基于WSO-LSTM的風電功率預測技術研究
滕云雷,李桓
(國網山東省電力公司臨沂供電公司,山東 臨沂 276000)
摘 要 :為了確保電力系統的可靠運行與持續供電,準確的風電功率預測顯得尤為重要。提出了一種新的白鯊優化算法-長短期記憶網絡 (WSO-LSTM) 模型,用于短期風電功率的預測 ;利用LSTM在自動學習 序列數據特征方面的優勢,同時借助WSO的全局優化策略對LSTM層的窗口大小及神經元數量進行優化。 通過標準性能指標,將WSO-LSTM的預測結果與實際功率以及現有模型的預測結果進行了對比,結果表明, WSO-LSTM能夠為歐洲 4 個風電場提供準確、可靠且穩健的風電功率預測,預測精度平均提升了 20%~47%。
關鍵詞 : 白鯊優化算法 ;長短期記憶網絡 ;風電功率預測 ;機器學習 ;特征提取
中圖分類號 :TM614 ;TM715 文獻標識碼 :A 文章編號 :1007-3175(2025)12-0022-07
The Research on Wind Power Prediction Technology Based on WSO-LSTM
TENG Yun-lei, LI Huan
(State Grid Shandong Electric Power Company Linyi Power Supply Company, Linyi 276000, China)
Abstract: To ensure the reliable operation and continuous power supply of the power system, accurate wind power prediction is particularly important. This paper proposes a novel white shark optimization algorithm-long short-term memory network (WSO-LSTM) model for short-term wind power prediction. By taking advantage of the strengths of LSTM in automatically learning the features of sequential data, and with the help of the global optimization strategy of WSO, the window size and the number of neurons of the LSTM layer are optimized. Through standard performance indicators, the prediction results of WSO-LSTM were compared with the actual power and the prediction results of existing models. The results show that WSO-LSTM can provide accurate, reliable and robust wind power prediction for four wind farms in Europe, achieving an average improvement in prediction accuracy ranging from 20% to 47%.
Key words: white shark optimization algorithm; long short-term memory network; wind power prediction; machine learning; feature extraction
參考文獻
[1] 楊群力,潘學萍,顧晨,等 . 風電機組仿真模型準確性評估試驗與方法 [J]. 電工電氣,2024(11) :1-7.
[2] 林海東,匡洪海,王俊,等 . 永磁直驅風電低電壓穿越與功率平抑策略研究 [J]. 電工電氣,2024(11) :8-14.
[3] 蘇向敬,程子凡,聶良釗,等 . 基于AGCN-LSTM模型的海上風電場功率概率預測 [J]. 電力系統自動化, 2024,48(22) :140-149.
[4] FOLEY A M, LEAHY P G, MARVUGIIA A, et al. Current methods and advances in forecasting of wind power generation[J].Renewable Energy,2012,37(1) :1-8.
[5] 盧雪平,董存,王錚,等 . 低溫寒潮天氣下的風電短期功率預測技術研究 [J] . 電網技術,2024,48(12) : 4833-4843.
[6] MARCIUKAITIS M, KATINAS V, KAVALIAUSKAS A. Wind power usage and predictionprospects in Lithuania[J].Renewable and Sustainable Energy Reviews,2008,12(1) :265-277.
[7] LOUKA P, GALANIS G, SIEBERT N, et al. Improvements in wind speed forecasts for wind power prediction purposes using Kalman filtering[J].Journal of Wind Engineering and Industrial Aerodynamics:The Journal of the International Association for Wind Engineering, 2008,96(12) :2348-2362.
[8] NIELSEN T S, MADSEN H, NIELSEN H A, et al. Zephyr-The Prediction Models[M].Copenhagen: WIP-Renewable Energies/ETA,2006.
[9] 林鵬 . 基于隨機模糊理論的風電功率預測 [D]. 北京 : 華北電力大學,2015.
[10] 王欣,李勝剛,秦斌,等 . 基于模糊支持向量機的風 電場功率預測 [J]. 新型工業化,2014,4(9) :50-55.
[11] 周松林,茆美琴,蘇建徽 . 基于主成分分析與人工神經網絡的風電功率預測 [J]. 電網技術,2011, 35(9) :128-132.
[12] 張智峰 . 基于改進 FCM與模糊馬爾可夫鏈的風電功率 短期預測方法研究 [D]. 銀川 :北方民族大學,2021.
[13] 繆銘狄 . 區間二型 TSK模糊邏輯系統在風電功率預測中的應用 [D]. 蘭州 :蘭州交通大學,2024.