为提高数值天气预报（NWP）预测风速的精确性,将NWP风速与实际风电场风速输入到全局搜索策略鲸鱼算法（GSWOA）优化的变分模态分解（VMD）进行分解。分解后的实际风速分量作为训练目标,对应的NWP风速分量则输入基于牛顿-拉夫逊优化算法-长短期记忆网络加注意力机制（NRBO-LSTM-Attention）模型,将输出的各分量线性叠加后替换原NWP风速。之后,通过孤立森林和Ransac算法等对修正后的NWP与风电场数据进行异常值清洗,最终输入NRBO-LSTM-Attention模型,用于预测未来功率。仿真结果表明：修正后的NWP风速更接近实际风速,评估指标平均绝对误差（MAE）和均方根误差（RMSE）分别降低11.45%和19.82%,R2提升31.24%;预测功率模型的性能更优,MAE和RMSE分别降低11.36%和10.43%,R2提升3.42%。
Abstract
To enhance the accuracy of wind speed prediction in numerical weather prediction （NWP）, both NWP-derived wind speeds and actual wind farm wind speeds are input into a variational mode decomposition （VMD） optimized by the global search strategy whale optimization algorithm （GSWOA） for decomposition. The decomposed actual wind speed components serve as training targets, while the corresponding NWP wind speed components are fed into the Newton-Raphson-based optimizer-long short-term memory with attention mechanism （NRBO-LSTM-Attention） model. The output components are linearly superimposed to replace the original NWP wind speeds. Subsequently, the corrected NWP data and wind farm data undergo outlier cleaning using algorithms such as isolation forest and Ransac. The processed data is finally input into the NRBO-LSTM-Attention model to predict future power output. Simulation results demonstrate that the corrected NWP wind speeds are closer to actual wind speeds, with evaluation metrics MAE and RMSE decreasing by 11.45% and 19.82%, respectively, and R2 increasing by 31.24%. Additionally, the power prediction model exhibits superior performance, with MAE and RMSE reduced by 11.36% and 10.43%, respectively, and R2 improved by 3.42%.
关键词
风电场 /
风速 /
变分模态分解 /
神经网络 /
牛顿-拉夫逊优化算法 /
注意力机制 /
功率预测
Key words
wind farm /
wind speed /
variational mode decomposition /
neural networks /
Newton-Raphson-based optimizer /
attention mechanism /
power forecasting
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 许晓艳, 马烁, 黄越辉, 等.大规模风电接入对电网静态电压稳定性影响机理研究[J]. 中国电力, 2015, 48(3): 139-143.
XU X Y, MA S, HUANG Y H, et al.Research on the influence mechanism of large-scale wind power access on the static voltage stability of power grid[J]. Electric power, 2015, 48(3): 139-143.
[2] WANG Y R, WANG D C, TANG Y.Clustered hybrid wind power prediction model based on ARMA, PSO-SVM, and clustering methods[J]. IEEE access, 2020, 8: 17071-17079.
[3] 兰昆, 吴战波, 赵泽妮, 等. 基于多因子的太阳辐照度预测方法研究进展[J]. 太阳能学报, 2024, 45(5): 593-601.
LAN K, WU Z B, ZHAO Z N, et al.Advance in prediction methods for solar irradiance based on multiple-factors[J]. Acta energiae solaris sinica, 2024, 45(5): 593-601.
[4] 赵泽妮, 云斯宁, 贾凌云, 等. 基于统计模型的短期风能预测方法研究进展[J]. 太阳能学报, 2022, 43(11): 224-234.
ZHAO Z N, YUN S N, JIA L Y, et al.Recent progress in short-term forecasting of wind energy based on statistical models[J]. Acta energiae solaris sinica, 2022, 43(11): 224-234.
[5] WANG D Y, CUI X W, NIU D X.Wind power forecasting based on LSTM improved by EMD-PCA-RF[J]. Sustainability, 2022, 14(12): 7307.
[6] WU B R, WANG L, ZENG Y R.Interpretable wind speed prediction with multivariate time series and temporal fusion transformers[J]. Energy, 2022, 252: 123990.
[7] GUO J X, YUN S N, MENG Y, et al.Prediction of heating and cooling loads based on light gradient boosting machine algorithms[J]. Building and environment, 2023, 236: 110252.
[8] SOWMYA R, PREMKUMAR M, JANGIR P.Newton-raphson-based optimizer: a new population-based metaheuristic algorithm for continuous optimization problems[J]. Engineering applications of artificial intelligence, 2024, 128: 107532.
[9] 贾凌云, 云斯宁, 赵泽妮, 等. 神经网络短期光伏发电预测的应用研究进展[J]. 太阳能学报, 2022, 43(12): 88-97.
JIA L Y, YUN S N, ZHAO Z N, et al.Recent progress of short-term forecasting of photovoltaic generation based on artificial neural networks[J]. Acta energiae solaris sinica, 2022, 43(12): 88-97.
[10] DENG X, SHAO H J, HU C L, et al.Wind power forecasting methods based on deep learning: a survey[J]. Computer modeling in engineering & sciences, 2020, 122(1): 273-301.
[11] MENG Y, YUN S N, ZHAO Z N, et al.Short-term electricity load forecasting based on a novel data preprocessing system and data reconstruction strategy[J]. Journal of building engineering, 2023, 77: 107432.
[12] ZHAO Z N, YUN S N, JIA L Y, et al.Hybrid VMD-CNN-GRU-based model for short-term forecasting of wind power considering spatio-temporal features[J]. Engineering applications of artificial intelligence, 2023, 121: 105982.
[13] 刘吉成, 朱玺瑞, 于晶.基于IWOA-SA-Elman神经网络的短期风电功率预测[J]. 太阳能学报, 2024, 45(1):143-150.
LIU J C, ZHU X R, YU J.Short-term wind power prediction based on IWOA-SA-Elman neural network[J]. Acta energiae solaris sinica, 2024, 45(1):143-150.
[14] SHAHID F, ZAMEER A, MUNEEB M.Predictions for COVID-19 with deep learning models of LSTM, GRU and Bi-LSTM[J]. Chaos, solitons & fractals, 2020, 140: 110212.
[15] 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 prediction with attention mechanism[J]. Acta energiae solaris sinica, 2024, 45(6):444-450.
[16] 史林军, 戴滔, 劳文洁, 等. 基于改进KNN算法的新能源发电单元运行状态识别[J]. 电力自动化设备, 2024, 44(5):65-72.
SHI L J, DAI T, LAO W J, et al.Identification of running status of renewable energy power generation units based on improved KNN algorithm[J]. Electric power automation equipment, 2024, 44(5): 65-72.
[17] BICA A M, CURILĂ POPESCU D.The Akima’s fitting method for quartic splines[J]. Journal of numerical analysis and approximation theory, 2022, 51(2): 155-166.
[18] 余长宏, 许孔豪, 张泽, 等. 基于分割点改进孤立森林的网络入侵检测方法[J]. 计算机工程, 2024, 50(6):148-156.
YU C H, XU K H, ZHANG Z, et al.Improving network intrusion detection method for isolated forest based on segmentation points[J]. Computer engineering, 2024, 50(6): 148-156.
[19] 王可, 贾松敏, 李秀智. 基于全概率更新的改进RANSAC算法[J]. 控制与决策, 2017, 32(3): 427-434.
WANG K, JIA S M, LI X Z.Improved RANSAC algorithm based on full probability update[J]. Control and decision, 2017, 32(3): 427-434.
[20] BAAK M, KOOPMAN R, SNOEK H, et al.A new correlation coefficient between categorical, ordinal and interval variables with Pearson characteristics[J]. Computational statistics & data analysis, 2020, 152: 107043.
[21] WAN A P, CHANG Q, AL-BUKHAITI K, et al.Short-term power load forecasting for combined heat and power using CNN-LSTM enhanced by attention mechanism[J]. Energy, 2023, 282: 128274.
[22] XU H, MANNOR S.Robustness and generalization[J]. Machine learning, 2012, 86(3): 391-423.
基金
国家自然科学基金（62373142）; 湖南省自然科学基金（2024JJ7135）; 湖南省教育厅科学研究项目（24B0524）
PDF(1719 KB)
