| 220 | 1 | 217 |
| 下载次数 | 被引频次 | 阅读次数 |
准确预测光伏发电功率对于可再生能源的利用至关重要。现有很多模型难以有效捕捉目标变量和协变量之间的复杂关系，且对时间动态和多变量数据之间的相互作用捕捉不精准。因此，提出一种新的模型架构，利用iTransformer和双向门控循环单元（BiGRU）从中提取特征，对于模型的融合输出，通过整合多头注意力机制和柯尔莫哥洛夫-阿诺德网络映射来增强表征能力。利用公开数据集对模型的有效性进行验证，结果表明，该模型能有效捕捉光伏发电的变化，其中春季指标的提升效果最优，相较iTransformer模型预测结果的平均绝对误差下降了36.8%，均方根误差下降了29.8%。
Abstract：Accurate prediction of photovoltaic(PV) generation is critical for renewable energy utilization. Many existing models struggle to effectively capture the complex relationships between target variables and covariates, and fail to accurately capture temporal dynamics and interactions between multivariate data. Therefore, a new model architecture was proposed. This architecture utilizes iTransformer and bi-directional gated recurrent units to extract features from them, and for the fused output of the model, the characterization was enhanced by integrating the multi-head attention mechanism and Kolmogorov-Arnold network mapping. The model's effectiveness was validated using publicly available datasets. The results demonstrate that the proposed model effectively captures variations in photovoltaic(PV) power generation, among which the improvement effect of the spring indicators is the most optimal. Compared to the iTransformer model's predictions, the mean absolute error(MAE) was reduced by 36. 8% and the root mean square error(RMSE) by 29. 8%.
[1]王其良，周恩泽，屠丽娟，等.基于太阳能补热的多源互补供暖系统优化研究[J].太阳能学报，2021,42(11):178-185.WANG Q L, ZHOU E Z, TU L J, et al. Optimization research on multi-source complementary heating system based on solar heating[J]. Acta Energiae Solaris Sinica, 2021, 42(11):178-185.(in Chinese)
[2] BAI S J, KOLTER J Z, KOLTUN V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[EB/OL].[2025-06-26]. https：∥arxiv. org/abs/1803. 01271.
[3] HOCHREITER S, SCHMIDHUBER J. Long shortterm memory[J]. Neural Computation, 1997, 9(8):1735-1780.
[4] CHO K, MERRIENBOER B V, GULCEHRE C,et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[C]∥Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing(EMNLP).Doha:Association for Computational Linguistics,2014:1724-1734.
[5]郝建华.基于深度学习的时间序列预测算法研究[D].济南：山东师范大学，2024:2-7.HAO J H. Research on time series forecasting algorithms based on deep learning[D]. Jinan:Shandong Normal University, 2024:2-7.(in Chinese)
[6] ZHOU H Y, ZHANG S H, PENG J Q, et al.Informer:beyond efficient transformer for long sequence time-series forecasting[C]∥Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto, California:AAAI Press, 2021:11106-11115.
[7] LIM B, AR?K S?，LOEFF N, et al. Temporal Fusion Transformers for interpretable multi-horizon time series forecasting[J]. International Journal of Forecasting, 2021, 37(4):1748-1764.
[8] WU H X, XU J H, WANG J M, et al. Autoformer:decomposition transformers with auto-correlation for long-term series forecasting[C]∥Advances in Neural Information Processing Systems 34(NeurIPS 2021).Red Hook, New York:Curran Associates, Inc.,2021:22419-22430.
[9] NIE Y, NGUYEN N H, SINTHONG P, et al. A time series is worth 64 words:long-term forecasting with transformers[EB/OL].[2025-06-26]. https：∥arxiv. org/abs/2211. 14730.
[10] SHUAI H, LI F X. Physics-informed KolmogorovArnold networks for power system dynamics[EB/OL].[2024-09-05]. https：∥arxiv. org/abs/2408.06650v1.
[11]谭海旺，杨启亮，邢建春，等.基于XGBoost-LSTM组合模型的光伏发电功率预测[J].太阳能学报，2022, 43(8):75-81.TAN H W, YANG Q L, XING J C, et al. Photovoltaic power prediction based on combined XGBoost-lstm model[J]. Acta Energiae Solaris Sinica, 2022,43(8):75-81.(in Chinese)
[12]孙跃飞.基于深度学习的智能电网用户异常用电检测方法研究[D].恩施：湖北民族大学，2023:10-11.SUN Y F. Research on detection method of abnormal power consumption of smart grid users based on deep learning[D]. Enshi:Hubei Minzu University, 2023:10-11.(in Chinese)
[13]吕玉坤，赵润一，周庆文，等.环境因素对光伏组件的温度及输出特性影响研究[J].太阳能学报，2024,45(11):194-203.LüY K, ZHAO R Y, ZHOU Q W, et al. Investigation on influence of environmental factors on temperature and output characteristics of solar PV modules[J].Acta Energiae Solaris Sinica, 2024, 45(11):194-203.(in Chinese)
[14]黄旭锐，于丰源，杨波，等.基于Transformer网络和多任务学习的园区综合能源系统电-热短期负荷预测方法[J].南方电网技术，2023, 17(1):152-160.HUANG X R, YU F Y, YANG B, et al. Short-term electric-thermal load forecasting method for park-level integrated energy system based on transformer network and multi-task learning[J]. Southern Power System Technology, 2023, 17(1):152-160.(in Chinese)
[15] LIU Y, HU T G, ZHANG H R, et al. iTransformer:inverted transformers are effective for time series forecasting[EB/OL].[2025-06-26]. https：∥arxiv. org/html/2310. 06625v3.
[16] LIU J, YANG F, YAN K, et al. Household energy consumption forecasting based on adaptive signal decomposition enhanced iTransformer network[J].Energy and Buildings, 2024, 324:114894.
[17]张建辉，滕婕，李秀慧，等.基于气象特征和改进Transformer的光伏功率短期预测[J].热能动力工程，2024, 39(8):146-154.ZHANG J H, TENG J, LI X H, et al. Short-term prediction of photovoltaic power based on meteorological features and improved transformer[J]. Journal of Engineering for Thermal Energy and Power, 2024,39(8):146-154.(in Chinese)
[18]范士雄，李东琦，郭剑波，等.基于时变滤波经验模态分解-重构和独立自注意力机制的iTransformer超短期负荷预测方法[J].电网技术，2025, 49(6):2436-2445.FAN S X, LI D Q, GUO J B, et al. Ultra-short-term load forecasting method based on time-varying filter empirical mode decomposition-reconstruction and iTransformer with stand-alone self-attention mechanism[J]. Power System Technology, 2025, 49(6):2436-2445.(in Chinese)
[19]谢荣怡，梁春，林倍民，等.基于改进型Transformer模型的光伏出力实时估计方法研究[J].价值工程，2025, 44(1):143-145.XIE R Y, LIANG C, LIN B M, et al. Research on real-time estimation method of PV output based on improved Transformer model[J]. Value Engineering,2025, 44(1):143-145.(in Chinese)
[20]王丽娟，王彦骁，张坤宁，等.基于天气场景动态识别的CO-MHA-LSTM光伏发电功率预测[J/OL].南方电网技术，2025:1-13.[2025-06-26]. https：∥kns. cnki. net/kcms/detail/44. 1643. TK. 20250623.0911. 012. html.WANG L J, WANG Y X, ZHANG K N, et al.Power prediction of CO-MHA-LSTM photovoltaic generation based on dynamic identification of weather scenes[J/OL]. Southern Power System Technology,2025:1-13.[2025-06-26]. https：∥kns. cnki. net/kcms/detail/44. 1643. TK. 20250623. 0911. 012. html.(in Chinese)
基本信息:
中图分类号:TM615
引用信息:
[1]董慧,刘清惓,谷祥宇,等.基于iTransformer-BiGRU优化的超短期光伏功率预测[J].中国科技论文,2025,20(10):813-822.
基金信息:
国家自然科学基金资助项目(42275143)