[1]卿会,郭军红,李薇,等.利用SVM-LSTM-DBN的短期光伏发电预测方法[J].华侨大学学报(自然科学版),2022,43(3):371-378.[doi:10.11830/ISSN.1000-5013.202104018]
　QING Hui,GUO Junhong,LI Wei,et al.Short-Term Photovoltaic Power Forecasting Method Based on SVM-LSTM-DBN[J].Journal of Huaqiao University(Natural Science),2022,43(3):371-378.[doi:10.11830/ISSN.1000-5013.202104018]
点击复制

利用SVM-LSTM-DBN的短期光伏发电预测方法()
分享到：

《华侨大学学报（自然科学版）》[ISSN:1000-5013/CN:35-1079/N]

卷:
第43卷
期数:
2022年第3期
页码:
371-378
栏目:
出版日期:
2022-05-10

文章信息/Info

Title:
Short-Term Photovoltaic Power Forecasting Method Based on SVM-LSTM-DBN
文章编号:
1000-5013(2022)03-0371-08
作者:
卿会1; 2;  郭军红1; 2;  李薇1; 2;  亢朋朋3;  王金明4;  潘张榕1; 2
1. 华北电力大学 环境科学与工程学院, 北京 102206;2. 华北电力大学 资源环境系统优化教育部重点实验室, 北京 102206;3. 国网新疆电力有限公司, 新疆 乌鲁木齐 830002;4. 国网新疆电力有限公司 阿勒泰供电公司, 新疆 阿勒泰 836500
Author(s):
QING Hui1; 2;  GUO Junhong1; 2;  LI Wei1; 2;  KANG Pengpeng3;  WANG Jinming4;  PAN Zhangrong1; 2
1. College of Environmental Science and Engineering, North China Electric Power University, Beijing 102206, China; 1. Key Laboratory of Resources and Environment System Optimization of Ministry of Education, North China Electric Power University, Beijing
关键词:
光伏发电;  光伏出力预测模型;  支持向量机;  长短期记忆神经网络;  深度信念网络
Keywords:
photovoltaic power generation;  photovoltaic output prediction model;  support vector machine;  long and short-term memory neural network;  deep belief network
分类号:
TM615;TP181
DOI:
10.11830/ISSN.1000-5013.202104018
文献标志码:
A
摘要:
为解决传统预测算法的不足,利用深度信念网络(DBN)耦合支持向量机(SVM)和长短期记忆神经网络(LSTM),提出一种新的光伏功率组合预测方法.分别构建以高斯径向基函数为核函数的支持向量机预测模型、4层长短期记忆神经网络为单项预测模型,通过深度信念网络组合,优化预测结果并输出.根据实际出力和预测结果的误差,利用DBN动态调整以获得最优值,进一步验证SVM-LSTM-DBN模型的有效性和准确性,并以新疆维吾尔自治区某光伏电站的实测数据进行仿真验证.结果表明:基于SVM-LSTM-DBN组合的光伏出力预测模型与单一模型相比,预测精度明显提高.
Abstract:
In order to solve the shortcomings of traditional forecasting algorithms, a new combination prediction method of photovoltaic power is proposed by using deep belief network(DBN)coupled support vector machine(SVM)and long short-term memory neural network(LSTM). The support vector machine prediction model with the kernel function of Gaussian radial basis function and the 4-layer long-short-term memory neural network as a single prediction model. Through the combination of deep belief networks, the prediction results are optimized and output. According to the actual output and the error of the prediction results, the DBN is used for dynamic adjustment to obtain optimal value, to further verify the validity and accuracy of the SVM-LSTM-DBN model. To take simulate and verify the actual measurement data of a photovoltaic power station in Xinjiang Uygur Autonomous Region. The results show that: compare the photovoltaic output prediction model based on the combination of SVM-LSTM-DBN and a single model, the prediction accuracy is significantly improved.

参考文献/References:

[1] 李旭.基于典型日出力特性分析的光伏电站功率预测研究[D].北京:华北电力大学,2016.DOI:10.7666/d.Y3114962.
[2] 李松威.基于神经网络的光伏发电功率预测研究[D].沈阳:沈阳工程学院,2017.
[3] 郑凯文,杨超.基于迭代决策树(GBDT)短期负荷预测研究[J].贵州电力技术,2017,20(2):82-84.DOI:10.19317/j.cnki.1008-083x.2017.02.019.
[4] DE GIORGI M G,CONGEDO P M,MALVONI M.Photovoltaic power forecasting using statistical methods: Impact of weather data[J].IET Science Measurement and Technology,2014,8(3):90-97.DOI:10.1049/IET-SMT.2013.0135.
[5] AHMAD M W,MOURSHED M,REZGUI Y.Tree-based ensemble methods for predicting PV power generation and their comparison with support vector regression[J].Energy,2018,164:465-474.DOI:10.1016/j.energy.2018.08.207.
[6] 姜恩宇,季亮,夏能弘,等.基于支持向量机的光伏发电功率预测[J].上海电力学院学报,2015,31(6):511-513,524 DOI:10.3969/j.issn.1006-4729.2015.06.002.
[7] 张雨金,周杭霞.Stacking-SVM的短期光伏发电功率预测[J].中国计量大学学报,2018,29(2):121-127.DOI:10.3969/j.issn.2096-2835.2018.02.002.
[8] JIN Lianghai,,XIONG Caiquan,LIU Hong.Improved bilateral filter for suppressing mixed noise in color images[J].Digital Signal Processing,2012,22(6):903-912.DOI:10.1016/j.dsp.2012.06.012.
[9] DU Nan,DAI Hanjun,TRIVEDI R,et al.Recurrent marked temporal point processes: Embedding event history to vector[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York:ACM,2016:1555-1564.DOI:10.1145/2939672.2939875.
[10] GONZáLEZ-HIDALGO M,MASSANET S,MIR A,et al.Improving salt and pepper noise removal using a fuzzy mathematical morphology-based filter[J].Applied Soft Computing,2018,63:167-180.DOI:10.1016/j.asoc.2017.11.030.
[11] 张春露,白艳萍.基于TensorFlow的LSTM模型在太原空气质量AQI指数预测中的应用[J].重庆理工大学学报(自然科学),2018,32(8):137-141.DOI:10.3969/j.issn.1674-8425(z).2018.08.021
[12] 赵淑芳,董小雨.基于改进的LSTM深度神经网络语音识别研究[J].郑州大学学报(工学版),2018,39(5):63-67.DOI:10.13705/j.issn.1671-6833.2018.02.004
[13] 陈卓,孙龙祥.基于深度学习LSTM网络的短期电力负荷预测方法[J].电子技术设计与应用,2018,01(001):39-41.DOI:10.3969/j.issn.1000-0755.2018.01.011.
[14] 李飞,高晓光,万开方.基于动态Gibbs采样的RBM训练算法研究[J].自动化学报,2016,42(6):931-942.DOI:10.16383/j.aas.2016.c150645.
[15] YUN Luo.An islanding detection method for photovoltaic power generation system using fluctuation characteristic of PCC harmonic voltage[J].Advanced Materials Research,2014,998/999:574-577.DOI:10.4028/www.scientific.net/AMR.998-999.574.
[16] 耿博,高贞彦,白恒远,等.结合相似日GA-BP神经网络的光伏发电预测[J].电力系统及其自动化学报,2017,29(6):118-123.DOI:10.3969/j.issn.1003-8930.2017.06.019.
[17] 贾俊平,何晓群,金勇进.统计学[M].7版.北京:中国人民大学出版社,2018.
[18] 孔祥玉,郑锋,鄂志君,等.基于深度信念网络的短期负荷预测方法[J].电力系统自动化,2018,42(5):133-139.DOI:10.7500/AEPS20170826002
[19] 林大贵.TensorFlow+Keras深度学习人工智能实践应用[M].北京:清华大学出版社,2018:193-196.

备注/Memo

备注/Memo:
收稿日期: 2021-04-12
通信作者: 李薇(1974-),女,教授,博士,博士生导师,主要从事能源环境污染控制、环境影响评价、环境规划与管理、节能减排优化、能源与环境系统分析等研究.E-mail:925657837@qq.com.
基金项目: 国家重点研发计划项目-战略性国际科技创新合作重点专项(2018YFE0208400)
更新日期/Last Update: 2022-05-20