针对传统风电功率预测方法在处理复杂信号时的局限性,提出一种结合集成分解与多模型融合的风电功率预测新方法。首先,采用融合混合策略的混沌-莱维黏菌（SCL）算法优化改进的完全自适应噪声集合经验模态分解方法（ICEEMDAN）,应用于风电功率数据的频率分解,将信号分解为高频、中频和低频分量。针对不同频率的分量,分别应用Transformer、TCN-BIGRU和OSELM进行单独建模,以充分捕捉不同频率分量的时序特征和非线性规律。随后利用多头注意力机制对各频率分量的预测结果进行加权融合,充分考虑不同分量对风电功率预测的贡献。为获得更具鲁棒性的风电功率预测结果,利用改进的马尔可夫链蒙特卡洛（IMCMC）方法对预测结果进行区间预测。实验结果表明,所提出的方法相比于传统单一模型预测方法,R2分别提高了14.2% 与7.15%,在95%置信区间PICP分别为95.78%与95.93%。
Abstract
Aiming at the limitations of traditional wind power prediction methods in dealing with complex signals, a new method for wind power prediction combining integrated decomposition and multi-model fusion is proposed. First, a Slime Chaos Levy （SCL） algorithm incorporating a hybrid strategy is proposed for optimizing the Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise （ICEEMDAN） method, which is applied to the frequency decomposition of wind power data to decompose the signals into high, medium and low frequency components. Transformer, temporal convolutional network and bidirectional gated recurrent unit （TCN-BIGRU）, and online sequential extreme learning machine （OSELM） are applied to represent different frequency components in order to completely capture their temporal features and nonlinear patterns. Subsequently, to fully account for the contribution of various components to the wind power forecast, the prediction results of each frequency component are then weighted and fused using the multi-head attention mechanism. In order to obtain more robust wind power prediction results, the improved Markov chain Monte Carlo （IMCMC） method is utilized to predict the prediction results in intervals. The experimental findings demonstrate that the target method increases the R2 by 14.2% and 7.15%, respectively, and the PICP is 95.78% and 95.93% at 95% confidence intervals, respectively, when compared to the typical single-model prediction method.
关键词
风电功率 /
优化算法 /
分解方法 /
多头注意力机制 /
区间预测
Key words
wind power /
optimization algorithms /
decomposition methods /
multi head attention mechanism /
interval prediction
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 国家能源局. 全国电力工业统计数据[EB/OL].(202512-26)[20260428]. https://www.nea.gov.cn20251226/640306962d7d421b921b902f48a04b47/c.html.
NEA. National electric power industry statistics[EB/OL]. (20251226)[20260428]. https://www.nea.gov.cn20251226/640306962d7d421b921b902f48a04b47/c.html.
[2] 国家发展和改革委员会, 国家能源局. “十四五”现代能源体系规划[EB/OL]. (2022-01-29)[2024-12-28]. https://www.gov.cn/zhengce/zhengceku/2022-03/23/5680759/files/ccc7dffca8f24880a80af12755558f4a.pdf
NDRC, NEA. The 14th Five-Year Plan for Modern Energy System[EB/OL]. (2022-01-29)[2024-12-28]. https://www.gov.cn/zhengce/zhengceku/2022-03/23/5680759/files/ccc7dffca8f24880a80af12755558f4a.pdf.
[3] 冯双磊, 王伟胜, 刘纯, 等. 风电场功率预测物理方法研究[J]. 中国电机工程学报, 2010, 30(2): 1-6.
FENG S L, WANG W S, LIU C, et al.Study on the physical approach to wind power prediction[J]. Proceedings of the CSEE, 2010, 30(2): 1-6.
[4] YUNUS K, THIRINGER T, CHEN P Y.ARIMA-based frequency-decomposed modeling of wind speed time series[J]. IEEE transactions on power systems, 2016, 31(4): 2546-2556.
[5] TRIZOGLOU P, LIU X L, LIN Z.Fault detection by an ensemble framework of Extreme Gradient Boosting (XGBoost) in the operation of offshore wind turbines[J]. Renewable energy, 2021, 179: 945-962.
[6] 马志侠, 张林鍹, 巴音塔娜, 等. 基于自适应二次分解与CNN-BiLSTM的超短期风电功率预测[J]. 太阳能学报, 2024, 45(6): 429-435.
MA Z X, ZHANG L X, BA Y, et al.Ultra-short-term wind power prediction based on adaptive quadratic mode decomposition and cnn-BiLSTM[J]. Acta energiae solaris sinica, 2024, 45(6): 429-435.
[7] 钟吴君, 李培强, 涂春鸣. 基于尾流关联的动态超图风电功率超短期预测方法[J]. 中国电机工程学报, 2025, 45(12): 4693-4706.
ZHONG W J, LI P Q, TU C M.Dynamic hypergraph wind power ultra short term prediction method based on wake correlation[J]. Proceedings of the CSEE, 2025, 45(12): 4693-4706.
[8] 徐武, 范鑫豪, 沈智方, 等. 多尺度特征提取的Transformer短期风电功率预测[J]. 太阳能学报, 2025, 46(2): 640-648.
XU W, FAN X H, SHEN Z F, et al.Short-term wind power prediction using transformer with multi-scale feature extraction[J]. Acta energiae solaris sinica, 2025, 46(2): 640-648.
[9] 赵卿, 高文华, 石慧, 等. 基于WT和黏菌算法的LSSVM短期风功率预测[J]. 计算机仿真, 2024, 41(7): 166-170, 226.
ZHAO Q, GAO W H, SHI H, et al.Short-term wind power prediction using LSSVM based on WT and slime mould algorithm[J]. Computer simulation, 2024, 41(7): 166-170, 226.
[10] 陈运蓬, 景超, 白静波, 等. 基于集成学习的新能源发电功率预测[J]. 太阳能学报, 2024, 45(6): 412-421.
CHEN Y P, JING C, BAI J B, et al.New energy power forecasting based on ensemble learning[J]. Acta energiae solaris sinica, 2024, 45(6): 412-421.
[11] 张昊立, 张菁, 倪建辉, 等. 引入注意力机制的LSTM-FCN海上风电功率预测[J]. 太阳能学报, 2024, 45(6): 444-450.
ZHANG H L, ZHANG J, NI J H, et al.Lstm-fcn offshore wind power forecasting with introduction of attention mechanism[J]. Acta energiae solaris sinica, 2024, 45(6): 444-450.
[12] 谷学静, 陈洪磊, 孙泽贤, 等. 基于VMD-RL-LSTM的短期风功率预测[J]. 计算机仿真, 2023, 40(4): 89-93, 309.
GU X J, CHEN H L, SUN Z X, et al.Short-term wind power prediction based on VMD-RL-LSTM[J]. Computer simulation, 2023, 40(4): 89-93, 309.
[13] 陈子含, 滕伟, 胥学峰, 等. 基于图卷积网络和风速差分拟合的中长期风功率预测[J]. 中国电力, 2023, 56(10): 96-105.
CHEN Z H, TENG W, XU X F, et al.Medium and long term wind power prediction based on graph convolutional network and wind velocity differential fitting[J]. Electric power, 2023, 56(10): 96-105.
[14] 荆志宇, 李培强, 林文婷. 结合贝叶斯优化及通道注意力的双端优化时序式风功率预测模型[J]. 电力系统及其自动化学报, 2024, 36(8): 39-47, 59.
JING Z Y, LI P Q, LIN W T.Double-side optimized time-series wind power prediction model combining Bayesian optimization and channel attention[J]. Proceedings of the CSU-EPSA, 2024, 36(8): 39-47, 59.
[15] LI S M, CHEN H L, WANG M J, et al.Slime mould algorithm: a new method for stochastic optimization[J]. Future generation computer systems, 2020, 111: 300-323.
[16] GEISEL T, FAIREN V.Statistical properties of chaos in Chebyshev maps[J]. Physics letters A, 1984, 105(6): 263-266.
[17] 程彦琳, 李书琴. 基于混沌映射和莱维飞行扰动的蛇形优化算法[J]. 计算机工程与设计, 2024, 45(9): 2658-2668.
CHENG Y L, LI S Q.Snake optimization algorithm based on chaotic reverse and Levy flight[J]. Computer engineering and design, 2024, 45(9): 2658-2668.
[18] BOMMIDI B S, TEEPARTHI K, KOSANA V.Hybrid wind speed forecasting using ICEEMDAN and transformer model with novel loss function[J]. Energy, 2023, 265: 126383.
[19] YU D J, CHENG J S, YANG Y.Application of EMD method and Hilbert spectrum to the fault diagnosis of roller bearings[J]. Mechanical systems and signal processing, 2005, 19(2): 259-270.
[20] VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017, 30: 5998-6008.
[21] ZHU R J, LIAO W L, WANG Y S.Short-term prediction for wind power based on temporal convolutional network[J]. Energy reports, 2020, 6: 424-429.
[22] 盛四清, 金航, 刘长荣. 基于VMD-WSGRU的风电场发电功率中短期及短期预测[J]. 电网技术, 2022, 46(3): 897-904.
SHENG S Q, JIN H, LIU C R.Short-term and mid-short-term wind power forecasting based on VMD-WSGRU[J]. Power system technology, 2022, 46(3): 897-904.
[23] XU M J, WANG K, LI M, et al.An adaptive on-board real-time model with residual online learning for gas turbine engines using adaptive memory online sequential extreme learning machine[J]. Aerospace science and technology, 2023, 141: 108513.
[24] ZHANG Y G, ZHAO Y P, SHEN X Y, et al.A comprehensive wind speed prediction system based on Monte Carlo and artificial intelligence algorithms[J]. Applied energy, 2022, 305: 117815.
基金
广国家自然科学基金联合基金项目（U22B6006）; 中央高校基本科研业务费学科交叉创新专项项目（2023JC006）
PDF(2450 KB)
