基于可变形CNN-LSTM-Attention模型的跨场景短期电力负荷预测

颜湘武; 王泽华; 卢俊达; 曹贺杨; 贾焦心

doi:10.19912/j.0254-0096.tynxb.2025-0558

PDF(2062 KB)

太阳能学报 ›› 2026, Vol. 47 ›› Issue (9) : 403-416. DOI: 10.19912/j.0254-0096.tynxb.2025-0558

基于可变形CNN-LSTM-Attention模型的跨场景短期电力负荷预测

颜湘武, 王泽华, 卢俊达, 曹贺杨, 贾焦心

作者信息 +

CROSS-SCENARIO SHORT-TERM POWER LOAD FORECASTING BASED ON DEFORMABLE CNN-LSTM-ATTENTION MODEL

Yan Xiangwu, Wang Zehua, Lu Junda, Cao Heyang, Jia Jiaoxin

Author information +

文章历史 +

摘要

针对现有短期电力负荷预测模型结构固化导致的跨场景适应能力不足问题,提出一种基于改进麻雀搜索算法（ISSA）优化的可变形卷积神经网络-长短期记忆网络-注意力（CNN-LSTM-Attention）模型,具备面向不同数据场景动态构建最优配置的能力。首先,为实现模型结构与超参数的自适应调整,引入ISSA联合优化CNN与LSTM的网络层数、学习率和批次大小。其中,ISSA融合拉丁超立方抽样法、余弦退火、动态螺旋策略以及莱维飞行机制,以提升优化质量与收敛性能。然后,依据ISSA在不同场景特征下得到的最优配置,构建可变形CNN-LSTM-Attention模型。最后,在美国纽约州和中国南方地区的电力负荷数据上进行仿真验证,结果表明所提模型在多时间分辨率和不同地区季节条件下,决定系数较CNN-LSTM等主流预测模型至少提升1.18%,均方根误差至少降低45.03%,展现出良好的预测性能与跨场景适应性。

Abstract

Addressing the limited cross-scenario adaptability resulting from the inflexible structures of current short-term power load forecasting models, this paper proposes a deformable convolutional neural network （CNN）-long short-term memory （LSTM）-attention model optimized using an improved sparrow search algorithm （ISSA）. This approach enables the dynamic construction of optimal configurations tailored to diverse data scenarios. Firstly, to achieve adaptive adjustment of both model architecture and hyperparameters, the ISSA jointly optimizes the number of CNN and LSTM layers, learning rate, and batch size. The ISSA incorporates Latin hypercube sampling, cosine annealing, a dynamic spiral strategy, and a Lévy flight mechanism to enhance optimization quality and convergence. Subsequently, a deformable CNN-LSTM-Attention model is constructed based on the optimal configurations identified by the ISSA for different scenario characteristics. Finally, simulation and validation are conducted using power load data from New York State （USA） and South China. The results demonstrate that, across multiple time resolutions and varying regional and seasonal conditions, the proposed model improves the coefficient of determination by at least 1.18% and reduces the root mean square error by at least 45.03% compared to mainstream forecasting models such as CNN-LSTM, demonstrating excellent forecasting performance and cross-scenario adaptability.

导出引用

颜湘武, 王泽华, 卢俊达, 曹贺杨, 贾焦心. 基于可变形CNN-LSTM-Attention模型的跨场景短期电力负荷预测[J]. 太阳能学报. 2026, 47(9): 403-416 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0558

Yan Xiangwu, Wang Zehua, Lu Junda, Cao Heyang, Jia Jiaoxin. CROSS-SCENARIO SHORT-TERM POWER LOAD FORECASTING BASED ON DEFORMABLE CNN-LSTM-ATTENTION MODEL[J]. Acta Energiae Solaris Sinica. 2026, 47(9): 403-416 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0558

中图分类号： TM715

参考文献

[1] Bashir T, Chen H Y, Tahir M F, et al.Short term electricity load forecasting using hybrid prophet-LSTM model optimized by BPNN[J]. Energy Reports, 2022, 8: 1678-1686.
[2] 孔祥玉, 马玉莹, 艾芊, 等. 新型电力系统多元用户的用电特征建模与用电负荷预测综述[J]. 电力系统自动化, 2023, 47(13): 2-17.
Kong X Y, Ma Y Y, Ai Q, et al.Review on electricity consumption characteristic modeling and load forecasting for diverse users in new power system[J]. Automation of Electric Power Systems, 2023, 47(13): 2-17.
[3] Nti I K, Teimeh M, Nyarko-Boateng O, et al.Electricity load forecasting: a systematic review[J]. Journal of electrical Systems and Information Technology, 2020, 7(1): 13.
[4] 罗凤章, 张旭, 杨欣, 等. 基于深度学习的综合能源配电系统负荷分析预测[J]. 高电压技术, 2021, 47(1): 23-32.
Luo F Z, Zhang X, Yang X, et al.Load analysis and prediction of integrated energy distribution system based on deep learning[J]. High Voltage Engineering, 2021, 47(1): 23-32.
[5] 赵洪山, 吴雨晨, 潘思潮, 等. 基于TCN和高斯过程残差建模学习的净负荷概率预测方法[J]. 太阳能学报, 2024, 45(12): 588-595.
Zhao H S, Wu Y C, Pan S C, et al.Probabilistic net load forecasting based on TCN and Gaussian process-enabled residual modeling learning approach[J]. Acta Energiae Solaris Sinica, 2024, 45(12): 588-595.
[6] 范杏蕊, 李元诚. 基于改进Autoformer模型的短期电力负荷预测[J]. 电力自动化设备, 2024, 44(4): 171-177.
Fan X R, Li Y C.Short-term power load forecasting based on improved Autoformer model[J]. Electric Power Automation Equipment, 2024, 44(4): 171-177.
[7] 韩富佳, 王晓辉, 乔骥, 等. 基于人工智能技术的新型电力系统负荷预测研究综述[J]. 中国电机工程学报, 2023, 43(22): 8569-8591.
Han F J, Wang X H, Qiao J, et al.Review on artificial intelligence based load forecasting research for the new-type power system[J]. Proceedings of the CSEE, 2023, 43(22): 8569-8591.
[8] Wang S M, Wang S X, Chen H W, et al.Multi-energy load forecasting for regional integrated energy systems considering temporal dynamic and coupling characteristics[J]. Energy, 2020, 195: 116964.
[9] Bashiri Behmiri N, Fezzi C, Ravazzolo F.Incorporating air temperature into mid-term electricity load forecasting models using time-series regressions and neural networks[J]. Energy, 2023, 278: 127831.
[10] Wang Y Z, Zhang N Q, Chen X.A short-term residential load forecasting model based on LSTM recurrent neural network considering weather features[J]. Energies, 2021, 14(10): 2737.
[11] Haque A, Rahman S.Short-term electrical load forecasting through heuristic configuration of regularized deep neural network[J]. Applied Soft Computing, 2022, 122: 108877.
[12] 薛东, 段立强, 高统彤, 等. 考虑多重特征与不确定性度量的综合能源系统负荷预测研究[J]. 太阳能学报, 2024, 45(7): 379-388.
Xue D, Duan L Q, Gao T T, et al.Study of integrated energy system load forecasting considering multiple characteristics and uncertainty measures[J]. Acta Energiae Solaris Sinica, 2024, 45(7): 379-388.
[13] 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.
[14] Chen J C, Zheng L, Che W Y, et al.A method for measuring carbon emissions from power plants using a CNN-LSTM-Attention model with Bayesian optimization[J]. Case Studies in Thermal Engineering, 2024, 63: 105334.
[15] Wan A P, Chang Q, Al-Bukhaiti K, et al.Short-term power load forecasting for combined heat and power using CNN-LSTM enhanced by attention mechanism[J]. Energy, 2023, 282: 128274.
[16] Liu H M, Tang Y, Pu Y, et al.Short-term load forecasting of multi-energy in integrated energy system based on multivariate phase space reconstruction and support vector regression mode[J]. Electric Power Systems Research, 2022, 210: 108066.
[17] Zhang W, Cai Y, Zhan H Y, et al.Multi-energy load forecasting for small-sample integrated energy systems based on neural network Gaussian process and multi-task learning[J]. Energy Conversion And Management, 2024, 321: 119027.
[18] 王延峰, 曹育晗, 孙军伟. 基于多策略改进金豺算法优化LSTM的短期电力负荷预测[J]. 电力系统保护与控制, 2024, 52(14): 95-102.
Wang Y F, Cao Y H, Sun J W.Short-term power load forecasting based on multi-strategy improved golden jackal algorithm-optimized LSTM[J]. Power System Protection and Control, 2024, 52(14): 95-102.
[19] MagalhãEs B, Bento P, Pombo J, et al. Short-term load forecasting based on optimized random forest and optimal feature selection[J]. Energies, 2024, 17(8): 1926.
[20] 薛建凯. 一种新型的群智能优化技术的研究与应用[D]. 上海: 东华大学, 2020.
Xue J K.Research and application of a novel swarm intelligence optimization technique[D]. Shanghai: Donghua University, 2020.
[21] 王玲芝, 李晨阳, 刘婧, 等. 基于GRO-SSA-LSTM的短期光伏发电功率预测[J]. 太阳能学报, 2025, 46(2): 401-409.
Wang L Z, Li C Y, Liu J, et al.Short term photovoltaic power prediction based on GRO-SSA-LSTM[J]. Acta Energiae Solaris Sinica, 2025, 46(2): 401-409.
[22] 金伟勇, 卢丽娜, 赖欢欢, 等. 基于功率特征的K-ISSA-LSTM光伏功率预测[J]. 太阳能学报, 2024, 45(2): 429-434.
Jin W Y, Lu L N, Lai H H, et al.K-ISSA-LSTM photovoltaic power prediction based on power characteristic[J]. Acta Energiae Solaris Sinica, 2024, 45(2): 429-434.
[23] 周泽楷, 侯宏娟, 孙莉, 等. 基于CNN和BiLSTM神经网络模型的太阳能供暖负荷预测研究[J]. 太阳能学报, 2024, 45(10): 415-422.
Zhou Z K, Hou H J, Sun L, et al.Research on solar heating load forecasting based on CNN and BiLSTM neural network model[J]. Acta Energiae Solaris Sinica, 2024, 45(10): 415-422.
[24] 崔杨, 朱晗, 王议坚, 等. 基于CNN-SAEDN-Res的短期电力负荷预测方法[J]. 电力自动化设备, 2024, 44(4): 164-170.
Cui Y, Zhu H, Wang Y J, et al.Short-term power load forecasting method based on CNN-SAEDN-Res[J]. Electric Power Automation Equipment, 2024, 44(4): 164-170.
[25] 陈纬楠, 胡志坚, 岳菁鹏, 等. 基于长短期记忆网络和LightGBM组合模型的短期负荷预测[J]. 电力系统自动化, 2021, 45(4): 91-97.
Chen W N, Hu Z J, Yue J P, et al.Short-term load prediction based on combined model of long short-term memory network and light gradient boosting machine[J]. Automation of Electric Power Systems, 2021, 45(4): 91-97.
[26] 欧阳静, 杨吕, 尹康, 等. 基于ALIF-LSTM多任务学习的综合能源系统短期负荷预测[J]. 太阳能学报, 2022, 43(9): 499-507.
Ouyang J, Yang L, Yin K, et al.Short-term load forecasting method for integrated energy system based on ALIF-LSTM and multi-task learning[J]. Acta Energiae Solaris Sinica, 2022, 43(9): 499-507.
[27] 庞昊, 高金峰, 杜耀恒. 基于多神经网络融合的短期负荷预测方法[J]. 电力自动化设备, 2020, 40(6): 37-42.
Pang H, Gao J F, Du Y H.Short-term load forecasting method based on fusion of multiple neural networks[J]. Electric Power Automation Equipment, 2020, 40(6): 37-42.
[28] 田书欣, 刘浪, 魏书荣, 等. 基于改进灰狼优化算法的配电网动态重构[J]. 电力系统保护与控制, 2021, 49(16): 1-11.
Tian S X, Liu L, Wei S R, et al.Dynamic reconfiguration of a distribution network based on an improved grey wolf optimization algorithm[J]. Power System Protection And Control, 2021, 49(16): 1-11.
[29] 赵鑫, 王东丽, 彭泓, 等. 基于多策略改进蜣螂算法优化的变压器故障诊断[J]. 电力系统保护与控制, 2024, 52(6): 120-130.
Zhao X, Wang D L, Peng H, et al.Transformer fault diagnosis based on a multi-strategy improved dung beetle optimizer[J]. Power System Protection and Control, 2024, 52(6): 120-130.
[30] 张国治, 陈康, 方荣行, 等. 基于DGA与鲸鱼算法优化LogitBoost-决策树的变压器故障诊断方法[J]. 电力系统保护与控制, 2023, 51(7): 63-72.
Zhang G Z, Chen K, Fang R H, et al.Transformer fault diagnosis based on DGA and a whale algorithm optimizing a LogitBoost-decision tree[J]. Power System Protection and Control, 2023, 51(7): 63-72.
[31] 郑聪, 周海峰, 郑东强, 等. 基于改进多元宇宙算法的主动配电网故障定位方法研究[J]. 电力系统保护与控制, 2023, 51(2): 169-179.
Zheng C, Zhou H F, Zheng D Q, et al.An active distribution network fault location method based on improved multi-universe algorithm[J]. Power System Protection and Control, 2023, 51(2): 169-179.
[32] 王育飞, 付玉超, 薛花. 基于Chaos-EEMD-PFBD分解和GA-BP神经网络的光伏发电功率超短期预测法[J]. 太阳能学报, 2020, 41(12): 55-62.
Wang Y F, Fu Y C, XUE H.Ultra-short-term forecasting method of photovoltaic power generation based on Chaos-EEMD-PFBD decomposition and GA-BP neural networks[J]. Acta Energiae Solaris Sinica, 2020, 41(12): 55-62.
[33] 邹智, 吴铁洲, 张晓星, 等. 基于贝叶斯优化CNN-BiGRU混合神经网络的短期负荷预测[J]. 高电压技术, 2022, 48(10): 3935-3945.
Zou Z, Wu T Z, Zhang X X, et al.Short-term load forecast based on Bayesian optimized CNN-BiGRU hybrid neural networks[J]. High Voltage Engineering, 2022, 48(10): 3935-3945.
[34] 张雲钦, 程起泽, 蒋文杰, 等. 基于EMD-PCA-LSTM的光伏功率预测模型[J]. 太阳能学报, 2021, 42(9): 62-69.
Zhang Y Q, Cheng Q Z, Jiang W J, et al.Photovoltaic power prediction model based on EMD-PCA-LSTM[J]. Acta Energiae Solaris Sinica, 2021, 42(9): 62-69.
[35] 王晓霞, 俞敏, 冀明, 等. 基于气候相似性与SSA-CNN-LSTM的光伏功率组合预测[J]. 太阳能学报, 2023, 44(6): 275-283.
Wang X X, Yu M, Ji M, et al.Photovoltaic power combination forecasting based on climate similarity and SSA-CNN-LSTM[J]. Acta Energiae Solaris Sinica, 2023, 44(6): 275-283.