基于混合分位数回归长短期记忆神经网络的风电功率短期区间预测

杨茂, 张书天, 王勃, 于欣楠

太阳能学报 ›› 2025, Vol. 46 ›› Issue (2) : 582-590.

PDF(2521 KB)
欢迎访问《太阳能学报》官方网站,今天是
PDF(2521 KB)
太阳能学报 ›› 2025, Vol. 46 ›› Issue (2) : 582-590. DOI: 10.19912/j.0254-0096.tynxb.2023-1676

基于混合分位数回归长短期记忆神经网络的风电功率短期区间预测

  • 杨茂1, 张书天1, 王勃2, 于欣楠1
作者信息 +

SHORT-TERM WIND POWER INTERVAL PREDICTION BASED ON MIXED QUANTILE REGRESSION LONG AND SHORT-TERM MEMORYNEURAL NETWORK

  • Yang Mao1, Zhang Shutian1, Wang Bo2, Yu Xinnan1
Author information +
文章历史 +

摘要

为进一步提升风电功率区间预测精度,提出一种基于混合分位数回归长短期记忆神经网络的风电功率短期区间预测方法。通过同时考虑复合、平滑和非交叉3个特点对传统分位数回归模型进行改进,首先使用平滑函数代替弹球损失函数,使长短期记忆神经网络更易于拟合分位数回归模型。然后构建复合目标函数,使其能在给出多个分位数的条件下不重复训练多个独立模型。接着利用ReLU罚函数进行非交叉约束来避免分位数交叉现象的发生。最后将改进后的分位数回归与长短期记忆神经网络相结合并应用于中国甘肃省某风电场,运行结果表明所提模型在不同置信水平下对应PICP和PIAW分别提高了4.17个百分点和降低了2.31 MW,验证了方法的有效性。

Abstract

In order to further improve the accuracy of wind power interval prediction, a short-term interval prediction method of wind power based on mixed quantile regression long-term and short-term memory neural network is proposed. The traditional quantile regression model is improved by considering the three characteristics of compound, smoothing and non-crossing. Firstly, the smoothing function is used instead of the pinball loss function, which makes it easier for long-term and short-term memory neural networks to fit the quantile regression model. Then, a composite objective function is constructed so that it does not train multiple independent models repeatedly under the condition of giving multiple quantiles. thirdly, ReLU penalty function is used for non-cross constraint to avoid the occurrence of quantile crossing. Finally, the improved quantile regression is combined with the long-term and short-term memory neural network and applied to a wind farm in Gansu Province, China. The operation results show that the PICP and PIAW corresponding to thg proposed model increases the PICP increase by 4.17 percentage points and decrease by 2.31 MW respectively at different confidence levels, which verifies the effectiveness of the method.

关键词

风电功率 / 深度学习 / 区间预测 / 复合非交叉 / 分位数回归 / ReLU罚函数

Key words

wind power / deep learning / interval prediction / compound non-crossing / quantile regression / ReLU penalty function

引用本文

导出引用
杨茂, 张书天, 王勃, 于欣楠. 基于混合分位数回归长短期记忆神经网络的风电功率短期区间预测[J]. 太阳能学报. 2025, 46(2): 582-590 https://doi.org/10.19912/j.0254-0096.tynxb.2023-1676
Yang Mao, Zhang Shutian, Wang Bo, Yu Xinnan. SHORT-TERM WIND POWER INTERVAL PREDICTION BASED ON MIXED QUANTILE REGRESSION LONG AND SHORT-TERM MEMORYNEURAL NETWORK[J]. Acta Energiae Solaris Sinica. 2025, 46(2): 582-590 https://doi.org/10.19912/j.0254-0096.tynxb.2023-1676
中图分类号: TM614   

参考文献

[1] 国家能源局. 2022年一季度风电并网运行情况[EB/OL].https://wind.in-en.com/html/wind-2415018.shtml State Energy Agency. Wind power grid operation in the second quarter of 2021 [EB/OL].https://wind.in-en.com/html/wind-2415018.shtml.
[2] 杨茂, 于欣楠. 考虑风电场功率爬坡的超短期组合预测[J]. 东北电力大学学报, 2022, 42(1): 63-70.
YANG M, YU X N.An ultra-short term combined prediction considering wind farm power climbing[J]. Journal of northeast electric power university, 2022, 42(1): 63-70.
[3] 杨茂, 张书天, 王天硕, 等. 基于IKLIEP-四分位模型的风电场异常数据识别算法[J]. 高电压技术, 2023, 49(7): 2952-2960.
YANG M, ZHANG S T, WANG T S, et al.Identification algorithm of wind farm abnormal data based on IKLIEP-quartile model[J]. High voltage engineering, 2023, 49(7): 2952-2960.
[4] 杨茂, 代博祉, 刘蕾. 风电功率概率预测研究综述[J]. 东北电力大学学报, 2020, 40(2): 1-6.
YANG M, DAI B Z, LIU L.A review of wind power probabilistic prediction[J]. Journal of Northeast Electric Power University, 2020, 40(2): 1-6.
[5] 杨茂, 杨琼琼, 苏欣. 基于风电场等效平均风速的风电功率日前预测研究[J]. 太阳能学报, 2020, 41(2): 85-92.
YANG M, YANG Q Q, SU X.Research of day-ahead wind power forecast based on wind farm equivalent mean wind speed[J]. Acta energiae solaris sinica, 2020, 41(2): 85-92.
[6] YANG M, SHI C Y, LIU H Y.Day-ahead wind power forecasting based on the clustering of equivalent power curves[J]. Energy, 2021, 218: 119515.
[7] 刘念璋, 杨健, 柳玉, 等. 分布函数差异化导向的风电功率预测误差气象条件概率建模方法[J]. 电力自动化设备, 2022, 42(12): 58-65.
LIU N Z, YANG J, LIU Y, et al.Probabilistic modeling method of weather condition for wind power forecasting error based on differentiation orientation of distribution function[J]. Electric power automation equipment, 2022, 42(12): 58-65.
[8] YUAN X H, CHEN C, JIANG M, et al.Prediction interval of wind power using parameter optimized Beta distribution based LSTM model[J]. Applied soft computing, 2019, 82: 105550.
[9] WAN C, LIN J, WANG J H, et al.Direct quantile regression for nonparametric probabilistic forecasting of wind power generation[J]. IEEE transactions on power systems, 2017, 32(4): 2767-2778.
[10] 李丹, 任洲洋, 颜伟, 等. 基于因子分析和神经网络分位数回归的月度风电功率曲线概率预测[J]. 中国电机工程学报, 2017, 37(18): 5238-5247, 5522.
LI D, REN Z Y, YAN W, et al.Month-ahead wind power curve probabilistic prediction based on factor analysis and quantile regression neural network[J]. Proceedings of the CSEE, 2017, 37(18): 5238-5247, 5522.
[11] 王森,孙永辉,侯栋宸,等. 基于多窗宽核密度估计的风电功率超短期自适应概率预测[J]. 高电压技术, 2024, 50(7): 3070-3079.
WANG S,SUN Y H,HOU D C,et al.Very short-term adaptive probabilistic forecasting of wind power based on multi-band width kernel density estimation[J]. High voltage engineering, 2024, 50(7): 3070-3079.
[12] 潘霄, 张明理, 刘德宝, 等. 基于鲁棒多标签生成对抗的风电场日前出力区间预测[J]. 电力系统自动化, 2022, 46(10): 216-223.
PAN X, ZHANG M L, LIU D B, et al.Interval prediction of wind farm day-ahead output based on robust multi-label generative adversarial[J]. Automation of electric power systems, 2022, 46(10): 216-223.
[13] 胡文慧, 苏欣, 姜林, 等. 考虑预测误差时空相依性的短期风电功率概率预测[J]. 南方电网技术, 2023, 17(2): 137-144.
HU W H, SU X, JIANG L, et al.Probabilistic prediction of short-term wind power considering the temporal and spatial dependence of prediction errors[J]. Southern power system technology, 2023, 17(2): 137-144.
[14] 郝晓弘, 薛泽华, 裴婷婷, 等. 基于KELM与PSO-LSSVM组合核方法的风电功率区间预测研究[J]. 自动化与仪表, 2022, 37(4): 10-14, 20.
HAO X H, XUE Z H, PEI T T, et al.Research on wind power interval prediction based on combination kernel method of KELM and PSO-LSSVM[J]. Automation & instrumentation, 2022, 37(4): 10-14, 20.
[15] 杨锡运, 邢国通, 马雪, 等. 一种核极限学习机分位数回归模型及风电功率区间预测[J]. 太阳能学报, 2020, 41(11): 300-306.
YANG X Y, XING G T, MA X, et al.A model of quantile regression with kernel extreme learning machine and wind power interval prediction[J]. Acta energiae solaris sinica, 2020, 41(11): 300-306.
[16] 殷豪, 黄圣权, 孟安波, 等. 基于长短期记忆网络分位数回归的短期风电功率概率密度预测[J]. 太阳能学报, 2021, 42(2): 150-156.
YIN H, HUANG S Q, MENG A B, et al.Short-term wind power probability density prediction based on long short term memory network quantile regression[J]. Acta energiae solaris sinica, 2021, 42(2): 150-156.
[17] 李彬, 彭曙蓉, 彭君哲, 等. 基于深度学习分位数回归模型的风电功率概率密度预测[J]. 电力自动化设备, 2018, 38(9): 15-20.
LI B, PENG S R, PENG J Z, et al.Wind power probability density forecasting based on deep learning quantile regression model[J]. Electric power automation equipment, 2018, 38(9): 15-20.
[18] CANNON A J.Non-crossing nonlinear regression quantiles by monotone composite quantile regression neural network, with application to rainfall extremes[J]. Stochastic environmental research and risk assessment, 2018, 32(11): 3207-3225.
[19] BEN TAIEB S, HUSER R, HYNDMAN R J, et al.Forecasting uncertainty in electricity smart meter data by boosting additive quantile regression[J]. IEEE transactions on smart grid, 2016, 7(5): 2448-2455.
[20] HAO L X, NAIMAN D.Quantile regression[M]. California: America SAGE Publications, Inc., 2007.
[21] GAN D H, WANG Y, ZHANG N, et al.Enhancing short-term probabilistic residential load forecasting with quantile long-short-term memory[J]. The journal of engineering, 2017, 2017(14): 2622-2627.
[22] ZHENG S F.Gradient descent algorithms for quantile regression with smooth approximation[J]. International journal of machine learning and cybernetics, 2011, 2(3): 191-207.
[23] TANG W L, SHEN G H, LIN Y Y, et al.Nonparametric quantile regression: non-crossing constraints and conformal prediction[EB/OL]. https://arxiv.org/abs/2210.10161v1.

基金

国家重点研发计划(2022YFB2403000)

PDF(2521 KB)

Accesses

Citation

Detail

段落导航
相关文章

/