JI Shan,姜巍,景鑫
JI Shan,JIANG Wei,JING Xin

• 1:杭州迈拓大数据服务有限公司
• 2:国网浙江省电力有限公司科技创新中心

摘要(Abstract)：

针对现有电力负荷预测模型难以深入提取时空关联特征，模型泛化能力弱，无法同时胜任短期和长期的电力负荷预测的问题，提出一种面向多用户的基于余弦相似度和全局-局部协同图卷积网络的电力负荷预测方法。首先，利用余弦相似度来学习不同节点负荷数据之间的相似模式，以提取深层次的时空关联特征。其次，对影响电力负荷变化趋势的静态全局因素和动态局部因素进行协同建模，以提升模型的泛化能力。最后，通过在一个实测数据集上进行的大量实验，验证了该方法在短期和长期负荷序列预测任务中同时具备有效性和稳定性。
To address the challenges of existing power load forecasting models, which struggle to deeply extract spatiotemporal correlation features and exhibit weak generalization capabilities—failing to simultaneously manage both short-term and long-term forecasting—this study proposes a multi-user power load forecasting method using cosine similarity and a global-local collaborative graph convolutional network. First, cosine similarity is utilized to learn similar patterns between load data from different nodes, allowing for the deep extraction of spatiotemporal correlation features. Second, a collaborative modeling approach is applied to static global factors and dynamic local factors that influence power load trends, enhancing the model's generalization ability. Finally, extensive experiments on a real-world dataset demonstrate the method's effectiveness and robustness in forecasting both short-term and longterm load series.

关键词(KeyWords)： 多节点电力负荷预测;时空关联特征;余弦相似度;图卷积
multi-node power load forecasting;spatiotemporal correlation features;cosine similarity;graph convolutional

Abstract:

Keywords:

基金项目(Foundation): 国网浙江省电力有限公司科技项目（B311JZ230003）

作者(Author): JI Shan,姜巍,景鑫
JI Shan,JIANG Wei,JING Xin

DOI: 10.19585/j.zjdl.202501007

参考文献(References)：

• [1]沈建良，来骏，张翼，等.基于ACMD和BiGRU-Attention的负荷预测模型研究[J].浙江电力，2023,42(6):70-77.SHEN Jianliang,LAI Jun,ZHANG Yi,et al.Research on a load forecasting model based on ACMD and BiGRUAttention[J].ZheJiang Electric Power,2023(6):70-77.
• [2]封钰，宋佑斌，金晟，等.基于随机森林算法和粗糙集理论的改进型深度学习短期负荷预测模型[J].发电技术，2023,44(6):889-895.FENG Yu,SONG Youbin,JIN Sheng,et al. Improved deep learning model for forecasting short-term load based on random forest algorithm and rough set theory[J].Power Generation Technology,2023,44(6):889-895.
• [3]王涛，朱海南，李丰硕，等.基于高速公路神经网络和注意力机制的短期负荷预测方法[J].电力电容器与无功补偿，2023,44(4):72-79.WANG Tao,ZHU Hainan,LI Fengshuo,et al.Short-term load forecast method based on highway neural network and attention mechanism[J].Power Capacitor&Reactive Power Compensation,2023,44(4):72-79.
• [4]陈逸枞，张大海，李宇欣，等.基于DWT-MOSMA-SVM的多目标优化短期母线负荷预测[J].电力建设，2023,44(3):49-55.CHEN Yicong,ZHANG Dahai,LI Yuxin,et al. Multiobjective optimization based on DWT-MOSMA-SVM for short-term bus load forecasting[J]. Electric Power Construction,2023,44(3):49-55.
• [5]李振坤，黄滢，李谅，等.计及需求侧响应的主动配电网多时间尺度优化调度[J].电力建设，2023,44(3):36-48.LI Zhenkun,HUANG Ying,LI Liang,et al. Multi-time scale optimal dispatching of active distribution network considering demand-side response[J].Electric Power Construction,2023,44(3):36-48.
• [6] ZAREMBA W,SUTSKEVER I,VINYALS O. Recurrent neural network regularization[J]. ArXiv e-Prints,2014:arXiv:1409.2329.
• [7] MA X L,TAO Z M,WANG Y H,et al.Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J]. Transportation Research Part C-emerging Technologies,2015,54:187-197.
• [8] HUANG Z H,XU W,YU K.Bidirectional LSTM-CRF models for sequence tagging[EB/OL].[2024-06-02].https：//arxiv.org/abs/1508.01991v1.
• [9]王鑫，李安桂，李扬，等.基ARIMA-LSTM模型的综合能源系统负荷与风光资源预测[J].西安建筑科技大学学报（自然科学版），2022,54(5):762-769.WANG Xin,LI Angui,LI Yang,et al. Multivariate load prediction and wind-solar resource characteristic quantity prediction of integrated energy system based on ARIMALSTM model[J]. Xi'an Univ. of Arch.&Tech.(Natural Science Edition),2022,54(5):762-769.
• [10]黄冬梅，张宁宁，胡安铎，等.基于双层XGBoost和数据增强的空间负荷预测方法[J].电力工程技术，2023,42(1):201-208.HUANG Dongmei,ZHANG Ningning,HU Anduo,et al.Spatial load forecasting method based on double-layer XGBoost and data enhancement[J].Electric Power Engineering Technology,2023,42(1):201-208.
• [11]唐晓，陈芳，许强，等.改进鲸鱼算法优化的多维度深度极限学习机短期负荷预测[J].山东电力技术，2023,50(1):1-7.TANG Xiao,CHEN Fang,XU Qiang,et al. Short-term load forecasting based on multi-dimensional deep extreme learning machine optimized by improved whale algorithm[J].Shandong Electric Power,2023,50(1):1-7.
• [12] LECUN Y,BOTTOU L,BENGIO Y,et al. Gradientbased learning applied to document recognition[J]. Proceedings of the IEEE,86(1998):2278-2324.
• [13]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 Jixiang,ZHANG Qipei,YANG Zhihong,et al.Shortterm load forecasting method based on CNN-LSTM hybrid neural network model[J]. Automation of Electric Power Systems,2019,43(8):131-137.
• [14]王裕鑫.基于CNN-BiLSTM算法的电力负荷预测方法研究[D].西安：西安理工大学，2022.WANG Yuxin. Research on the power load forecasting method based on CNN-BiLSTM[D].Xi’an:Xi’an University of Technology,2022.
• [15] KIPF T N,WELLING M.Semi-supervised classification with graph convolutional networks[J]. ArXiv e-Prints,2016:arXiv:1609.02907.
• [16] ZHAO L,SONG Y J,ZHANG C,et al.T-GCN:A temporal graph convolutional network for traffic prediction[J].IEEE Transactions on Intelligent Transportation Systems,2020,21(9):3848-3858.
• [17]史含笑，王雷春.结合LSTM和自注意力机制的图卷积网络短期电力负荷预测[J].计算机应用，2024,44(1):311-317.SHI Hanxiao,WANG Leichun. Short-term power load forecasting by graph convolutional network combining LSTM and self-attention mechanism[J].Journal of Computer Applications,2024,44(1):311-317.
• [18] LUO C J,ZHAN J F,XUE X H,et al.Cosine normalization:using cosine similarity instead of dot product in neural networks[M]//Lecture Notes in Computer Science.Cham:Springer International Publishing,2018:382-391.
• [19] YU F,KOLTUN V. Multi-scale context aggregation by dilated convolutions[EB/OL].[2024-06-25]. https：//arxiv.org/abs/1511.07122v3.
• [20] WU Z H,PAN S R,LONG G D,et al.Graph WaveNet for deep spatial-temporal graph modeling[J]. ArXiv ePrints,2019:arXiv:1906.00121.
• [21] DAUPHIN N Y,FAN A,AULI M,et al.Language modeling with gated convolutional networks[J]. ArXiv ePrints,2016:arXiv:1612.08083.
• [22] SHAO Z Z,ZHANG Z,WANG F,et al.Spatial-temporal identity:a simple yet effective baseline for multivariate time series forecasting[C]//Proceedings of the 31st ACM International Conference on Information&Knowledge Management.Atlanta GA USA.ACM,2022:4454-4458.
• [23] ZENG A L,CHEN M X,ZHANG L,et al.Are transformers effective for time series forecasting[J].Proceedings of the AAAI Conference on Artificial Intelligence,2023,37(9):11121-11128.
• [24] NIE Y Q,NGUYEN N H,SINTHONG P,et al.A time series is worth 64 words:long-term forecasting with transformers[EB/OL].[2024-06-21]. https：//arxiv. org/pdf/2211.14730.