù)據(jù)組成的數(shù)據(jù)集,采用長(zhǎng)短期記憶網(wǎng)絡(luò)算法訓(xùn)練預(yù)測(cè)模型,利用線路實(shí)際運(yùn)行數(shù)據(jù)對(duì)模型進(jìn)行優(yōu)化和評(píng)估。實(shí)驗(yàn)結(jié)果表明,所提方法可準(zhǔn)確、有效地實(shí)現(xiàn)線路覆冰發(fā)展情況的預(yù)測(cè),預(yù)測(cè)誤差僅4.2%。
    關(guān)鍵詞:輸電線路;覆冰預(yù)測(cè);氣象信息;長(zhǎng)短期記憶網(wǎng)絡(luò)
    中圖分類號(hào):TM74     文獻(xiàn)標(biāo)識(shí)碼:A     文章編號(hào):1007-3175(2020)03-0005-07
 
Study of an Icing Prediction Model for Transmission Line Based on Long and Short-Term Memory Network
 
CHEN Yu-ge1, GAO Wei1, LIN Hong-wei2, RUAN Zhao-hua2, ZHENG Wei-cou2, LIN Fu2, CHEN Jin-zhi2
(1 College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108, China;
2 Ningde Power Supply Company of State Grid Fujian Electric Power Co., Ltd, Ningde 352100, China)
 
    Abstract: Ice disasters on power transmission lines can easily lead to accidents, endanger the safe operation of the power grid and bring economic losses. Therefore, it is important to make short-term predictions of icing conditions on transmission lines. A time series model prediction method combining meteorological factors and the amount of icing on the wire is proposed. First, a data set consisting of five meteorological factors and the amount of ice covered by the traverse is established. Secondly, the long and short-term memory network algorithm is used to train the prediction model. Finally, the model is optimized and evaluated using the actual operating data of the line. The experimental results show that the proposed method can accurately and effectively predict the development of icing on the line, and the prediction error is only 4.2%.
    Key words: transmission line; icing prediction; meteorological information; long and short-term memory network
 
參考文獻(xiàn)
[1] 李慶峰,范崢,吳穹,等. 全國(guó)輸電線路覆冰情況調(diào)研及事故分析[J]. 電網(wǎng)技術(shù),2008,32(9):33-36.
[2] 陸佳政, 蔣正龍, 雷紅才, 等. 湖南電網(wǎng)2008年冰災(zāi)事故分析[J]. 電力系統(tǒng)自動(dòng)化,2008,32(11):16-19.
[3] 李偉,馬佳,王世蓉,等. 精細(xì)化氣象要素下輸電線路覆冰預(yù)測(cè)預(yù)警研究[J]. 電力大數(shù)據(jù),2018,21(2):1-7.
[4] 黃新波,劉家兵,蔡偉,等. 電力架空線路覆冰雪的國(guó)內(nèi)外研究現(xiàn)狀[J]. 電網(wǎng)技術(shù),2008,32(4):23-28.
[5] 苑吉河,蔣興良,易輝,等. 輸電線路導(dǎo)線覆冰的國(guó)內(nèi)外研究現(xiàn)狀[J]. 高電壓技術(shù),2004,30(1):6-9.
[6] 趙陽(yáng), 安佳坤, 孟斌, 等. 基于覆冰動(dòng)態(tài)增長(zhǎng)的電力系統(tǒng)風(fēng)險(xiǎn)評(píng)估方法[J]. 科技創(chuàng)新與應(yīng)用,2016(33):193-194.
[7] SAKAMOTO Y. Snow accretion on overhead wires[J].Philosophical Transactions of the Royal Society A:Mathematical, Physical and Engineering Sciences,2000,358(1776):2941-2970.
[8] LUO Y, LI Y, YAO Y, et al.Research on power transmission line ice prediction system based on BP neural network[C]//IEEE International Conference on Measurement,2012.
[9] 韓葉良,蘇國(guó)鋒,袁宏永,等. 基于粗糙集的電網(wǎng)覆冰事故預(yù)警模型[J]. 清華大學(xué)學(xué)報(bào)( 自然科學(xué)版),2010,50(12):1930-1933.
[10] 劉春城,劉佼. 輸電線路導(dǎo)線覆冰機(jī)理及雨凇覆冰模型[J]. 高電壓技術(shù),2011,37(1):241-248.
[11] 黃新波,王玉鑫,朱永燦,等. 基于遺傳算法與模糊邏輯融合的線路覆冰預(yù)測(cè)[J]. 高電壓技術(shù),2016,42(4):1228-1235.
[12] 莊文兵,祁創(chuàng),王建,等. 基于微氣象監(jiān)測(cè)的輸電線路覆冰動(dòng)態(tài)過(guò)程估計(jì)模型[J]. 電力系統(tǒng)保護(hù)與控制,2019,47(14):87-94.
[13] HOCHREITER S, SCHMIDHUBER J.Long short-term memory[J].Neural Computation,1997,9(8):1735-1780.
[14] 石曉文,蔣洪迅. 面向高精度與強(qiáng)魯棒的空氣質(zhì)量預(yù)測(cè)LSTM模型研究[J]. 統(tǒng)計(jì)與決策,2019(16):49-53.