针对传统风功率预测方法通常基于固定时间粒进行研究,但该类方法往往忽略了其他时间粒度对风功率的影响的问题,提出一种基于多粒度时间卷积网络（MGTCN）的超短期风功率预测方法,使用时间卷积网络来挖掘多粒度视角下的风力机数据特征,并设计多粒度特征融合模块来增强模型的鲁棒性,提高风功率预测精度。首先,利用随机森林算法（RF）得到与输出功率相关性较强的部分特征数据;然后,对筛选后的特征数据进行多粒度划分,通过时间卷积网络（TCN）提取各个粒度的独立特征。最后,使用挤压激励网络（SENet）对不同粒度特征进行自适应加权融合,得到最终预测值。采用中国某风场数据进行算例分析,结果表明相较于其他方法,所提方法在24步预测任务和6步预测任务上取得了最佳的预测性能,具有较高的准确性和稳定性。在24步预测任务上归一化均方根误差、归一化平均绝对值误差和决定系数指标分别为0.152、0.108和0.7214,在6步预测任务上各指标分别为0.1027, 0.0683和0.8717。
Abstract
To address the problem that traditional wind power prediction methods are usually based on the fixed time granularity and often ignore the influence of other time granularity. an ultra-short-term wind power prediction method based on multi-granularity time convolution network （MGTCN） is proposed. In our proposed method, the temporal convolution network （TCN） is used to mine the characteristics of wind turbine data from a multi-granularity perspective, and a multi-granularity feature fusion module is designed to enhance the robustness of the model and improve the accuracy of wind power prediction. Firstly, the random forest algorithm （RF） is used to select related feature data with strong correlation with the output power. Then, the filtered feature data is divided into multiple granularities, and the independent features of each granularity are extracted through TCN. Finally, the squeeze and excitation network （SENet） is used to adaptively weight the fusion of different granularity features to obtain the final prediction value. A wind field data in China is used for example analysis. The results show that compared with other methods, our proposed method achieved the best performance with high accuracy and stability on both the twenty-four-step prediction task and the six-step prediction task. Specifically, in terms of three common prediction performance metrics, including normalized root mean square error, normalize mean absolute error and R2, our proposed method obtained the best performance with 0.152, 0.108 and 0.7214, respectively on the twenty-four-step prediction task. For the six-step prediction task, it achieved 0.1027, 0.0683 and 0.8717, respectively.
关键词
风功率 /
预测 /
随机森林 /
多粒度计算 /
时间卷积网络 /
挤压激励网络
Key words
wind power /
prediction /
random forests /
multi-granularity computing /
temporal convolution network /
squeeze and excitation network
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] LIU H, CHEN C, LYU X W, et al.Deterministic wind energy forecasting: a review of intelligent predictors and auxiliary methods[J]. Energy conversion and management, 2019, 195: 328-345.
[2] 钱政, 裴岩, 曹利宵, 等. 风电功率预测方法综述[J]. 高电压技术, 2016, 42(4): 1047-1060.
QIAN Z, PEI Y, CAO L X, et al.Review of wind power forecasting method[J]. High voltage engineering, 2016, 42(4): 1047-1060.
[3] 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.
[4] 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.
[5] YE L, ZHAO Y N, ZENG C, et al.Short-term wind power prediction based on spatial model[J]. Renewable energy, 2017, 101: 1067-1074.
[6] 卢继平, 曾燕婷, 喻华, 等. 基于改进AWNN的风电功率超短期多步预测[J]. 太阳能学报, 2021, 42(1): 166-173.
LU J P, ZENG Y T, YU H, et al.Ultra-short-term wind power multi-step forecasting based on improved awnn[J]. Acta energiae solaris sinica, 2021, 42(1): 166-173.
[7] 胡梦月, 胡志坚, 仉梦林, 等. 基于改进AdaBoost.RT和KELM的风功率预测方法研究[J]. 电网技术, 2017, 41(2): 536-542.
HU M Y, HU Z J, ZHANG M L, et al.Research on wind power forecasting method based on improved AdaBoost.RT and KELM algorithm[J]. Power system technology, 2017, 41(2): 536-542.
[8] LI L L, CEN Z Y, TSENG M L, et al.Improving short-term wind power prediction using hybrid improved cuckoo search arithmetic-support vector regression machine[J]. Journal of cleaner production, 2021, 279: 123739.
[9] YANG M, CHEN X X, DU J, et al.Ultra-short-term multistep wind power prediction based on improved EMD and reconstruction method using run-length analysis[J]. IEEE access, 2018, 6: 31908-31917.
[10] YUAN X H, TAN Q X, LEI X H, et al.Wind power prediction using hybrid autoregressive fractionally integrated moving average and least square support vector machine[J]. Energy, 2017, 129: 122-137.
[11] ZHANG Y Z, XIONG R, HE H W, et al.Long short-term memory recurrent neural network for remaining useful life prediction of lithium-ion batteries[J]. IEEE transactions on vehicular technology, 2018, 67(7): 5695-5705.
[12] 王增平, 赵兵, 纪维佳, 等. 基于GRU-NN模型的短期负荷预测方法[J]. 电力系统自动化, 2019, 43(5): 53-58.
WANG Z P, ZHAO B, JI W J, et al.Short-term load forecasting method based on GRU-NN model[J]. Automation of electric power systems, 2019, 43(5): 53-58.
[13] 朱乔木, 李弘毅, 王子琪, 等. 基于长短期记忆网络的风电场发电功率超短期预测[J]. 电网技术, 2017, 41(12): 3797-3802.
ZHU Q M, LI H Y, WANG Z Q, et al.Short-term wind power forecasting based on LSTM[J]. Power system technology, 2017, 41(12): 3797-3802.
[14] DING M, ZHOU H, XIE H, et al.A gated recurrent unit neural networks based wind speed error correction model for short-term wind power forecasting[J]. Neurocomputing, 2019, 365: 54-61.
[15] BAI S, KOLTER J Z, KOLTUN V.An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[EB/OL]. https://arxiv.org/abs/1803.01271.
[16] 黄睿, 杜文娟, 王海风. 计及湍流强度的风电功率短期预测[J]. 电网技术, 2019, 43(6): 1907-1914.
HUANG R, DU W J, WANG H F.Short-term prediction of wind power considering turbulence intensity[J]. Power system technology, 2019, 43(6): 1907-1914.
[17] 王国胤, 傅顺, 杨洁, 等. 基于多粒度认知的智能计算研究[J]. 计算机学报, 2022, 45(6): 1161-1175.
WANG G Y, FU S, YANG J, et al.A review of research on multi-granularity cognition based intelligent computing[J]. Chinese journal of computers, 2022, 45(6): 1161-1175.
[18] HU J, SHEN L, SUN G.Squeeze-and-excitation networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA, 2018: 7132-7141.
[19] 杨晓梅, 郭林明, 肖先勇, 等. 基于可调品质因子小波变换和随机森林特征选择算法的电能质量复合扰动分类[J]. 电网技术, 2020, 44(8): 3014-3020.
YANG X M, GUO L M, XIAO X Y, et al.Classification of multiple power quality disturbances based on TQWT and random forest feature selection algorithm[J]. Power system technology, 2020, 44(8): 3014-3020.
[20] 赵荣臻, 文云峰, 叶希, 等. 基于改进堆栈降噪自动编码器的预想事故频率指标评估方法研究[J]. 中国电机工程学报, 2019, 39(14): 4081-4093.
ZHAO R Z, WEN Y F, YE X, et al.Research on frequency indicators evaluation of disturbance events based on improved stacked denoising autoencoders[J]. Proceedings of the CSEE, 2019, 39(14): 4081-4093.
[21] 庞昊, 高金峰, 杜耀恒. 基于时间卷积网络分位数回归的短期负荷概率密度预测方法[J]. 电网技术, 2020, 44(4): 1343-1350.
PANG H, GAO J F, DU Y H.A short-term load probability density prediction based on quantile regression of time convolution network[J]. Power system technology, 2020, 44(4): 1343-1350.
[22] 石重托, 姚伟, 黄彦浩, 等. 基于SE-CNN和仿真数据的电力系统主导失稳模式智能识别[J]. 中国电机工程学报, 2022, 42(21): 7719-7731.
SHI Z T, YAO W, HUANG Y H, et al.Power system dominant instability mode identification based on convolutional neural networks with squeeze and excitation block and simulation data[J]. Proceedings of the CSEE, 2022, 42(21): 7719-7731.
[23] Q/GDW 588—2011, 风电功率预测功能规范[S].
Q/GDW 588—2011, Function specification of wind power forcasting[S].
[24] 唐新姿, 顾能伟, 黄轩晴, 等. 风电功率短期预测技术研究进展[J]. 机械工程学报, 2022, 58(12): 213-236.
TANG X Z, GU N W, HUANG X Q, et al.Progress on short term wind power forecasting technology[J]. Journal of mechanical engineering, 2022, 58(12): 213-236.
[25] NIU Z W, YU Z Y, TANG W H, et al.Wind power forecasting using attention-based gated recurrent unit network[J]. Energy, 2020, 196: 117081.
基金
国家自然科学基金（62273299）; 中央引导地方科技发展资金项目（216Z2101G）; 河北省自然科学基金（F2021203009）; 秦皇岛市科学技术研究与发展计划项目（201902A032）
PDF(1910 KB)
