Ultra Short-term Power Load Forecasting Based on Combined LSTM-XGBoost Model

被引：0
|
作者：
Chen Z. [1 ,2 ]
Liu J. [3 ]
Li C. [4 ]
Ji X. [4 ]
Li D. [1 ]
Huang Y. [1 ]
Di F. [1 ]
Gao X. [5 ]
Xu L. [6 ]
机构：
[1] Beijing Key Laboratory of Research and System Evaluation of Power Dispatching Automation Technology, China Electric Power Research Institute, Haidian District, Beijing
[2] Big Data Center, SGCC, Xicheng District, Beijing
[3] National Power Dispatch and Control Center, SGCC, Xicheng District, Beijing
[4] School of Information Engineering, China University of Geosciences, Haidian District, Beijing
[5] Institute of Microelectronics, Chinese Academy of Sciences, Chaoyang District, Beijing
[6] State Grid Zhejiang Electric Power Company, Hangzhou, 310007, Zhejiang Province
来源：
Dianwang Jishu/Power System Technology | 2020年 / 44卷 / 02期
关键词：
Combined model; Load forecasting; LSTM network; Power load; Ultra short-term; XGBoost;
D O I：
10.13335/j.1000-3673.pst.2019.1566
中图分类号：
学科分类号：
摘要：
Accurate power load forecasting provides effective and reliable guidance for power grid construction and operation, and plays a very important role in power system. In order to improve accuracy of power load forecasting, this paper proposes a combined forecast model based on LSTM (Long Short Term Memory Network) and XGBoost (eXtreme Gradient Boosting). The LSTM forecast model and the XGBoost forecast model are firstly established and the power load is predicted using the two models respectively. Then the combined model is used to predict the power load, using error reciprocal method to combine the results from the two single models. Through experimental verification of the power load data of 2016 Electrician Mathematical Contest in Modeling, the MAPE (mean absolute percentage error) error of the combined model is 0.57%, significantly lower than any single forecast model. Compared with GRU (gated recurrent unit) and XGBoost combined forecasting model, the combined forecasting model proposed in this paper has higher accuracy for ultra-short-term power load. © 2020, Power System Technology Press. All right reserved.
引用
收藏
页码：614 / 620
页数：6
相关论文
共 39 条
  • [1] Tao H., Shu F., Probabilistic electric load forecasting: a tutorial review, International Journal of Forecasting, 32, 3, pp. 914-938, (2016)
  • [2] Shi D., Application of outlier mining in power load forecasting, International Conference on Computer Application & System Modeling(ICCASM 2010), pp. V5-174-V5-176, (2010)
  • [3] Li H., Short term load forecasting by adaptive neural network, IOP Conference Series: Materials Science and Engineering
  • [4] Wang N., Fu P., Chen D., Et al., Application of big data method in optimal load dispatching of power plant, Proceedings of the CSEE, 35, 1, (2015)
  • [5] Zhang Z., Yang Z., Load derivation in short term forecasting using weather factor, Proceedings of the CSU-EPSA, 18, 5, pp. 79-83, (2006)
  • [6] Zhang Y., Qiu C., He X., Et al., A short-term load forecasting based on LSTM neural network, Electric Power Information and Communication Technology, 15, 9, pp. 19-25, (2017)
  • [7] Baharudin Z., Kamel N., Autoregressive method in short term load forecast, 2008 IEEE 2nd International Power and Energy Conference, pp. 1603-1608, (2008)
  • [8] Ji P., Xiong D., Wang P., Et al., A study on exponential smoothing model for load forecasting, 2012 Asia-Pacific Power and Energy Engineering Conference, pp. 2884-2887, (2012)
  • [9] Zhu X., Shen M., Based on the ARIMA model with grey theory for short term load forecasting model, 2012International Conference on Systems and Informatics(ICSAI2012), pp. 564-567, (2012)
  • [10] Ma W.M.W., Power system short-term load forecasting based on improved support vector machines, 2008 International Symposium on Knowledge Acquisition and Modeling, pp. 658-662, (2008)
1234