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

Article retrieval

文章檢索

首頁 >> 文章檢索 >> 文章瀏覽排名

基于CNN-LSTM網(wǎng)絡的短期電力負荷預測

來源:電工電氣發(fā)布時間:2022-09-26 16:26 瀏覽次數(shù):384

基于CNN-LSTM網(wǎng)絡的短期電力負荷預測

簡定輝,李萍,黃宇航,梁志洋
(寧夏大學 物理與電子電氣工程學院,寧夏 銀川 750021)
 
    摘 要:傳統(tǒng)的神經(jīng)網(wǎng)絡在時間相關性較強的負荷預測中精度不高。為了有效提高短期電力負荷預測精度,提出了一種基于卷積神經(jīng)網(wǎng)絡 CNN 和長短時記憶網(wǎng)絡 LSTM 相結(jié)合的負荷預測方法。采集 5 維負荷特征數(shù)據(jù),以 CNN 卷積層和池化層作為特征提取單元,提取數(shù)據(jù)空間耦合交互特征;將重構數(shù)據(jù)輸入到 LSTM 網(wǎng)絡挖掘負荷時序特征,采用 Dropout 技術增加模型泛化能力;利用適應性矩估計 (Adam) 優(yōu)化器訓練模型;將測試數(shù)據(jù)輸入訓練后的神經(jīng)網(wǎng)絡模型,預測未來 1h 和 12h 電負荷。實驗結(jié)果表明,該負荷預測模型收斂速度和預測精度均優(yōu)于改進的 BP 神經(jīng)網(wǎng)絡、LSTM 等預測模型,其 1h 負荷預測精度達到98.66%,12h 負荷預測精度達到96.81%,提高了短期電力負荷預測精度。
    關鍵詞: 長短時記憶網(wǎng)絡;短期負荷預測;Dropout 技術;卷積神經(jīng)網(wǎng)絡;適應性矩估計
    中圖分類號:TM715     文獻標識碼:A     文章編號:1007-3175(2022)09-0001-06
 
Short-Term Power Load Forecasting Based on CNN-LSTM
 
JIAN Ding-hui, LI Ping, HUANG Yu-hang, LIANG Zhi-yang
(School of Physics and Electronic-Electrical Engineering, Ningxia University, Yinchuan 750021, China)
 
    Abstract: The traditional neural network has low accuracy in load forecasting with strong time dependence. This paper provided a load prediction method based on the convolutional neural network (CNN) and the long short-term memory network (LSTM) to improve the accuracy of short-term power load.Moreover, it collected 5-dimensional load characteristic data and extracted spatial coupling interaction features of data by using CNN convolution layer and pooling layers feature extraction units.In addition, it inputted the reconstructed data into the LSTM network to mine the load timing characteristics and used dropout technology to increase the model generalization ability. Besides, it used an adaptive moment estimation (Adam) optimizer to train the model. It entered the test data into the trained neural network model to predict the electric load in the next 1 h and 12 h. The experimental results show that the proposed model is better than the improved neural network forecasting models,such as improved BP neural network and LSTM, from the convergence speed and forecasting accuracy perspective. The prediction accuracy of 1 h load forecasting is 98.66%, and the 12 h load forecasting accuracy is 96.81%, which improves the accuracy of short-term power load forecasting.
    Key words: long short-term memory network; short-term load forecasting; Dropout technology; convolutional neural network; adaptive moment estimation
 
參考文獻
[1] 商立群,李洪波,侯亞東,黃辰浩,張建濤. 基于特征選擇和優(yōu)化極限學習機的短期電力負荷預測[J]. 西安交通大學學報,2022,56(4) :165-175.
[2] 王艷松,趙惺,李強,李雪,魏澈. 基于油氣開采的海上油田中長期電力負荷預測[J] . 中國石油大學學報(自然科學版),2021,45(2) :127-133.
[3] 王繼東,杜沖. 基于Attention-BiLSTM 神經(jīng)網(wǎng)絡和氣象數(shù)據(jù)修正的短期負荷預測模型[J] . 電力自動化設備,2022,42(4):172-177.
[4] 李婷婷,畢海權,王宏林,王曉亮,周遠龍. 基于 BP 神經(jīng)網(wǎng)絡的地鐵站廳空調(diào)負荷預測[J] . 計算機科學,2019,46(S2):590-594.
[5] 廖慶陵,竇震海,孫鍇,朱亞玲. 基于自適應粒子群算法優(yōu)化支持向量機的負荷預測[J] . 現(xiàn)代電子技術,2022,45(3):125-129.
[6] 王健,易姝慧,劉俊杰,劉儉. 基于隨機森林算法和穩(wěn)態(tài)波形的非介入式工業(yè)負荷辨識[J] . 中國電力,2022,55(2):82-89.
[7] 何桂雄,金璐,李克成,何偉,閆華光. 基于改進 DaNN 的綜合能源系統(tǒng)多能負荷預測[J] . 電力工程技術,2021,40(6):25-33.
[8] 荀超, 陳伯建, 吳翔宇, 項康利, 林可堯, 肖芬,易楊. 基于改進 K-means 算法的電力短期負荷預測方法研究[J] . 電力科學與技術學報,2022,37(1):90-95.
[9] 尹春杰,肖發(fā)達,李鵬飛,趙欽. 基于 LSTM 神經(jīng)網(wǎng)絡的區(qū)域微網(wǎng)短期負荷預測[J] . 計算機與現(xiàn)代化,2022(4):7-11.
[10] 魏驁,茅大鈞,韓萬里,呂彬. 基于 EMD 和長短期記憶網(wǎng)絡的短期電力負荷預測研究[J] . 熱能動力工程,2020,35(4):203-209.
[11] 胡欣球,馬立新. VMD-LSTM 算法在短期負荷預測中的應用[J] . 電力科學與工程,2018,34(6):9-13.
[12] 宋珊珊,潘文林,王嘉梅,梁志茂. 基于 CNN-BiLSTM-Attention 的超短期電力負荷預測[J]. 云南民族大學學報(自然科學版),2022,31(2):235-240.
[13] 方娜,余俊杰,李俊曉,萬暢. 基于 CNN-BIGRU-ATTENTION 的短期電力負荷預測[J]. 計算機仿真,2022,39(2):40-44.
[14] 程江洲,潘飛,鮑剛,何艷,陳奕睿. 基于 MAC-WD-CNN-MCNN 模型的超短期負荷預測[J]. 水電能源科學,2021,39(9):205-209.
[15] 朱凌建,荀子涵,王裕鑫,崔強,陳文義,婁俊超. 基于 CNN-BiLSTM 的短期電力負荷預測[J]. 電網(wǎng)技術,2021,45(11):4532-4539.
[16] 張林,賴向平,仲書勇,李柯沂. 基于正交小波和長短期記憶神經(jīng)網(wǎng)絡的用電負荷預測方法[J] . 現(xiàn)代電力,2022,39(1):72-79.
[17] 孫俊峰,李志斌. 基于 LSTM 的滾動預測算法的電纜纜芯溫度的研究[J] . 電子測量技術,2021,44(21):84-88.
[18] TANG Yehui, WANG Yunhe, XU Yixing, SHI Boxin, XU Chao, XU Chunjing, XU Chang.Beyond Dropout:Feature Map Distortion to Regularize Deep Neural Networks[J].Proceedings of the AAAI Conference on Artificial Intelligence,2020,34(4):5964-5971.
[19] WEI Yuqin, WENG Zhengxin.Research on TE process fault diagnosis method based on DBN and dropout[J].The Canadian Journal of Chemical Engineering,2020,98(6) :1293-1306.