准确预测综合能源系统中的各种负荷需求对其经济稳定运行至关重要。然而,用能形式的多样化发展使负荷演化趋势日益复杂,为准确预测带来严峻挑战。为此,提出一种结合时序重构技术、组合神经网络与集成学习的预测架构。时序重构技术由改进自适应噪声完备集合经验模态分解（ICEEMDAN）和样本熵算法（SE）组成,可将复杂的原始序列分解并重构为更平滑的子序列,以降低预测难度。组合神经网络结合了时间卷积网络（TCN）、门控循环单元（GRU）和注意力机制（AM）的优势,用于捕捉子序列的长期依赖关系。同时,通过AdaBoost算法增强模型的鲁棒性和泛化能力。最后,将所有子序列的预测结果累加得出最终结果。对比实验结果表明,所提模型的电、气、热和冷负荷预测效果均优于其他对比模型。此外,相较于GRU基准模型,所提模型的平均绝对百分比误差与均方根误差分别降低40.59%与40.03%。
Abstract
Accurate prediction of various load demands in an integrated energy system is crucial to its stable economic operation. However, the diversified development of energy use forms makes the load evolution trend increasingly complex, which brings severe challenges to accurate prediction. To this end, a prediction architecture combining time series reconstruction technology, a combined neural network and ensemble learning is proposed. The time series reconstruction technology is composed of improved complete ensemble empirical mode decomposition with adaptive noise （ICEEMDAN） and sample entropy algorithm （SE）, which decomposes and reconstructs the complex original sequence into smoother subsequences to reduce the difficulty of prediction. The combined neural network combines the advantages of temporal convolutional network （TCN）, gated recurrent unit （GRU） and attention mechanism（AM） to capture the long-term dependencies of subsequences. At the same time, the robustness and generalization ability of the model are enhanced by the adaptive Boosting algorithm. Finally, the prediction results of all subsequences are accumulated to obtain the final result. The results of comparative experiments show that the proposed model is superior to other comparative models in the prediction of electrical, gas, heat and cooling loads. In addition, compared with the GRU benchmark model, the mean absolute percentage error and root mean square error of the proposed model are reduced by 40.59% and 40.03%, respectively.
关键词
负荷预测 /
综合能源系统 /
时间序列 /
神经网络 /
集成学习
Key words
load forecasting /
integrated energy system /
time series /
neural network /
ensemble learning
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 倪建辉, 张菁. 基于多任务双层注意力优化的TCN-BiGRU综合能源负荷短期预测[J]. 控制工程, 2024, 31(11): 1924-1936.
Ni J H, Zhang J.Short-term load forecasting for integrated energy system based on TCN-BiGRU optimized by multi-task dual-layer attention[J]. Control Engineering of China, 2024, 31(11): 1924-1936.
[2] 黄文龙, 葛文超, 任洪波, 等. 全可再生能源多能互补系统优化配置与运行探索[J]. 太阳能学报, 2024, 45(5): 351-359.
Huang W L, Ge W C, Ren H B, et al.Exploration of optimal configuration and operation for all-renewable multi-energy complementary systems[J]. Acta Energiae Solaris Sinica, 2024, 45(5): 351-359.
[3] Naz A, Javaid N, Rasheed M B, et al.Game theoretical energy management with storage capacity optimization and photovoltaic cell generated power forecasting in micro grid[J]. Sustainability, 2019, 11(10): 2763.
[4] Fan J L, Zhuang W, Xia M, et al.Optimizing attention in a transformer for multihorizon, multienergy load forecasting in integrated energy systems[J]. IEEE Transactions on Industrial Informatics, 2024, 20(8): 10238-10248.
[5] 窦真兰, 张春雁, 许一洲, 等. 基于多变量相空间重构和径向基函数神经网络的综合能源系统电冷热超短期负荷预测[J]. 电网技术, 2024, 48(1): 121-128.
Dou Z L, Zhang C Y, Xu Y Z, et al.Ultra-short-term load forecasting of electrical, cooling and heating for integrated energy system based on multivariate phase space reconstruction and radial basis function neural network[J]. Power System Technology, 2024, 48(1): 121-128.
[6] 吕忠麟, 顾洁, 孟璐. 基于耦合特征与多任务学习的综合能源系统短期负荷预测[J]. 电力系统自动化, 2022, 46(11): 58-66.
Lyu Z L, Gu J, Meng L.Short-term load forecasting for integrated energy system based on coupling features and multi-task learning[J]. Automation of Electric Power Systems, 2022, 46(11): 58-66.
[7] Lee D H, Park S Y, Hong J C, et al.Analysis of the energy and environmental effects of green car deployment by an integrating energy system model with a forecasting model[J]. Applied Energy, 2013, 103: 306-316.
[8] 韩宝慧, 陆玲霞, 包哲静, 等. 基于多头概率稀疏自注意力模型的综合能源系统多元负荷短期预测[J]. 电力建设, 2024, 45(2): 127-136.
Han B H, Lu L X, Bao Z J, et al.Short-term forecasting of multienergy loads of integrated energy system based on multihead probabilistic sparse self-attention model[J]. Electric Power Construction, 2024, 45(2): 127-136.
[9] Yang Y H, Han L, Qiu C Y, et al.A short-term wave energy forecasting model using two-layer decomposition and LSTM-attention[J]. Ocean Engineering, 2024, 299: 117279.
[10] 苏磊. 基于GA-BP神经网络的微电网短期负荷预测[J]. 电工技术, 2023(12): 152-154.
Su L.Short term load forecasting of micro grid based on GA-BP neural network[J]. Electric Engineering, 2023(12): 152-154.
[11] 文彦飞, 王万雄. 基于FA-SVR-LSTM组合模型的短期电力负荷预测[J]. 电子科技, 2023, 36(9): 1-7.
Wen Y F, Wang W X.Short-term power load forecasting based on FA-SVR-LSTM combined model[J]. Electronic Science and Technology, 2023, 36(9): 1-7.
[12] Banik R, Biswas A.Enhanced renewable power and load forecasting using RF-XGBoost stacked ensemble[J]. Electrical Engineering, 2024, 106(4): 4947-4967.
[13] Abumohsen M, Owda A Y, Owda M.Electrical load forecasting using LSTM, GRU, and RNN algorithms[J]. Energies, 2023, 16(5): 2283.
[14] 吴晗. 基于改进多任务学习的区域综合能源系统负荷预测研究[D]. 徐州: 中国矿业大学, 2022.
Wu H.Research on forecasting loads of a regional integrated energy system based on improved multi-task learning[D]. Xuzhou: China University of Mining and Technology, 2022.
[15] 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.
[16] Wen X Y, Liao J C, Niu Q Y, et al.Deep learning-driven hybrid model for short-term load forecasting and smart grid information management[J]. Scientific Reports, 2024, 14: 13720.
[17] Mounir N, Ouadi H, Jrhilifa I.Short-term electric load forecasting using an EMD-BI-LSTM approach for smart grid energy management system[J]. Energy and Buildings, 2023, 288: 113022.
[18] Zhu Z W, Zhou M R, Hu F, et al.A day-ahead industrial load forecasting model using load change rate features and combining FA-ELM and the AdaBoost algorithm[J]. Energy Reports, 2023, 9: 971-981.
[19] 陈景文, 杨淇, 兰天一, 等. 基于集成经验模态分解和极限学习机的质子交换膜燃料电池寿命预测[J]. 太阳能学报, 2025, 46(2): 135-141.
Chen J W, Yang Q, Lan T Y, et al.Lifetime prediction of PEMFC based on ensemble empirical mode decomposition and extreme learning machine[J]. Acta Energiae Solaris Sinica, 2025, 46(2): 135-141.
[20] 张群, 侯玉强, 许剑冰, 等. 基于CEEMDAN和时间卷积网络的风向预测算法[J]. 太阳能学报, 2024, 45(10): 512-520.
Zhang Q, Hou Y Q, Xu J B, et al.Wind direction prediction algorithm based on CEEMDAN and timporal convolutional network[J]. Acta Energiae Solaris Sinica, 2024, 45(10): 512-520.
[21] 赵兵, 王增平, 纪维佳, 等. 基于注意力机制的CNN-GRU短期电力负荷预测方法[J]. 电网技术, 2019, 43(12): 4370-4376.
Zhao B, Wang Z P, Ji W J, et al.A short-term power load forecasting method based on attention mechanism of CNN-GRU[J]. Power System Technology, 2019, 43(12): 4370-4376.
[22] Lyu Y, Shi Y J, Li Z Y, et al.Short-term load forecasting method of IES based on RLA neural network with dual signal processing[J]. Energy and Buildings, 2024, 309: 114074.
PDF(4086 KB)
