针对电力负荷波动性与非线性增强的问题,提出一种基于软动态时间规整-改进中心点聚类（SDTW-IPAM）和Informer的短期电力负荷预测方法。首先,引入Soft-DTW距离度量方法以准确刻画负荷曲线的动态时序特征,有效识别负荷曲线的局部形变;并结合Gap统计量和K-均值++初始化策略对PAM聚类算法进行改进,将原始负荷划分为双峰负荷、高峰负荷和平稳负荷,使得模型训练更有针对性。其次,对各类典型负荷应用最大互信息数（MIC）进行特征选择,识别出负荷的关键影响因素,实现差异化特征提取。同时,为提高模型的预测性能,引入Informer模型分别构建专属预测模型。最后,以新疆乌鲁木齐实际负荷数据进行实例验证,结果表明所提组合预测模型可有效提升短期负荷预测的准确率,具有较强的实用价值。
Abstract
A short-term power load forecasting method based on soft dynamic time warping-improved partitioning around medoids （SDTW-IPAM） and Informer is proposed to address the problem of power load volatility and nonlinearity enhancement. Firstly, the soft-DTW distance metric is introduced to accurately portray the dynamic time-series characteristics of the load curve and effectively identify the local deformation of the load curve; and the PAM clustering algorithm is improved by combining the Gap statistic and the K-means++ initialization strategy, which classifies the original loads into bimodal loads, peak loads and smooth loads, so that the model training is more targeted. Secondly, the maximum information coefficient （MIC） is applied to the typical loads of each category for feature selection, which identifies the key influencing factors of the loads and achieves differentiated feature extraction. Meanwhile, in order to improve the prediction performance of the model, the Informer model is introduced to construct exclusive prediction models respectively. Finally, the actual load data of Urumqi, Xinjiang is used as an example for validation, and the results show that the proposed combined forecasting model can effectively improve the accuracy of short-term load forecasting and has strong practical value.
关键词
负荷预测 /
聚类算法 /
注意力机制 /
Informer /
软动态时间规整
Key words
electric load forecasting /
clustering algorithms /
attention mechanism /
Informer /
soft dynamic time warping
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 王印松, 吕率豪. 基于改进时间卷积网络的微电网超短期负荷预测[J]. 太阳能学报, 2024, 45(6): 255-263.
Wang Y S, Lyu S H.Ultra-short-term power load prediction ofmicro-grid based on improved temporalconvolutional network[J]. Acta Energiae Solaris Sinica, 2024, 45(6): 255-263.
[2] 李练兵, 郭兴辰, 曾四鸣, 等. 基于GLDSC-ConvAutoformer模型的区域电动汽车短期充电负荷预测[J]. 太阳能学报, 2025, 46(2): 90-98.
Li L B, Guo X C, Zeng S M, et al.Short-term charging load prediction of regional electric vehicles based on GLDSC-ConvAutoformer model[J]. Acta Energiae Solaris Sinica, 2025, 46(2): 90-98.
[3] 冯佳威, 王海鑫, 杨子豪, 等. 基于数字孪生与多模型融合的多元负荷短期预测[J]. 太阳能学报, 2024, 45(10): 97-106.
Feng J W, Wang H X, Yang Z H, et al.Short-term forecasting of multivariate load based on digital twin and multi-model fusion[J]. Acta Energiae Solaris Sinica, 2024, 45(10): 97-106.
[4] Habbak H, Mahmoud M, Metwally K, et al.Load forecasting techniques and their applications in smart grids[J]. Energies, 2023, 16(3): 1480.
[5] Dudek G, Pełka P, Smyl S.A hybrid residual dilated LSTM and exponential smoothing model for midterm electric load forecasting[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022, 33(7): 2879-2891.
[6] 肖白, 赵晓宁, 姜卓, 等. 利用模糊信息粒化与支持向量机的空间负荷预测方法[J]. 电网技术, 2021, 45(1): 251-258.
Xiao B, Zhao X N, Jiang Z, et al.Spatial load forecasting method using fuzzy information granulation and support vector machine[J]. Power System Technology, 2021, 45(1): 251-258.
[7] Abumohsen M, Owda A Y, Owda M.Electrical load forecasting using LSTM, GRU, and RNN algorithms[J]. Energies, 2023, 16(5): 2283.
[8] 胡威, 张新燕, 李振恩, 等. 基于优化的VMD-mRMR-LSTM模型的短期负荷预测[J]. 电力系统保护与控制, 2022, 50(1): 88-97.
Hu W, Zhang X Y, Li Z E, et al.Short-term load forecasting based on an optimized VMD-mRMR-LSTM model[J]. Power System Protection and Control, 2022, 50(1): 88-97.
[9] 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.
[10] 陆继翔, 张琪培, 杨志宏, 等. 基于CNN-LSTM混合神经网络模型的短期负荷预测方法[J]. 电力系统自动化, 2019, 43(8): 131-137.
Lu J X, Zhang Q P, Yang Z H, et al.Short-term load forecasting method based on CNN-LSTM hybrid neural network model[J]. Automation of Electric Power Systems, 2019, 43(8): 131-137.
[11] 魏健, 赵红涛, 刘敦楠, 等. 基于注意力机制的CNN-LSTM短期电力负荷预测方法[J]. 华北电力大学学报(自然科学版), 2021, 48(1): 42-47.
Wei J, Zhao H T, Liu D N, et al.Short-term power load forecasting method by Attention-based CNN-LSTM[J]. Journal of North China Electric Power University (natural science edition), 2021, 48(1): 42-47.
[12] Wu H X, Xu J H, Wang J M, et al.Autoformer: decomposition transformers with auto-correlation for long-term series forecasting[C]//Neural Information Processing Systems. Vancouver, Canada, 2021.
[13] Zhou T, Ma Z Q, Wang X, et al.Fedformer: frequency enhanced decomposed transformer for long-term series forecasting[M]//AI for Time Series. Boca Raton: CRC Press, 2026: 10-34.
[14] 石卓见, 冉启武, 徐福聪. 基于聚合二次模态分解及Informer的短期负荷预测[J]. 电网技术, 2024, 48(6): 2574-2583.
Shi Z J, Ran Q W, Xu F C.Short-term load forecasting based on aggregated secondary decomposition and informer[J]. Power System Technology, 2024, 48(6): 2574-2583.
[15] 叶林, 宫婷, 宋旭日, 等. 基于波动类型精细划分与聚类的短期负荷预测[J]. 电网技术, 2023, 47(3): 998-1009.
Ye L, Gong T, Song X R, et al.Short-term load forecasting based on fine division and clustering of fluctuation types[J]. Power System Technology, 2023, 47(3): 998-1009.
[16] 王宇飞, 杜桐, 边伟国, 等. 基于DTW K-medoids与VMD-多分支神经网络的多用户短期负荷预测[J]. 中国电力, 2024, 57(6): 121-130.
Wang Y F, Du T, Bian W G, et al.Short-term load forecasting based on DTW K-medoids and VMD multi-branch neural network for multiple users[J]. Electric Power, 2024, 57(6): 121-130.
[17] 董雷, 陈振平, 韩富佳, 等. 基于图卷积神经网络与K-means聚类的居民用户集群短期负荷预测[J]. 电网技术, 2023, 47(10): 4291-4301.
Dong L, Chen Z P, Han F J, et al.Short-term load forecasting of residential user groups based on graph convolutional neural network and K-means clustering[J]. Power Ssystem Technology, 2023, 47(10): 4291-4301.
[18] 张炀, 汪洋, 祝宇翔, 等. 基于PAM和ELM的电力短期负荷预测相似日选取算法[J]. 电力大数据, 2017, 20(12): 84-87.
Zhang Y, Wang Y, Zhu Y X, et al.Similarity day selection algorithm for power short-term load forecasting[J]. Power Systems and Big Data, 2017, 20(12): 84-87.
[19] 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.
[20] 崔树银, 汪昕杰. 基于最大信息系数和多目标Stacking集成学习的综合能源系统多元负荷预测[J]. 电力自动化设备, 2022, 42(5): 32-39, 81.
Cui S Y, Wang X J.Multivariate load forecasting in integrated energy system based on maximal information coefficient and multi-objective Stacking ensemble learning[J]. Electric Power Automation Equipment, 2022, 42(5): 32-39, 81.
基金
国网新疆电力有限公司科技项目（SGXJ0000TKJS2310341）
PDF(2171 KB)
