基于聚类和混合提取的BiGRU-MDSA光伏功率短期预测

周育才; 秦缘恒; 肖振江; 谢七月; 付强; 谭艳祥

doi:10.19912/j.0254-0096.tynxb.2025-0729

PDF(1613 KB)

太阳能学报 ›› 2026, Vol. 47 ›› Issue (8) : 244-251. DOI: 10.19912/j.0254-0096.tynxb.2025-0729

基于聚类和混合提取的BiGRU-MDSA光伏功率短期预测

周育才^1,2, 秦缘恒^1,3, 肖振江², 谢七月¹, 付强¹, 谭艳祥⁴

作者信息 +

SHORT-TERM PHOTOVOLTAIC POWER FORECASTING BASED ON CLUSTERING AND HYBRID FEATURE EXTRACTION USING BiGRU-MDSA

Zhou Yucai^1,2, Qin Yuanheng^1,3, Xiao Zhenjiang², Xie Qiyue¹, Fu Qiang¹, Tan Yanxiang⁴

Author information +

文章历史 +

摘要

为解决光伏发电功率的波动性和不确定性,提出一种基于模糊C均值（FCM）聚类算法和混合尺度特征提取（HSFE）,多头动态稀疏注意力机制（MDSA）的双向门控循环单元（BIGRU）光伏功率短期预测。首先对历史光伏数据进行预处理和异常值分析,对在不同气候下光伏功率输出的影响因素进行相关性分析,然后通过降维对FCM聚类方法对输入数据进行天气分类聚类,再将分类数据作为输入,在BiGRU模型的基础上增加混合尺度特征提取模型,加强提取不同时间尺度的信息能力;然后加入多头动态稀疏自注意力机制构建组合模型,动态调整模型对各时间步数据特征的聚焦程度。最终通过组合模型仿真结果对比试验表明,此方法具有更高的精准度和泛化能力。

Abstract

To address the volatility and uncertainty of photovoltaic （PV） power generation, this study proposes a short-term PV power forecasting method based on fuzzy C-means （FCM） clustering, hybrid scale feature extraction （HSFE）, and multi-head dynamic sparse attention （MDSA） mechanism integrated into a bidirectional gated recurrent unit （BiGRU） framework. First, historical PV data undergo preprocessing and outlier analysis, followed by a correlation analysis of the factors influencing PV output under different weather conditions. Then, dimensionality reduction is performed, and FCM is used to classify the input data based on weather conditions. The clustered data are fed into a BiGRU model enhanced with the HSFE module to enhance its capability in extracting information across different temporal scales. Furthermore, the multi-head dynamic sparse attention mechanism is incorporated to dynamically adjust the model’s focus on temporal features across time steps. Simulation results and comparative experiments demonstrate that the proposed composite model achieves superior accuracy and generalization performance compared to benchmark methods.

导出引用

周育才, 秦缘恒, 肖振江, 谢七月, 付强, 谭艳祥. 基于聚类和混合提取的BiGRU-MDSA光伏功率短期预测[J]. 太阳能学报. 2026, 47(8): 244-251 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0729

Zhou Yucai, Qin Yuanheng, Xiao Zhenjiang, Xie Qiyue, Fu Qiang, Tan Yanxiang. SHORT-TERM PHOTOVOLTAIC POWER FORECASTING BASED ON CLUSTERING AND HYBRID FEATURE EXTRACTION USING BiGRU-MDSA[J]. Acta Energiae Solaris Sinica. 2026, 47(8): 244-251 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0729

中图分类号： TM615

参考文献

[1] 韩梦瑶, 熊焦, 刘卫东. 中国光伏发电的时空分布、竞争格局及减排效益[J]. 自然资源学报, 2022, 37(5): 1338-1351.
Han M Y, Xiong J, Liu W D.Spatio-temporal distribution, competitive development and emission reduction of China’s photovoltaic power generation[J]. Journal of Natural Resources, 2022, 37(5): 1338-1351.
[2] 王宏伟, 朱雪婷, 殷晨曦. 中国光伏产业发展及电价补贴政策影响研究[J]. 数量经济技术经济研究, 2022, 39(7): 90-112.
Wang H W, Zhu X T, Yin C X.Research on the development and the feed-in tariff policy impact of China’s photovoltaic industry[J]. Journal of Quantitative & Technical Economics, 2022, 39(7): 90-112.
[3] 陈艳, 周园媛, 纪雅星. 产业政策对企业绩效的影响及作用机制研究: 来自中国光伏产业的经验数据[J]. 科技进步与对策, 2021, 38(22): 68-75.
Chen Y, Zhou Y Y, Ji Y X.A research on the impact of industrial policy on enterprise performance and its mechanism: empirical data from China’s photovoltaic industry[J]. Science & Technology Progress and Policy, 2021, 38(22): 68-75.
[4] Armeanu D S, Joldes C C, Gherghina S C, et al.Understanding the multidimensional linkages among renewable energy, pollution, economic growth and urbanization in contemporary economies: Quantitative assessments across different income countries’ groups[J]. Renewable and Sustainable Energy Reviews, 2021, 142: 110818.
[5] Swilling M, Nygaard I, Kruger W, et al.Linking the energy transition and economic development: a framework for analysis of energy transitions in the Global South[J]. Energy Research & Social Science, 2022, 90: 102567.
[6] Sheraz M, Xu D Y, Sinha A, et al.The dynamic nexus among financial development, renewable energy and carbon emissions: moderating roles of globalization and institutional quality across BRI countries[J]. Journal of Cleaner Production, 2022, 343: 130995.
[7] Choudhary P, Srivastava R K.Sustainability perspectives- a review for solar photovoltaic trends and growth opportunities[J]. Journal of Cleaner Production, 2019, 227: 589-612.
[8] Ahmad E, Khan D, Anser M K, et al.The influence of grid connectivity, electricity pricing, policy-driven power incentives, and carbon emissions on renewable energy adoption: exploring key factors[J]. Renewable Energy, 2024, 232: 121108.
[9] 周孝信, 陈树勇, 鲁宗相. 电网和电网技术发展的回顾与展望: 试论三代电网[J]. 中国电机工程学报, 2013, 33(22): 1-11, 22.
Zhou X X, Chen S Y, Lu Z X.Review and prospect for power system development and related technologies: a concept of three-generation power systems[J]. Proceedings of the CSEE, 2013, 33(22): 1-11, 22.
[10] 门献伟, 郑晓亮. 基于深度学习的光伏发电功率短期预测[J]. 兰州文理学院学报(自然科学版), 2023, 37(6): 55-59.
Men X W, Zheng X L.Short-term prediction of photovoltaic power generation power based on deep learning[J]. Journal of LanZhou University of Arts and Science (Natural Sciences Edition), 2023, 37(6): 55-59.
[11] 刘振路, 郭军红, 李薇, 等. 基于FCM-LSTM的光热发电出力短期预测[J]. 工程科学学报, 2024, 46(1): 178-186.
Liu Z L, Guo J H, Li W, et al.Short-term prediction of concentrating solar power based on FCM-LSTM[J]. Chinese Journal of Engineering, 2024, 46(1): 178-186.
[12] 黄博阳, 何肖蒙, 肖小兵, 等. 基于FCM-WS-BP的光伏日前出力预测研究[J]. 控制工程, 2023, 30(12): 2254-2260.
Huang B Y, He X M, Xiao X B, et al.Research of photovoltaic day-ahead output power prediction based on FCM-WS-BP[J]. Control Engineering of China, 2023, 30(12): 2254-2260.
[13] 周浩杰, 杨建卫, 王尊, 等. 基于LSTM 光伏发电功率超短期预测模型研究[J]. 电源技术, 2023, 47(6): 785-789.
Zhou H J, Yang J W, Wang Z, et al.Research on ultra-short term prediction model of photovoltaic power generation based on LSTM[J]. Chinese Journal of Power Sources, 2023, 47(6): 785-789.
[14] 徐先峰, 刘阿慧, 陈雨露, 等. 基于气象因素充分挖掘的BiLSTM光伏发电短期功率预测[J]. 计算机系统应用, 2020, 29(7): 205-211.
Xu X F, Liu A H, Chen Y L, et al.BiLSTM short-term forecasting method for photovoltaic power generation based on fully exploiting meteorological factors[J]. Computer Systems & Applications, 2020, 29(7): 205-211.
[15] 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.
Dong X, Zhao H W, Zhao S X, 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.
[16] 刘刚, 闵金, 宋伟, 等. 基于K-Means和GBRT的分布式光伏中短期发电量预测[J]. 能源与环保, 2023, 45(3): 210-215, 221.
Liu G, Min J, Song W, et al.Medium-short term power generation prediction of distributed photovoltaic based on K-Means and GBRT[J]. China Energy and Environmental Protection, 2023, 45(3): 210-215, 221.
[17] 吴家葆, 曾国辉, 张振华, 等. 基于K-means分层聚类的TCN-GRU和LSTM动态组合光伏短期功率预测[J]. 可再生能源, 2023, 41(8): 1015-1022.
Wu J B, Zeng G H, Zhang Z H, et al.Dynamic combination of TCN-GRU and LSTM photovoltaic short-term power prediction based on K-means hierarchical clustering[J]. Renewable Energy Resources, 2023, 41(8): 1015-1022.
[18] 王瑞, 张璐婷, 逯静. 基于新型相似日选取和VMD-NGO-BiGRU的短期光伏功率预测[J]. 湖南大学学报(自然科学版), 2024, 51(2): 68-80.
Wang R, Zhang L T, Lu J.Short term photovoltaic power prediction based on new similar day selection and VMD-NGO-BiGRU[J]. Journal of Hunan University (Natural Sciences), 2024, 51(2): 68-80.
[19] 李青, 张新燕, 马天娇, 等. 基于SSA-CNN-BiGRU-Attention的超短期风电功率预测模型[J]. 电机与控制应用, 2023, 50(5): 61-71.
Li Q, Zhang X Y, Ma T J, et al.Ultra-short term forecasting model of wind power based on SSA-CNN-BiGRU-attention[J]. Electric Machines & Control Application, 2023, 50(5): 61-71.
[20] 王家乐, 张耀, 林帆, 等. 基于自注意力特征提取的光伏功率组合概率预测[J]. 太阳能学报, 2024, 45(12): 123-131.
Wang J L, Zhang Y, Lin F, et al.Combining probabilistic prediction of PV power based on self-attention feature extraction mechanism[J]. Acta Energiae Solaris Sinica, 2024, 45(12): 123-131.