基于相似天气波动分型的短期光伏发电功率预测

doi:10.16180/j.cnki.issn1007-7820.2025.08.011

摘要/Abstract

摘要：

针对光伏出力受天气波动影响较大,天气输入特征分布分散导致预测精度不高的问题,文中提出一种考虑相似天气波动分型的短期光伏发电功率预测模型。通过相关性分析筛选出强相关性气象因子作为输入数据,利用主成分分析法将多维天气特征数据聚合成一维综合气象数据,利用综合气象数据的波动特征来代表大部分天气数据的波动特征。为了使聚类后的波动特征更集中,采用综合气象数据的5个统计指标作为K-means聚类算法的聚类特征,对天气数据进行分型。使用基于双通道卷积神经网络(Convolutional Neural Networks, CNN)的双向门控循环网络(Bidirectional Gated Recurrent Unit, BiGRU)-Attention预测模型对3种天气类型下的光伏发电数据分别进行预测。通过对天气数据分型使具有相似波动的天气特征更集中。相较于传统算法,所提预测方法的精度更高,验证了所提方法和模型的有效性。

关键词: 光伏发电, 天气分型, 相关性分析, 主成分分析, 卷积神经, 统计指标, K-means聚类, BiGRU-ATT模型

Abstract:

In view of the problem that photovoltaic power output is greatly affected by weather fluctuations and the distribution of weather input features is scattered resulting in low prediction accuracy, this study proposes a short-term photovoltaic power prediction model considering similar weather fluctuation fractals. Strongly correlated weather factors are screened as input data through correlation analysis. Multi-dimensional weather features are aggregated into one-dimensional comprehensive weather data using principal component analysis, and the fluctuation features of the comprehensive weather data are used to represent the fluctuation features of most of the weather data. To make the fluctuation characteristics after clustering more concentrated, five statistical indicators of the comprehensive weather data were used as the clustering features of the K-means clustering algorithm to typify the weather data. A BiGRU(Bidirectional Gated Recurrent Unit)-Attention prediction model based on two-channel CNN(Convolutional Neural Networks) is used to predict the photovoltaic power generation data under three weather types. Compared with the traditional algorithm, the accuracy of the proposed prediction method is higher, which verifies the effectiveness of the proposed method and model.

Key words: photovoltaic power, weather typing, correlation analysis, principal component analysis, convolutional neural, statistical indicators, K-means clustering, BiGRU-ATT model

中图分类号:

TP183

王林, 王林先, 李伟硕, 王涛, 雒志恒, 朱家君. 基于相似天气波动分型的短期光伏发电功率预测[J]. 电子科技, 2025, 38(8): 79-86.

WANG Lin, WANG Linxian, LI Weishuo, WANG Tao, LUO Zhiheng, ZHU Jiajun. Short-Term Photovoltaic Power Generation Forecasting Based on Similar Weather Fluctuation Patterns[J]. Electronic Science and Technology, 2025, 38(8): 79-86.

图/表 15

图1

图2

表1

图3

图4

图5

图6

图7

图8

图9

表2

图10

图11

图12

表3

参考文献 20

[1]	黎静华, 骆怡辰, 杨舒惠, 等. 可再生能源电力不确定性预测方法综述[J]. 高电压技术, 2021, 47(4):1144-1155.
	Li Jinghua, Luo Yichen, Yang Shuhui, et al. Review of uncertainty forecasting methods for renewable energy power[J]. High Voltage Technology, 2021, 47(4):1144-1155.
[2]	Rafi S H, Deeba S R, Hossain E, et al. A short-term load forecasting method using integrated CNN and LSTM network[J]. IEEE Access, 2021, 9(1):32436-32448.
[3]	简献忠, 顾洪志, 王如志. 一种基于双通道CNN和LSTM的短期光伏功率预测方法[J]. 电力科学与工程, 2019, 35(5):7-11. doi: 1672-0792(2019)05-0007-05
	Jian Xianzhong, Gu Hongzhi, Wang Ruzhi. A short-term photovoltaic power prediction method based on dual-channel CNN and LSTM[J]. Electric Power Science and Engineering, 2019, 35(5):7-11. doi: 1672-0792(2019)05-0007-05
[4]	黄圆, 魏云冰, 童东兵, 等. 基于VMD和改进TCN的短期光伏发电功率预测[J]. 电子科技, 2023, 36(3):42-49.
	Huang Yuan, Wei Yunbing, Tong Dongbing, et al. Short-term photovoltaic power prediction based on VMD and improved TCN[J]. Electronic Science and Technology, 2023, 36(3):42-49.
[5]	叶兴, 薛家祥. 改进型LSTM网络光伏发电功率预测研究[J]. 中国测试, 2019, 45(11):14-20.
	Ye Xing, Xue Jiaxiang. Research on photovoltaic power generation prediction based on improved LSTM network[J]. China Measurement and Test, 2019, 45(11):14-20.
[6]	董雪, 赵宏伟, 赵生校, 等. 基于SOM聚类和二次分解的BiGRU超短期光伏功率预测[J]. 太阳能学报, 2022, 43(11):85-93. doi: 10.19912/j.0254-0096.tynxb.2021-0518
	Dong Xue, Zhao Hongwei, Zhao Shengxiao, et al. Ultra-short-term forecasting method of photovoltaic power based on SOM clustering,secondary decomposition and BiGRU[J]. Acta Energiae Solaris Sinica, 2022, 43(11):85-93. doi: 10.19912/j.0254-0096.tynxb.2021-0518
[7]	Fang X, Han S H, Li Juan, et al. A FCM-XGBoost-GRU model for short-term photovoltaic power forecasting based on weather classification[C]. Chengdu: The Fifth Asia Energy and Electrical Engineering Symposium, 2023:1444-1449.
[8]	冯先丁, 魏镜弢, 吴张永, 等. 基于PCA-PSO-SVM的球磨机负荷预测研究[J]. 电子科技, 2022, 35(1):29-34.
	Feng Xianding, Wei Jingtao, Wu Zhangyong, et al. Research on load forecast of ball mill based on PCA-PSO-SVM[J]. Electronic Science and Technology, 2022, 35(1): 29-34.
[9]	Lin P J, Peng Z N, Lai Y F, et al. Short-term power prediction for photovoltaic power plants using a hybrid improved K-means-GRA-Elman model based on multivariate meteorological factors and historical power datasets[J]. Energy Conversion and Management, 2018, 177(1):704-717.
[10]	吕伟杰, 方一帆, 程泽. 基于模糊C均值聚类和样本加权卷积神经网络的日前光伏出力预测研究[J]. 电网技术, 2022, 46(1):231-238.
	Lü Weijie, Fang Yifan, Cheng Ze. Research of day-ahead photovoltaic output based on FCM-WA-CNN[J]. Power System Technology, 2022, 46(1):231-238.
[11]	王桢, 王海祥. 不同天气下光伏电池全日输出电压与功率研究[J]. 能源与节能, 2020, 25(8):51-54,66.
	Wang Zhen, Wang Haixiang. Study on full-day output voltage and power of photovoltaic cells under different weather conditions[J]. Energy and Energy Conservation, 2020, 25(8):51-54,66.
[12]	任建吉, 位慧慧, 邹卓霖, 等. 基于CNN-BiLSTM-Attention的超短期电力负荷预测[J]. 电力系统保护与控制, 2022, 50(8):109-116.
	Ren Jianji, Wei Huihui, Zou Zhuolin, et al. Ultra-short-term power load forecasting based on CNN-BiLSTM-Attention[J]. Power System Protection and Control, 2022, 50(8):109-116.
[13]	李荣. 利用卷积神经网络的显著性区域预测方法[J]. 重庆邮电大学学报(自然科学版), 2019, 31(1):37-43.
	Li Rong. A significant regional prediction method using convolutional neural network[J]. Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), 2019, 31(1):37-43.
[14]	曾囿钧, 肖先勇, 徐方维, 等. 基于CNN-BiGRU-NN模型的短期负荷预测方法[J]. 中国电力, 2021, 54(9):17-23.
	Zeng Youjun, Xiao Xianyong, Xu Fangwei, et al. Short-term load forecasting method based on CNN-BiGRU-NN model[J]. Electric Power, 2021, 54(9):17-23.
[15]	王童. 基于MSWOA改进Attention-BiGRU模型的电力负荷预测[J]. 软件导刊, 2023, 22(10):84-89.
	Wang Tong. Power load forecasting based on MSWOA improved Attention-BiGRU model[J]. Software Guide, 2023, 22(10):84-89.
[16]	常青松, 杨昭, 杨熠辉, 等. 基于相似日聚类的超短期光伏功率组合预测模型[J]. 热力发电, 2023, 52(11):123-131.
	Chang Qingsong, Yang Zhao, Yang Yihui, et al. Ultrashort term photovoltaic power combinatorial forecasting model based on similar day clustering[J]. Thermal Power Generation, 2023, 52(11):123-131.
[17]	叶林, 宫婷, 宋旭日, 等. 基于波动类型精细划分与聚类的短期负荷预测[J]. 电网技术, 2023, 47(3):998-1013.
	Ye Lin, Gong Ting, Song Xuri, et al. Short-term load forecasting based on fine division and clustering of fluctuation types[J]. Power System Technology, 2023, 47(3):998-1013.
[18]	孙辉, 杨帆, 高正男, 等. 考虑特征重要性值波动的MI-BILSTM短期负荷预测[J]. 电力系统自动化, 2022, 46(8):95-103.
	Sun Hui, Yang Fan, Gao Zhengnan, et al. Short-term load forecasting based on mutual information and Bi-direction long short-term memory network considering fluctuation in importance values of features[J]. Automation of Electric Power Systems, 2022, 46(8):95-103.
[19]	陈美珍. 基于聚类算法及组合模型的光伏发电功率短期预测[D]. 福州: 福建工程学院, 2022:1-68.
	Chen Meizhen. Short-term prediction of photovoltaic power generation based on clustering algorithm and combined model[D]. Fuzhou: Fujian University of Technology, 2022:1-68.
[20]	胡兵, 詹仲强, 陈洁, 等. 基于PCA-GA-Elman的短期光伏出力预测研究[J]. 太阳能学报, 2020, 41(6):256-263.
	Hu Bing, Zhan Zhongqiang, Chen Jie, et al. Prediction research on short-term photovoltaic output based on PCA-GA-Elman[J]. Acta Energiae Solaris Sinica, 2020, 41(6):256-263.

变量序号	气象变量	MIC相关系数值
1	直接辐射	0.92
2	相对湿度	0.46
3	环境气温	0.36
4	散射辐射	0.33
5	风速	0.23
6	风向	0.17

评分指标	晴天			多云及阴雨天气			极端天气
评分指标	RMSE/kW	MAPE/%	R²/%	RMSE/kW	MAPE/%	R²/%	RMSE/kW	MAPE/%	R²/%
SVR	2.832 5	12.40	92.22	5.465 0	30.23	76.95	4.198 3	32.31	62.59
BP	1.478 4	6.14	97.87	3.905 7	16.02	88.23	4.042 0	25.64	65.32
GRU	1.294 1	4.62	98.37	3.670 4	15.42	89.60	3.859 8	23.33	68.38
CNN-BiGRU	1.311 8	5.86	98.32	3.577 4	14.93	90.12	3.665 4	21.68	71.48
CNN-LSTM	1.427 5	5.50	98.02	3.648 7	18.12	89.72	3.874 2	22.84	68.14
双通道CNN-LSTM	1.326 2	4.99	98.29	3.547 0	15.40	90.29	3.706 3	21.67	70.84
双通道CNN-BiGRU	1.230 4	4.61	98.53	3.533 3	15.24	90.36	3.654 9	21.01	71.65
双通道CNN-BiGRU-ATT	1.149 9	4.16	98.71	3.494 2	14.17	90.57	3.408 6	19.93	75.34

评分指标	晴天		多云及阴雨天气		极端天气
评分指标	RMSE/kW	MAPE/%	RMSE/kW	MAPE/%	RMSE/kW	MAPE/%
SVR	3.229 3	12.43	6.693 6	46.21	10.588 9	75.14
BP	3.398 1	16.17	6.379 9	35.70	9.796 6	64.38
双通道CNN-BiGRU-ATT	2.018 0	7.65	3.538 0	14.89	9.204 5	59.85