Side-by-side Chinese-English

基于深度长短时记忆网络的区域级超短期负荷预测方法

【摘要】超短期负荷预测为实时电力市场运行提供重要依据, 预测准确度的提升对于揭示负荷变化的不确定性以及日前预测偏差具有重要意义。基于电力系统中含有的丰富大数据资源, 提出了一种针对区域级负荷的深度长短时记忆网络超短期预测方法, 该方法包括输入数据的预处理、深度长短时记忆 (long short-termmemory, LSTM) 网络的构建以及模型的训练和超参数的寻找等步骤。其中采用随机搜索的方法寻找最优超参数, 并在该超参数下选择泛化能力最优的模型, 与前沿机器学习预测算法进行对比。实验结果证实, 深度LSTM网络可以取得更好的预测效果, 适合于离线训练实时预测。此外, 通过对隐藏层激活向量的可视化展示和相关关系定量计算, 首次直观展示了深度LSTM算法对负荷数据中含有的抽象特征提取情况, 证实了深度LSTM具有对输入负荷数据特征学习以及长短期相关性挖掘的能力。

【关键词】 超短期负荷预测;深度LSTM;循环神经网络;可视化;相关性;

【DOI】

【基金资助】 国家电网公司总部科技项目“基于多能互补的微网规划与运行控制技术研究及应用” (52060018000N) ;

A Very Short-term Load Forecasting Method Based on Deep LSTM RNN at Zone Level

【Abstract】Very short-term load forecasting provides a basis for a real-time electricity market. Improvement of its forecasting accuracy is important for revealing the uncertainties of the load and the deviations of day-ahead load forecasting results. Based on big data resources in power systems, a deep long short-term memory (LSTM) load forecasting method at a small area level is proposed, including preprocessing of input data, construction of a deep LSTM network, and search for hyper-parameters with the random search method. Under such hyper-parameters, comparison of forecasting performance between the proposed method and machine-learning method is conducted. Results prove that the proposed method can achieve better forecasting accuracy, which is suitable for offline training and online predicting. Also, through visualization of hidden-layer activation units and calculation of correlation, the ability of feature learning and correlation mining of a deep LSTM network is proven.

【Keywords】 very short-term load forecasting; deep LSTM; recurrent neural network (RNN) ; visualization; correlation analysis;

【DOI】

【Funds】 Science and Technology Foundation of SGCC (52060018000N) ;

Download this article
    References

    [1] Hong T, Fan S. Probabilistic electric load forecasting: a tutorial review [J]. International Journal of Forecasting, 2016, 32 (3): 914–938.

    [2] Wang Q, Zhang C, Ding Y, et al. Review of real-time electricity markets for integrating distributed energy resources and demand response [J]. Applied Energy, 2015 (138): 695–706.

    [3] Hong T, Wang P, White L. Weather station selection for electric load forecasting [J]. International Journal of Forecasting, 2015, 31 (2): 286–295.

    [4] Guo Siqi, Yuan Yue, ZhangXinsong, et, al. Energy management strategy of isolated microgird based on multi-time scale coordinated control [J]. Transactions of China Electrotechnical Society, 2014, 29 (2): 122–129 (in Chinese)..

    [5] Liu Siyuan, Ai Qian, Zhen Jianping, et, al. Bi-level coordination mechanism and operation strategy of multi-time scale multiple virtual power plants [J]. Proceedings of the CSEE, 2018, 38 (3): 753–761 (in Chinese).

    [6] Vaghefi A, Jafari M A, Emmanuel B, et, al. Modeling and forecasting of cooling and electricity load demand [J]. Applied Energy, 2014 (136): 186–196.

    [7] Zhao Feng, Sun Bo, Zhang Chenghui. Cooling, heating and electrical load forecasting method for CCHP system based on multivariate phase space reconstruction and Kalman Filter [J]. Proceedings of the CSEE, 2016, 36 (2): 399–406 (in Chinese).

    [8] Ma Jingbo, Yang Honggeng. Application of adaptive Kalman Filter in power system short-term load forecasting [J]. Power System Technology, 2005, 29 (1): 75–79 (in Chinese).

    [9] Cui Herui, Peng Xu. Summer short-term load forecasting based on ARIMAX model [J]. Power System Protection and Control, 2015, 43 (4): 108–114 (in Chinese).

    [10] Kong Xiangyu, Zheng Feng, E Zhijun, et al. Short-term load forecasting based on deep belief network [J]. Automation of Electric Power Systems, 2018, 42 (5): l33–139 (in Chinese).

    [11] Wang Gang, Jiang Jie, Tang Kunming, et, al. Ultra-short-term load forecasting based on adaptive bidirectional weighted least squares support vector machines [J]. Power System Protection and Control, 2010, 38 (19): 142–146 (in Chinese).

    [12] Wang Dewen, Sun Zhiwei. Big data analysis and parallel load forecasting of electric power user side [J]. Proceedings of the CSEE, 2015, 33 (3): 527–537 (in Chinese).

    [13] Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the inception architecture for computer vision [C] //The IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2016: 2818–2826.

    [14] Singh S P, Kumar A, Darbari H, et al. Machine translation using deep learning: an overview [C] //International Conference on Computer, Communications and Electronics. IEEE, 2017: 162–167.

    [15] Litjens G, Kooi T, Bejnordi B E, et al. A survey on deep learning in medical image analysis [J]. Medical Image Analysis, 2017, 42 (9): 60–88.

    [16] Lecun Y, Bengio Y, Hinton G. Deep learning [J]. Nature, 2015, 521 (7553): 436–444.

    [17] He K, Zhang X, Ren S, et al. Delving deep into rectifiers: surpassing human-level performance on image net classification [C] //The IEEEInternational Conference on Computer Vision (ICCV). IEEE, 2015: 1026–1034.

    [18] Lv Y, Duan Y, Kang W, et al. Traffic flow prediction with big data: a deep learning approach [J]. IEEE Transactions on Intelligent Transportation Systems, 2015, 16 (2): 865–873.

    [19] Shi Jiaqi, Tan Tao, Guo Jing, et al. Multi-task learning based on deep architecture for various types of load forecasting in regional energy system integration [J]. Power System Technology, 2018, 42 (3): 698–706 (in Chinese).

    [20] Mocanu E, Mocanu D C, Nguyen P H, et al. On-line building energy optimization using deep reinforcement learning [J]. IEEE Transactions on Smart Grid (Early Access), DOI: 10. 1109/TSG. 2018. 2834219.

    [21] Liu Wei, Zhang Dongxia, Wang Xinying, et, al. A decision making strategy for generating unit tripping under emergency circumstances based on deep reinforcement learning [J]. Proceedings of the CSEE, 2018, 38 (1): 109–119 (in Chinese).

    [22] Peng Xiaosheng, Deng Diyuan, Cheng Shijie, et, al. Key technologies of electric power big data and its application prospects in smart grid [J]. Proceedings of the CSEE, 2015, 35 (3): 503–511 (in Chinese).

    [23] Greff K, Srivastava R K, Koutnik J, et al. LSTM: a search space odyssey [J]. IEEE Transactions on Neural Networks & Learning Systems, 2016, 28 (10): 2222–2232.

    [24] Zhang Yuhang, Qiu Caiming, He Xing, et, al. A short-term load forecasting based on LSTM neural network [J]. Electric Power ICT, 2017, 15 (9): 19–25 (in Chinese).

    [25] Heng Shi, Minghao Xu, Ran Li. Deep learning for household load forecasting—a novel pooling deep RNN [J]. IEEE Transactions on Smart Grid (Early Access), DOI: 10. 1109/TSG. 2017. 2686012.

    [26] Kong W, Dong Z Y, Hill D J, et al. Short-term residential load forecasting based on resident behavior learning [J]. IEEE Transactions on Power Systems, 2018, 33 (1): 1087–1088.

    [27] Hong T, Pinson P, Fan S. Global energy forecasting competition2012 [J]. International Journal of Forecasting, 2014, 30 (2): 357–363.

    [28] James B, Yoshua B. Random search for hyper-parameter optimization [J]. Journal of Machine Learning Research, DOI: 10.1016/j.chemolab.2011.12.002.

    [29] Gregoire M, Genevieve B O, Klaus-Robert M, et, al. Neural networks: tricks of the trade [S]. Springer, 1998.

    [30] Pedregosa F, Varoquaux G, Gramfort A, et al. Scikit-learn: machine learning in Python [J]. Journal of Machine Learning Research, 2011 (12): 2825–2830

    [31] Van der M L, Hinton G. Visualizing data using t-SNE [J]. The Journal of Machine Learning Research, 2008 (9): 2579–2605.

This Article

ISSN:1000-3673

CN: 11-2410/TM

Vol 43, No. 06, Pages 1884-1892

June 2019

Downloads:0

Share
Article Outline

Abstract

  • 0 Introduction
  • 1 Process and data preprocessing of VSTLF
  • 2 Structure of load forecasting model based on deep LSTM
  • 3 Example analysis
  • 4 Experimental results
  • 5 Conclusion
  • Appendix A
  • References