针对传统风电功率预测方法在局部峰值附近的波动数据区间和突变数据点预测精度有限的问题,提出基于BiTCN-Informer融合时空特征的超短期风电功率预测方法。首先,利用双向时间卷积神经网络（BiTCN）分析过去观测值与未来协变量的信息特征,提取数据的空间特征;然后,采用Informer概率稀疏自注意力机制和自注意力蒸馏机制简化网络参数,减少计算复杂度,识别长序列数据的依赖关系,捕捉序列的时序特征;最后,引入全连接层整合时序和空间特征。算例分析结果表明,所提组合模型相比其他模型能有效提高预测精度。
Abstract
Aiming at the problem that the prediction accuracy of fluctuation data intervals and abrupt data points near the local peak value of traditional wind power prediction methods is limited, an ultra-short term wind power prediction method based on BiTCN-Informer fusion of temporal and spatial characteristics is proposed. Firstly, bidirectional temporal convolutional network （BiTCN） is used to analyze the information features of past observations and future covariates, and to extract the spatial features of the data. Then, the Informer probabilistic sparse self-attention mechanism and self-attention distillation mechanism are used to simplify the network parameters, reduce the computational complexity, identify the dependency relationship of long sequence data, and capture the temporal features of the sequence. Finally, a fully connected layer is introduced to integrate temporal and spatial features. The results of the case study show that the proposed combined model can effectively improve the prediction accuracy compared with other models.
关键词
风电功率预测 /
双向时间卷积神经网络 /
Informer /
概率稀疏自注意力机制 /
自注意力蒸馏机制
Key words
wind power forecasting /
bidirectional temporal convolutional network /
Informer /
probsparse self-attention /
self-attention distilling
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 韩丽, 程颖洁, 王施琪, 等. 基于扩散模型和爬坡趋势分类的风电功率自适应区间预测[J]. 电网技术, 2024, 48(6): 2448-2457.
HAN L, CHENG Y J, WANG S Q, et al.Adaptive interval forecasting of wind power based on diffusion model and ramping trend classification[J]. Power system technology, 2024, 48(6): 2448-2457.
[2] 臧海祥, 赵勇凯, 张越, 等. 基于低风速功率修正和损失函数改进的超短期风电功率预测[J]. 电力系统自动化, 2024, 48(7): 248-257.
ZANG H X, ZHAO Y K, ZHANG Y, et al.Ultra-short-term wind power prediction based on power correction under low wind speed and improved loss function[J]. Automation of electric power systems, 2024, 48(7): 248-257.
[3] 袁畅, 王森, 孙永辉, 等. 基于混合特征双重衍生和误差修正的风电功率超短期预测[J]. 电力系统自动化, 2024, 48(5): 68-76.
YUAN C, WANG S, SUN Y H, et al.Ultra-short-term forecasting of wind power based on dual derivation of hybrid features and error correction[J]. Automation of electric power systems, 2024, 48(5): 68-76.
[4] KISVARI A, LIN Z, LIU X L.Wind power forecasting-a data-driven method along with gated recurrent neural network[J]. Renewable energy, 2021, 163: 1895-1909.
[5] WANG L, TAO R, HU H L, et al.Effective wind power prediction using novel deep learning network: stacked independently recurrent autoencoder[J]. Renewable energy, 2021, 164: 642-655.
[6] 朱菊萍, 魏霞, 谢丽蓉, 等. 基于VMD和改进BiLSTM的短期风电功率预测[J]. 太阳能学报, 2024, 45(6): 422-428.
ZHU J P, WEI X, XIE L R, et al.Short-term wind power peediction based on VMD and improved BiLSTM[J]. Acta energiae solaris sinica, 2024, 45(6): 422-428.
[7] 乔宽龙, 董存, 车建峰, 等. 时空特性下基于图卷积神经网络的风电集群功率短期预测方法[J]. 太阳能学报, 2024, 45(5): 95-103.
QIAO K L, DONG C, CHE J F, et al.Short-term prediction method of wind power clusters based on graph convolution neural network under spitio-temporal characteristics[J]. Acta energiae solaris sinica, 2024, 45(5): 95-103.
[8] 刘沛汉, 尹翠, 贾娜, 等. 基于小生境遗传算法与径向基代理模型的短期风电功率预测[J]. 太阳能学报, 2024, 45(8): 554-564.
LIU P H, YIN C, JIA N, et al.Short-term wind power pridiction based on niche genetic algorithm and radial basis surrogate model[J]. Acta energiae solaris sinica, 2024, 45(8): 554-564.
[9] 吕云龙, 胡琴, 熊俊杰, 等. 基于时空注意力卷积模型的超短期风电功率预测[J]. 电网技术, 2024, 48(5): 2064-2073.
LYU Y L, HU Q, XIONG J J, et al.Ultra-short-term wind power prediction based on spatiotemporal attention convolution model[J]. Power system technology, 2024, 48(5): 2064-2073.
[10] 李练兵, 高国强, 吴伟强, 等. 考虑特征重组与改进Transformer的风电功率短期日前预测方法[J]. 电网技术, 2024, 48(4): 1466-1480.
LI L B, GAO G Q, WU W Q, et al.Short-term day-ahead wind power prediction considering feature recombination and improved transformer[J]. Power system technology, 2024, 48(4): 1466-1480.
[11] SPRANGERS O, SCHELTER S, DE RIJKE M.Parameter-efficient deep probabilistic forecasting[J]. International journal of forecasting, 2023, 39(1): 332-345.
[12] ZHOU H Y, ZHANG S H, PENG J Q, et al.Informer: beyond efficient transformer for long sequence time-series forecasting[J]. Proceedings of the AAAI conference on artificial intelligence, 2021, 35(12): 11106-11115.
[13] CHEN Y B, XU J J.Solar and wind power data from the Chinese State Grid renewable energy generation forecasting competition[J]. Scientific data, 2022, 9: 577.
基金
湖南省教育厅科学研究重点项目（23A0441）; 湖南省自然科学基金（2023JJ50176）
PDF(4581 KB)
