Short-Term Interval Forecasting of Photovoltaic Power Based on CEEMDAN-GSA-LSTM and SVR

doi:10.16183/j.cnki.jsjtu.2022.511

Abstract

Abstract:

Aimed at the intermittency and fluctuation of photovoltaic output power, a short-term interval prediction model of photovoltaic power is proposed. First, the model uses the complete ensemble empirical mode decomposition of adaptive noise (CEEMDAN) to decompose the historical photovoltaic output data into different components and define them as time-series components and random components according to their correlation with time-series features such as declination and time angles. Then, the long short-term memory (LSTM) neural network and the support vector regression (SVR) model optimized by the gravitational search algorithm (GSA) are used to predict the time series components and the random components respectively, and the prediction results of the time series components and the random components are superimposed to obtain the point prediction result. After the error is subjected to Johnson transformation and normal distribution modeling, the photovoltaic power interval prediction result is obtained. Finally, the effectiveness of the method is verified by an example. The comparison of the proposed model with other existing prediction models under different weather conditions suggests that the proposed model has a higher accuracy and a better robustness, which can provide precise confidence intervals based on point prediction values.

Key words: photovoltaic power prediction, interval prediction, complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), gravitational search algorithm (GSA), long short-term memory (LSTM), support vector regression (SVR), Johnson transformation

CLC Number:

TM615
P49

LI Fen, SUN Ling, WANG Yawei, QU Aifang, MEI Nian, ZHAO Jinbin. Short-Term Interval Forecasting of Photovoltaic Power Based on CEEMDAN-GSA-LSTM and SVR[J]. Journal of Shanghai Jiao Tong University, 2024, 58(6): 806-818.

Figures/Tables 17

Fig.1

Fig.2

Tab.1

Three types of Johnson transformation

分布系列	g(·)	约束条件
S_B	ln $X - ξ ξ + λ - X$	ξ<X<ξ+λ,λ>0
S_L	ln $X - ξ λ$	X>ξ,λ>0
S_U	arsinh $X - ξ λ$	λ>0

Tab.1

Fig.3

Fig.4

Fig.5

Tab.2

Tab.3

Tab.4

Fig.6

Fig.7

Tab.5

Fig.8

Fig.9

Fig.10

Tab.6

Tab.7

References 26

[1]	国家发展改革会, 国家能源局. 关于促进新时代新能源高质量发展的实施方案[EB/OL].(2022-05-30)[2022-05-30]. http://zfxxgk.nea.gov.cn/2022-05/30/c_1310608539.htm.
	National Development and Reform Commission, National Energy Administration. The implementation plan for promoting the high-quality development of new energy in a new era[EB/OL].(2022-05-30)[2022-05-30]. http://zfxxgk.nea.gov.cn/2022-05/30/c_1310608539.htm.
[2]	张雲钦, 程起泽, 蒋文杰, 等. 基于EMD-PCA-LSTM的光伏功率预测模型[J]. 太阳能学报, 2021, 42(9): 62-69.
	ZHANG Yunqin, CHENG Qize, JIANG Wenjie, et al. Photovoltaic power prediction model based on EMD-PCA-LSTM[J]. Acta Energiae Solaris Sinica, 2021, 42(9): 62-69.
[3]	叶林, 马明顺, 靳晶新, 等. 考虑风电-光伏功率相关性的因子分析-极限学习机聚合方法[J]. 电力系统自动化, 2021, 45(23): 31-40.
	YE Lin, MA Mingshun, JIN Jingxin, et al. Factor analysis-extreme learning machine aggregation method considering correlation of wind power and photovoltaic power[J]. Automation of Electric Power Systems, 2021, 45(23): 31-40.
[4]	黎嘉明, 艾小猛, 文劲宇, 等. 光伏发电功率持续时间特性的概率分布定量分析[J]. 电力系统自动化, 2017, 41(6): 30-36.
	LI Jiaming, AI Xiaomeng, WEN Jinyu, et al. Quantitative analysis of probability distribution for duration time characteristic of photovoltaic power[J]. Automation of Electric Power Systems, 2017, 41(6): 30-36.
[5]	李芬, 周尔畅, 孙改平, 等. 一种新型天气分型方法及其在光伏功率预测中的应用[J]. 上海交通大学学报, 2021, 55(12): 1510-1519. doi: 10.16183/j.cnki.jsjtu.2021.264
	LI Fen, ZHOU Erchang, SUN Gaiping, et al. A novel weather classification method and its application in photovoltaic power prediction[J]. Journal of Shanghai Jiao Tong University, 2021, 55(12): 1510-1519.
[6]	唐雅洁, 林达, 倪筹帷, 等. 基于XGBoost的双层协同实时校正超短期光伏预测[J]. 电力系统自动化, 2021, 45(7): 18-27.
	TANG Yajie, LIN Da, NI Chouwei, et al. XGBoost based Bi-layer collaborative real-time calibration for ultra-short-term photovoltaic prediction[J]. Automation of Electric Power Systems, 2021, 45(7): 18-27.
[7]	AGGA A, ABBOU A, LABBADI M, et al. CNN-LSTM: An efficient hybrid deep learning architecture for predicting short-term photovoltaic power production[J]. Electric Power Systems Research, 2022, 208: 107908.
[8]	万灿, 崔文康, 宋永华. 新能源电力系统概率预测:基本概念与数学原理[J]. 中国电机工程学报, 2021, 41(19): 6493-6509.
	WAN Can, CUI Wenkang, SONG Yonghua. Probabilistic forecasting for power systems with renewable energy sources: Basic concepts and mathematical principles[J]. Proceedings of the CSEE, 2021, 41(19): 6493-6509.
[9]	黎敏, 林湘宁, 张哲原, 等. 超短期光伏出力区间预测算法及其应用[J]. 电力系统自动化, 2019, 43(3): 10-16.
	LI Min, LIN Xiangning, ZHANG Zheyuan, et al. Interval prediction algorithm for ultra-short-term photovoltaic output and its Application[J]. Automation of Electric Power Systems, 2019, 43(3): 10-16.
[10]	杨茂, 王凯旋. 基于CEEMD-DBN模型的光伏出力日前区间预测[J]. 高电压技术, 2021, 47(4): 1156-1164.
	YANG Mao, WANG Kaixuan. Day-ahead interval forecasting of PV power based on CEEMD-DBN model[J]. High Voltage Engineering, 2021, 47(4): 1156-1164.
[11]	茆美琴, 龚文剑, 张榴晨, 等. 基于EEMD-SVM方法的光伏电站短期出力预测[J]. 中国电机工程学报, 2013, 33(34): 17-24.
	MAO Meiqin, GONG Wenjian, ZHANG Liuchen, et al. Short-term photovoltaic generation forecasting based on EEMD-SVM combined method[J]. Proceedings of the CSEE, 2013, 33(34): 17-24.
[12]	ZHANG C, HUA L, JI C L, et al. An evolutionary robust solar radiation prediction model based on WT-CEEMDAN and IASO-optimized outlier robust extreme learning machine[J]. Applied Energy, 2022, 322: 119518.
[13]	RASHEDI E, NEZAMABADI-POUR H, SARYAZDI S. GSA: A gravitational search algorithm[J]. Information Sciences, 2009, 179(13): 2232-2248.
[14]	ALI KHAN T, LING S H. A novel hybrid gravitational search particle swarm optimization algorithm[J]. Engineering Applications of Artificial Intelligence, 2021, 102: 104263.
[15]	杨晶显, 张帅, 刘继春, 等. 基于VMD和双重注意力机制LSTM的短期光伏功率预测[J]. 电力系统自动化, 2021, 45(3): 174-182.
	YANG Jingxian, ZHANG Shuai, LIU Jichun, et al. Short-term photovoltaic power prediction based on variational mode decomposition and long shortterm memory with dual-stage attention mechanism[J]. Automation of Electric Power Systems, 2021, 45(3): 174-182.
[16]	苏向敬, 周汶鑫, 李超杰, 等. 基于双重注意力LSTM神经网络的可解释海上风电出力预测[J]. 电力系统自动化, 2022, 46(7): 141-151.
	SU Xiangjing, ZHOU Wenxin, LI Chaojie, et al. Interpretable offshore wind power output forecasting based on long short-term memory neural network with dual-stage attention[J]. Automation of Electric Power Systems, 2022, 46(7): 141-151.
[17]	LIU R H, WEI J C, SUN G P, et al. A short-term probabilistic photovoltaic power prediction method based on feature selection and improved LSTM neural network[J]. Electric Power Systems Research, 2022, 210: 108069.
[18]	LI J L, SONG Z H, WANG X F, et al. A novel offshore wind farm typhoon wind speed prediction model based on PSO-Bi-LSTM improved by VMD[J]. Energy, 2022, 251: 123848.
[19]	董春曦, 饶鲜, 杨绍全, 等. 支持向量机参数选择方法研究[J]. 系统工程与电子技术, 2004, 26(8): 1117-1120.
	DONG Chunxi, RAO Xian, YANG Shaoquan, et al. Method for selecting the parameters of support vector machines[J]. Systems Engineering & Electronics, 2004, 26(8): 1117-1120.
[20]	徐浩, 姜新雄, 刘志成, 等. 基于概率预测的电网静态安全运行风险评估及主动调控策略[J]. 电力系统自动化, 2022, 46(1): 182-191.
	XU Hao, JIANG Xinxiong, LIU Zhicheng, et al. Probability prediction based risk assessment and proactive regulation and control strategy for static operation safety of power grid[J]. Automation of Electric Power Systems, 2022, 46(1): 182-191.
[21]	WANG H L, XU D X, MARTINEZ A. Parameter selection method for support vector machine based on adaptive fusion of multiple kernel functions and its application in fault diagnosis[J]. Neural Computing & Applications, 2020, 32(1): 183-193.
[22]	SLIFKER J F, SHAPIRO S S. The Johnson system: Selection and parameter estimation[J]. Technometrics, 1980, 22(2): 239-246.
[23]	HE Y Y, ZHENG Y Y. Short-term power load probability density forecasting based on Yeo-Johnson transformation quantile regression and Gaussian kernel function[J]. Energy, 2018, 154: 143-156.
[24]	汪冬华. 多元统计分析与SPSS应用[M]. 上海: 华东理工大学出版社, 2010.
	WANG Donghua. Multivariate statistical analysis and SPSS application[M]. Shanghai: East China University of Science and Technology Press, 2010.
[25]	LI F, LIN Y L, GUO J P, et al. Novel models to estimate hourly diffuse radiation fraction for global radiation based on weather type classification[J]. Renewable Energy, 2020, 157: 1222-1232.
[26]	DEMAIN C, JOURNÉE M, BERTRAND C. Evaluation of different models to estimate the global solar radiation on inclined surfaces[J]. Renewable Energy, 2013, 50: 710-721.

天气	分量	GSA-LSTM		LSTM		BP		SVR
天气	分量	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%
晴天	H_IMF1	20.3	3.7	27.5	4.8	54.4	15.0	50.5	13.2
晴天	H_IMF2	21.4	3.5	26.3	5.2	52.3	14.3	49.4	14.5
晴天	H_IMF3	18.7	2.6	21.9	3.7	47.2	12.1	48.8	13.6
晴天	H_IMF7	18.2	2.1	19.9	3.6	48.1	9.9	41.0	10.3
转折	H_IMF1	47.0	10.1	59.5	14.7	94.4	44.2	89.2	42.7
转折	H_IMF2	39.2	9.7	60.3	15.6	94.6	43.1	87.6	42.0
转折	H_IMF3	35.4	6.4	52.0	12.3	88.7	38.5	88.7	36.3
转折	H_IMF7	29.1	5.5	52.3	13.3	83.3	40.3	82.1	36.9
阴雨	H_IMF1	45.5	13.2	70.6	44.0	110.5	55.5	100.3	57.2
阴雨	H_IMF2	40.3	12.1	65.5	42.5	103.2	52.4	97.0	51.4
阴雨	H_IMF3	41.2	10.1	71.3	35.4	99.9	43.6	92.3	40.2
阴雨	H_IMF7	38.9	7.8	62.2	32.1	101.2	44.2	90.1	41.1

天气	分量	GSA-SVR		LSTM		BP		SVR
天气	分量	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%
晴天	H_IMF4	25.7	7.1	39.0	10.5	42.3	10.0	36.2	9.2
晴天	H_IMF5	24.3	6.8	38.7	10.7	37.0	9.6	27.9	9.8
晴天	H_IMF6	22.4	5.6	32.2	8.8	31.1	8.4	30.2	8.0
转折	H_IMF4	60.0	17.5	75.4	34.7	86.3	35.4	67.8	22.3
转折	H_IMF5	57.1	20.3	69.7	32.8	83.3	34.9	62.4	24.6
转折	H_IMF6	50.9	15.4	75.2	30.9	75.1	31.6	55.3	18.7
阴雨	H_IMF4	62.4	20.5	87.3	42.4	99.9	45.4	69.4	22.9
阴雨	H_IMF5	56.2	19.6	89.0	40.0	97.1	41.0	64.4	27.1
阴雨	H_IMF6	54.8	18.7	83.5	34.9	89.0	35.7	59.9	21.0

天气	CEEMDAN-GSA-LSTM 和SVR		EMD-GSA-LSTM		EMD-GSA-SVR		XGBoost-RC		CNN-LSTM
天气	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%
晴天	22.8	3.7	29.6	5.0	38.6	9.1	30.1	4.5	24.2	9.1
转折	40.0	9.1	61.8	16.0	64.1	22.2	47.1	21.2	55.4	10.1
阴雨	48.5	14.6	78.4	40.3	63.0	26.2	54.7	32.5	50.7	19.6

数据组别	CEEMDAN-GSA-LSTM和SVR		EMD-GSA-LSTM		EMD-GSA-SVR
数据组别	e_S/%	e_N/%	e_S/%	e_N/%	e_S/%	e_N/%
原始历史数据	48.5	14.6	78.4	40.3	63.0	26.2
D1	72.3	36.5	94.4	48.4	85.8	39.2
D2	73.4	38.9	98.9	53.3	91.1	57.1
D3	203.5	250.4	200.4	287.5	241.1	329.9
D4	212.3	233.9	240.1	298.8	242.9	310.5

置信水平/ %	晴天		转折		阴雨
置信水平/ %	Q_PICP	Q_PINAW	Q_PICP	Q_PINAW	Q_PICP	Q_PINAW
95	1.0000	0.1307	0.9697	0.2393	0.9394	0.2238
80	0.9394	0.0984	0.9394	0.1876	0.9091	0.1733
60	0.8485	0.0703	0.8182	0.1382	0.6970	0.1282
40	0.6364	0.0460	0.6364	0.0923	0.4848	0.0864