-
摘要:
短期负荷预测是电网合理调度和平稳运行的基础。为提高短期负荷预测精度,提出了一种基于Pearson相关系数(PCC)和长短期记忆(LSTM)神经网络的短期负荷预测方法。该方法运用Pearson相关性分析对原始多维输入变量组成的时间序列进行相关性分析,选取与电力负荷数据相关性较大的影响因素作为输入量,实现原始数据的降维和选优;再通过LSTM神经网络结合Adam优化算法,对与电力负荷相关性较大的影响因素和负荷实际输出序列之间的非线性关系建立网络模型。以嘉捷BOX和重庆丽苑维景国际大酒店的负荷数据作为实际算例,并与Prophet、LSTNet、门控循环(GRU)神经网络模型方法进行对比。结果表明:所提PCC-LSTM模型预测精度均在91%以上,最高可达95.44%,有效提高了负荷预测的精度。
-
关键词:
- Pearson相关系数 /
- 长短期记忆神经网络 /
- 负荷预测 /
- Adam算法 /
- 时间序列
Abstract:Short-term load forecasting is the basis for reasonable dispatch and smooth operation of the power grid. To improve the accuracy of short-term load forecasting, a method based on Pearson correlation coefficient (PCC) and long and short term memory (LSTM) neural network is proposed. This method uses Pearson correlation analysis to analyze the correlation of the time series composed of original multi-dimensional input variables, and selects the influencing factors with greater correlation with the power load data as the input to achieve the dimensionality reduction of the original data. Then, through the combination of LSTM neural network and Adam optimization algorithm, a network model is established for the nonlinear relationship between the influencing factors that have a greater correlation with the power load and the actual output sequence of the load. Taking the load data of Jiajie BOX and Chongqing International Grand Metropark Liyuan Hotel as calculation examples, the proposed model is compared with Prophet, LSTNet, and gated recurrent unit (GRU) models. Results show that the prediction accuracy of the proposed model is above 91%, with the highest up to 95.44%, thus effectively improving the accuracy of load forecasting.
-
表 1 相关系数范围
Table 1. Range of correlation coefficients
相关程度 系数范围 无关 0~0.2 弱相关 0.2~0.5 强相关 0.5~1 表 2 嘉捷BOX影响因素相关性分析
Table 2. Correlation analysis of influencing factors of Jiajie BOX
影响因素 工作日 节假日 天气 温度 湿度 用电量
相关性0.320** — -0.027 3 -0.222** -0.372** 注:“**”在0.01级别(双尾),相关性显著;“—”由于至少有一个变量为常量,因此无法进行计算。 表 3 重庆丽苑维景国际大酒店影响因素相关性分析
Table 3. Correlation analysis of influencing factors of Chongqing International Grand Metropark Liyuan Hotel
影响因素 工作日 节假日 天气 温度 湿度 用电量
相关性-0.022 7 — -0.089** 0.100** -0.313** 注:“**”在0.01级别(双尾),相关性显著;“—”由于至少有一个变量为常量,因此无法进行计算。 表 4 嘉捷BOX PCC-LSTM模型预测结果yMAPE和yRMSE对比
Table 4. Comparison of yMAPE and yRMSE of prediction results with PCC-LSTM model of Jiajie BOX
日期 yMAPE/% yRMSE 2020年12月1日 6.086 70.302 2020年12月2日 5.402 77.228 2020年12月3日 4.781 56.796 2020年12月4日 4.892 49.412 表 5 不同方法嘉捷BOX日负荷预测结果比较
Table 5. Comparison of results of daily load forecasting by different methods of Jiajie BOX
模型/方法 yMAPE/% yRMSE 计算时间/h Prophet 26.931 249.61 0.02 LSTNet 11.497 154.051 20.16 GRU 7.583 104.751 1.56 PCC-LSTM 4.562 54.685 2 表 6 重庆丽苑维景国际大酒店PCC-LSTM模型预测结果yMAPE和yRMSE对比
Table 6. Comparison of yMAPE and yRMSE of prediction results with PCC-LSTM model of Chongqing International Grand Metropark Liyuan Hotel
日期 yMAPE/% yRMSE 2020年12月3日 8.047 28.661 2020年12月4日 7.756 25.084 2020年12月5日 4.761 18.823 2020年12月6日 6.261 24.132 表 7 不同方法重庆丽苑维景国际大酒店日负荷预测结果比较
Table 7. Comparison of results of daily load forecasting by different methods of Chongqing International Grand Metropark Liyuan Hotel
模型/方法 yMAPE/% yRMSE 计算时间/h Prophet 13.021 49.532 0.02 LSTNet 18.689 60.934 21.1 GRU 5.069 23.03 2.53 PCC-LSTM 4.866 20.266 3.1 -
[1] 康重庆, 夏清, 刘梅. 电力系统负荷预测[M]. 北京: 中国电力出版社, 2007.KANG C Q, XIA Q, LIU M. Power system load forecasting[M]. Beijing: China Electric Power Press, 2007(in Chinese). [2] 杨英, 胡函武, 魏晗, 等. 电力系统负荷预测分类研究[J]. 电子世界, 2018(21): 109. https://www.cnki.com.cn/Article/CJFDTOTAL-ELEW201821072.htmYANG Y, HU H W, WEI H, et al. Research on classification of power system load forecast[J]. Electronics World, 2018(21): 109(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-ELEW201821072.htm [3] 康重庆, 夏清, 张伯明. 电力系统负荷预测研究综述与发展方向的探讨[J]. 电力系统自动化, 2004, 28(17): 1-11. doi: 10.3321/j.issn:1000-1026.2004.17.001KANG C Q, XIA Q, ZHANG B M. Review of power system load forecasting and its development[J]. Automation of Electric Power Systems, 2004, 28(17): 1-11(in Chinese). doi: 10.3321/j.issn:1000-1026.2004.17.001 [4] 程宇也. 基于人工神经网络的短期电力负荷预测研究[D]. 杭州: 浙江大学, 2017: 31-40.CHENG Y Y. Short-term electricity demand forecasting based on artificial neural network[D]. Hangzhou: Zhejiang University, 2017: 31-40(in Chinese). [5] YILDIZ B, BILBAO J I, SPROUL A B. A review and analysis of regression and machine learning models on commercial building electricity load forecasting[J]. Renewable and Sustainable Energy Reviews, 2017, 73: 1104-1122. doi: 10.1016/j.rser.2017.02.023 [6] BOX G G, BOX G, JENKINS G. Time series analysis: Forecasting and control[J]. Journal of Time, 1976, 31(4): 238-242. [7] 贺静, 韦钢, 熊玲玲. 负荷预测线性回归分析法的模糊改进[J]. 华东电力, 2003, 31(11): 21-23. https://www.cnki.com.cn/Article/CJFDTOTAL-HDDL200311007.htmHE J, WEI G, XIONG L L. Fuzzy improvement of linear regression analysis for load forecasting[J]. East China Electric Power, 2003, 31(11): 21-23(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-HDDL200311007.htm [8] 袁诚斌, 张涛, 朱建良. 基于MRA与回归分析法的短期电力负荷预测[J]. 信息技术, 2007, 31(10): 88-90. https://www.cnki.com.cn/Article/CJFDTOTAL-HDZJ200710030.htmYUAN C B, ZHANG T, ZHU J L. Short-term power load forecasting based on MRA and regression analysis[J]. Information Technology, 2007, 31(10): 88-90(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-HDZJ200710030.htm [9] 李强, 王中宇, 王卓然. 压力传感器动态校准不确定度评定[J]. 北京航空航天大学学报, 2015, 41(5): 847-856. doi: 10.13700/j.bh.1001-5965.2014.0356LI Q, WANG Z Y, WANG Z R. Uncertainty evaluation for the dynamic calibration of pressure transducer[J]. Journal of Beijing University of Aeronautics and Astronautics, 2015, 41(5): 847-856(in Chinese). doi: 10.13700/j.bh.1001-5965.2014.0356 [10] 郭松林, 水泉龙, 顾翔瑜. 灰色系统理论在负荷预测中运用综述[J]. 工业仪表与自动化装置, 2017(3): 24-27. https://www.cnki.com.cn/Article/CJFDTOTAL-GYZD201703006.htmGUO S L, SHUI Q L, GU X Y. Summary of application of gray system theory in load forecasting[J]. Industrial Instrumentation & Automation, 2017(3): 24-27(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-GYZD201703006.htm [11] 张国江, 邱家驹, 李继红. 基于模糊推理系统的多因素电力负荷预测[J]. 电力系统自动化, 2002, 26(5): 49-53. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT200205011.htmZHANG G J, QIU J J, LI J H. Multi-factor short-term load forecasting based on fuzzy inference system[J]. Automation of Electric Power Systems, 2002, 26(5): 49-53(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT200205011.htm [12] 艾欣, 周志宇, 魏妍萍, 等. 基于自回归积分滑动平均模型的可转移负荷竞价策略[J]. 电力系统自动化, 2017, 41(20): 26-31. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201720004.htmAI X, ZHOU Z Y, WEI Y P, et al. Bidding strategy for time-shiftable loads based on autoregressive integrated moving average model[J]. Automation of Electric Power Systems, 2017, 41(20): 26-31(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201720004.htm [13] 刘龙龙. 基于SARIMA与SVR的短期电力负荷预测[D]. 抚州: 东华理工大学, 2018: 7-14.LIU L L. Short-term power load forecasting based on SARIMA and SVR[D]. Fuzhou: East China Institute of Technology, 2018: 7-14(in Chinese). [14] 吴潇雨, 和敬涵, 张沛, 等. 基于灰色投影改进随机森林算法的电力系统短期负荷预测[J]. 电力系统自动化, 2015, 39(12): 50-55. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201512009.htmWU X Y, HE J H, ZHANG P, et al. Power system short-term load forecasting based on improved random forest with grey relation projection[J]. Automation of Electric Power Systems, 2015, 39(12): 50-55(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201512009.htm [15] 李元诚, 方廷健, 于尔铿. 短期负荷预测的支持向量机方法研究[J]. 中国电机工程学报, 2003, 23(6): 55-59. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGDC200306010.htmLI Y C, FANG T J, YU E K. Study of support vector machines for short-term load forecasting[J]. Proceedings of the Chinese Society for Electrical Engineering, 2003, 23(6): 55-59(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-ZGDC200306010.htm [16] 张涛, 赵登福, 周琳, 等. 基于RBF神经网络和专家系统的短期负荷预测方法[J]. 西安交通大学学报, 2001, 35(4): 331-334. https://www.cnki.com.cn/Article/CJFDTOTAL-XAJT200104000.htmZHANG T, ZHAO D F, ZHOU L, et al. Short-term load forecasting using radial basis function networks and expert system[J]. Journal of Xi'an Jiaotong University, 2001, 35(4): 331-334(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-XAJT200104000.htm [17] 李龙, 魏靖, 黎灿兵, 等. 基于人工神经网络的负荷模型预测[J]. 电工技术学报, 2015, 30(8): 225-230. https://www.cnki.com.cn/Article/CJFDTOTAL-DGJS201508028.htmLI L, WEI J, LI C B, et al. Prediction of load model based on artificial neural network[J]. Transactions of China Electrotechnical Society, 2015, 30(8): 225-230(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DGJS201508028.htm [18] 张倩, 马愿, 李国丽, 等. 频域分解和深度学习算法在短期负荷及光伏功率预测中的应用[J]. 中国电机工程学报, 2019, 39(8): 2221-2230. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGDC201908005.htmZHANG Q, MA Y, LI G L, et al. Applications of frequency domain decomposition and deep learning algorithms in short-term load and photovoltaic power forecasting[J]. Proceedings of the Chinese Society for Electrical Engineering, 2019, 39(8): 2221-2230(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-ZGDC201908005.htm [19] 彭文, 王金睿, 尹山青. 电力市场中基于Attention-LSTM的短期负荷预测模型[J]. 电网技术, 2019, 43(5): 1745-1751. https://www.cnki.com.cn/Article/CJFDTOTAL-DWJS201905032.htmPENG W, WANG J R, YIN S Q. Short-term load forecasting model based on attention-LSTM in electricity market[J]. Power System Technology, 2019, 43(5): 1745-1751(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DWJS201905032.htm [20] 杨甲甲, 刘国龙, 赵俊华, 等. 采用长短期记忆深度学习模型的工业负荷短期预测方法[J]. 电力建设, 2018, 39(10): 20-27. https://www.cnki.com.cn/Article/CJFDTOTAL-DLJS201810006.htmYANG J J, LIU G L, ZHAO J H, et al. A long short term memory based deep learning method for industrial load forecasting[J]. Electric Power Construction, 2018, 39(10): 20-27(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLJS201810006.htm [21] 陈锦鹏, 胡志坚, 陈纬楠, 等. 二次模态分解组合DBiLSTM-MLR的综合能源系统负荷预测[J]. 电力系统自动化, 2021, 45(13): 85-94. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT202113010.htmCHEN J P, HU Z J, CHEN W N, et al. Load prediction of integrated energy system based on combination of quadratic modal decomposition and deep bidirectional long short-term memory and multiple linear regression[J]. Automation of Electric Power Systems, 2021, 45(13): 85-94(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT202113010.htm [22] 赵倩, 郑贵林. 基于WD-LSSVM-LSTM模型的短期电力负荷预测[J/OL]. 电测与仪表, 2021(2021-02-26)[2021-03-21]. http://kns.cnki.net/kcms/detail/23.1202.TH.20210226.1041.004.html.ZHAO Q, ZHENG G L. Short-term load forecasting based on WD-LSSVM-LSTM model[J/OL]. Electrical Measurement and Instrumentation, 2021(2021-02-26)[2021-03-21]. http://kns.cnki.net/kcms/detail/23.1202.TH.20210226.1041.004.html(in Chinese). [23] 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. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201908018.htmLU J X, ZHANG Q P, YANG Z H, et al. Short-term load forecasting method based on CNN-LSTM hybrid neural network model[J]. Automation of Electric Power Systems, 2019, 43(8): 131-137(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201908018.htm [24] 王增平, 赵兵, 纪维佳, 等. 基于GRU-NN模型的短期负荷预测方法[J]. 电力系统自动化, 2019, 43(5): 53-58. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201905008.htmWANG Z P, ZHAO B, JI W J, et al. Short-term load forecasting method based on GRU-NN model[J]. Automation of Electric Power Systems, 2019, 43(5): 53-58(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201905008.htm [25] FENG M C, ZHENG J B, REN J C, et al. Big data analytics and mining for effective visualization and trends forecasting of crime data[J]. IEEE Access, 2019, 7: 106111-106123. [26] HUANG B, LIANG Y Y, QIU X L. Wind power forecasting using attention-based recurrent neural networks: A comparative study[J]. IEEE Access, 2021, 9: 40432-40444. [27] AFRASIABI M, MOHAMMADI M, RASTEGAR M, et al. Adv-anced deep learning approach for probabilistic wind speed forecasting[J]. IEEE Transactions on Industrial Informatics, 2021, 17(1): 720-727. -