基于特征加权Stacking集成学习的净负荷预测方法

doi:10.12204/j.issn.1000-7229.2022.09.011

电力建设 ›› 2022, Vol. 43 ›› Issue (9): 104-116.doi: 10.12204/j.issn.1000-7229.2022.09.011

• 新型电力系统下配电网规划与运行优化关键技术研究及应用·栏目主持王守相教授、赵倩宇博士· • 上一篇下一篇

基于特征加权Stacking集成学习的净负荷预测方法

鲍海波¹(), 吴阳晨²(), 张国应¹(), 李江伟¹(), 郭小璇³(), 黎静华²()

1.广西电网有限责任公司南宁供电局,南宁市 530031
2.广西电力系统最优化与节能技术重点实验室(广西大学),南宁市 530004
3.广西电网有限责任公司电力科学研究院,南宁市 530023

收稿日期:2022-01-23 出版日期:2022-09-01 发布日期:2022-08-31
作者简介:鲍海波(1988),男,博士,高级工程师,研究方向为电力系统优化运行,E-mail: baohaibop50514007@163.com;
吴阳晨(1997),女,硕士研究生,主要研究方向为新能源电力系统预测,E-mail: 425044089@qq.com ;
张国应(1990),男,本科,工程师,研究方向为配网运行管理,E-mail: 442319633@qq.com;
李江伟(1987),男,本科,工程师,研究方向为配网运行管理,E-mail: lijw0476@gx.csg.cn;
郭小璇(1986),女,博士,高级工程师,研究方向为电力需求侧管理、电能替代等,E-mail: guo_xiaoxuan@163.com;
黎静华(1982),女,博士,教授,博士生导师,主要研究方向为电力系统优化运行与控制、大规模新能源并网技术等,E-mail: happyjinghua@163.com。
基金资助:
中国南方电网公司科技项目(GXKJXM20190717)

Net Load Forecasting Method Based on Feature-Weighted Stacking Ensemble Learning

BAO Haibo¹(), WU Yangchen²(), ZHANG Guoying¹(), LI Jiangwei¹(), GUO Xiaoxuan³(), LI Jinghua²()

1. Nanning Power Supply Bureau, Guangxi Power Grid Corporation, Nanning 530031, China
2. Guangxi Key Laboratory of Power System Optimization and Energy Technology (Guangxi University), Nanning 530004, China
3. Electric Power Research Institute, Guangxi Power Grid Corporation, Nanning 530023, China

Received:2022-01-23 Online:2022-09-01 Published:2022-08-31
Supported by:
China Southern Power Grid Corporation of China Research Program(GXKJXM20190717)

摘要/Abstract

摘要：

当前用户侧接入大量新能源,通常采用实际电力负荷减去新能源发电功率后的负荷(称为“净负荷”)进行预测研究。由于新能源发电随机性强,净负荷具有不确定性强、规律性差的特点,难以准确预测。为此,提出了一种基于特征加权改进的Stacking集成学习净负荷预测方法。首先,通过对不同单一预测模型的预测性能和差异性的分析,优选出长短期记忆网络、Elman神经网络、随机森林树和最小二乘支持向量机作为Stacking集成的学习器。其次,针对传统Stacking集成预测由于忽略学习器之间差异性所导致的预测能力不足问题,根据预测精度对不同学习器进行特征加权,以修正各学习器所带来的整体预测误差。最后,基于德国TENNET区域实测数据进行算例分析,结果表明,相比于单一预测模型、传统Stacking集成预测方法,基于特征加权Stacking集成学习的净负荷预测方法在晴天、多云、多雨、多雪等天气情况下均具有更高的预测精度。

关键词: 新能源, 净负荷预测, Stacking集成算法, 特征加权

Abstract:

At present, a large amount of new energy is connected to the user side, and the actual power load minus the power generated by the new energy (hereinafter referred to as “net load”) is usually used for prediction research. Due to the strong randomness of new energy power generation, net load has strong uncertainty and poor regularity, which makes it difficult to predict accurately. To this end, this paper proposes a net load prediction method based on the feature-weighted Stacking ensemble algorithm. Firstly, through the analysis of the prediction performance and difference of different prediction models, this paper chooses Long Short-Term Memory Network, Elman Neural Network, Random Forest Tree and Least Squares Support Vector Machine as stacking ensemble learners. Secondly, because the traditional Stacking ensemble prediction model ignores the differences between learners, the model’s prediction ability is insufficient. Therefore, this paper weights the features of the learners according to the prediction accuracy to correct the prediction error introduced by different learners. Finally, the measured data in the German TENNET area is analyzed as an example. The simulation results show that, compared with the single forecasting model and the traditional Stacking integrated forecasting method, the payload prediction method based on feature-weighted stacking ensemble learning has higher forecasting accuracy in sunny, cloudy, rainy, snowy and other weather conditions.

Key words: renewable energy, net load forecasting, stacking ensemble learning, feature weighting

中图分类号:

TM714

鲍海波, 吴阳晨, 张国应, 李江伟, 郭小璇, 黎静华. 基于特征加权Stacking集成学习的净负荷预测方法[J]. 电力建设, 2022, 43(9): 104-116.

BAO Haibo, WU Yangchen, ZHANG Guoying, LI Jiangwei, GUO Xiaoxuan, LI Jinghua. Net Load Forecasting Method Based on Feature-Weighted Stacking Ensemble Learning[J]. ELECTRIC POWER CONSTRUCTION, 2022, 43(9): 104-116.

图/表 17

图1 Stacking集成预测框架过程

Fig.1 The framework for Stacking ensemble forecasting

图2 k折交叉验证流程图

Fig.2 Schematic diagram of k-fold cross validation

图3 基于特征加权Stacking集成净负荷预测方法

Fig.3 Net load forecasting based on feature-weighted Stacking framework

表1 仿真场景

Table 1 The simulation scene

表2 各单一预测模型的MAPE

Table 2 The MAPE of single prediction model %

图4 各模型预测误差的相关性分析

Fig.4 Correlation analysis of forecasting error for each model

表3 不同预测方法对12月净负荷预测的MAPE

Table 3 MAPE of different prediction models on December

图5 不同预测方法对12月24 h净负荷预测的MAPE

Fig.5 MAPE of net load forecast for the 24 hours of December by different forecasting methods

图6 不同预测方法对周末净负荷的预测结果

Fig.6 Comparison between various prediction models over the weekend

图7 不同预测方法对晴天净负荷的预测结果

Fig.7 Comparison between various prediction models on sunny days

表4 不同预测方法对晴天净负荷的预测误差

Table 4 The MAPE of prediction models for sunny days

图8 不同预测方法对多云日净负荷的预测结果

Fig.8 Comparison between various prediction models on cloudy days

表5 不同预测方法对多云天气净负荷的预测误差

Table 5 The MAPE of prediction models for cloudy days

图9 不同预测方法对多雪日净负荷的预测结果

Fig.9 Comparison between various prediction models on snowy days

表6 不同预测方法对多雪天气净负荷的预测误差

Table 6 The MAPE of prediction models for snowy days

图10 不同预测方法对多雨净负荷的预测结果

Fig.10 Comparison between various prediction models on rainy days

表7 不同预测方法对多雨天气净负荷的预测误差

Table 7 The MAPE of prediction models for rainy days

参考文献 25

[1]	韦善阳, 黎静华, 黄乾, 等. 考虑多重因素耦合的广义负荷特征曲线的模式分析[J]. 电力系统自动化, 2021, 45(1): 114-122.
	WEI Shanyang, LI Jinghua, HUANG Qian, et al. Pattern analysis of generalized load characteristic curve considering coupling of multiple factors[J]. Automation of Electric Power Systems, 2021, 45(1): 114-122.
[2]	周志华. 机器学习[M]. 北京: 清华大学出版社, 2016.
[3]	廖旎焕, 胡智宏, 马莹莹, 等. 电力系统短期负荷预测方法综述[J]. 电力系统保护与控制, 2011, 39(1): 147-152.
	LIAO Nihuan, HU Zhihong, MA Yingying, et al. Review of the short-term load forecasting methods of electric power system[J]. Power System Protection and Control, 2011, 39(1): 147-152.
[4]	黄静, 陆春良, 张俊, 等. 计及分布式能源影响的母线净负荷预测方法研究[J]. 电气自动化, 2018, 40(1): 63-66.
	HUANG Jing, LU Chunliang, ZHANG Jun, et al. Study on bus net load forecast method considering the influence of distributed energy[J]. Electrical Automation, 2018, 40(1): 63-66.
[5]	刘友波, 吴浩, 刘挺坚, 等. 集成经验模态分解与深度学习的用户侧净负荷预测算法[J]. 电力系统自动化, 2021, 45(24): 57-64.
	LIU Youbo, WU Hao, LIU Tingjian, et al. User-side net load forecasting method integrating empirical mode decomposition and deep learning[J]. Automation of Electric Power Systems, 2021, 45(24): 57-64.
[6]	颜宏文, 李欣然. 基于差分进化的含分布式电源母线净负荷预测[J]. 电网技术, 2013, 37(6): 1602-1606.
	YAN Hongwen, LI Xinran. Uncertainty analysis on net load forecasting for busbar containing distributed energy sources based on differential evolution and rough sets reduction[J]. Power System Technology, 2013, 37(6): 1602-1606.
[7]	史佳琪, 马丽雅, 李晨晨, 等. 基于串行-并行集成学习的高峰负荷预测方法[J]. 中国电机工程学报, 2020, 40(14): 4463-4472.
	SHI Jiaqi, MA Liya, LI Chenchen, et al. Daily peak load forecasting based on sequential-parallel ensemble learning[J]. Proceedings of the CSEE, 2020, 40(14): 4463-4472.
[8]	史佳琪, 张建华. 基于多模型融合Stacking集成学习方式的负荷预测方法[J]. 中国电机工程学报, 2019, 39(14): 4032-4042.
	SHI Jiaqi, ZHANG Jianhua. Load forecasting based on multi-model by stacking ensemble learning[J]. Proceedings of the CSEE, 2019, 39(14): 4032-4042.
[9]	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.
	LU Jixiang, ZHANG Qipei, YANG Zhihong, 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.
[10]	王哲, 万宝, 凌天晗, 等. 基于谱聚类和LSTM神经网络的电动公交车充电负荷预测方法[J]. 电力建设, 2021, 42(6): 58-66.
	WANG Zhe, WAN Bao, LING Tianhan, et al. Electric bus charging load forecasting method based on spectral clustering and LSTM neural network[J]. Electric Power Construction, 2021, 42(6): 58-66.
[11]	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.
	YANG Jiajia, LIU Guolong, ZHAO Junhua, et al. A long short term memory based deep learning method for industrial load forecasting[J]. Electric Power Construction, 2018, 39(10): 20-27.
[12]	KONG W, DONG Z Y, JIA Y, et ai. Short-term residential load forecasting based on LSTM recurrent neural network[J]. IEEE Transactions on Smart Grid, 2019, 10(1): 841-851. doi: 10.1109/TSG.2017.2753802 URL
[13]	李兵, 韩睿, 何怡刚, 等. 改进随机森林算法在电机轴承故障诊断中的应用[J]. 中国电机工程学报, 2020, 40(4): 1310-1319.
	LI Bing, HAN Rui, HE Yigang, et al. Applications of the improved random forest algorithm in fault diagnosis of motor bearings[J]. Proceedings of the CSEE, 2020, 40(4): 1310-1319.
[14]	李焱, 贾雅君, 李磊, 等. 基于随机森林算法的短期电力负荷预测[J]. 电力系统保护与控制, 2020, 48(21): 117-124.
	LI Yan, JIA Yajun, LI Lei, et al. Short term power load forecasting based on a stochastic forest algorithm[J]. Power System Protection and Control, 2020, 48(21): 117-124.
[15]	梅贵琴. 改进的Elman神经网络和网络参数优化算法研究[D]. 重庆: 西南大学, 2017.
	MEI Guiqin. The research on improved Elman neural network and parameter optimization algorithm[D]. Chongqing: Southwest University, 2017.
[16]	唐玮, 钟士元, 舒娇, 等. 基于GRA-LSSVM的配电网空间负荷预测方法研究[J]. 电力系统保护与控制, 2018, 46(24): 76-82.
	TANG Wei, ZHONG Shiyuan, SHU Jiao, et al. Research on spatial load forecasting of distribution network based on GRA-LSSVM method[J]. Power System Protection and Control, 2018, 46(24): 76-82.
[17]	YANG H Z, JIANG Z Y, LI M L, et al. Short-term load forecasting based on variational mode decomposition and least squares support vector machine by improved artificial fish swarm-shuffled frog jump algorithms[J]. International Journal of Performability Engineering, 2019, 15(12): 3117. doi: 10.23940/ijpe.19.12.p3.31173128 URL
[18]	李霄, 王昕, 郑益慧, 等. 基于改进最小二乘支持向量机和预测误差校正的短期风电负荷预测[J]. 电力系统保护与控制, 2015, 43(11): 63-69.
	LI Xiao, WANG Xin, ZHENG Yihui, et al. Short-term wind load forecasting based on improved LSSVM and error forecasting correction[J]. Power System Protection and Control, 2015, 43(11): 63-69.
[19]	徐慧丽. Stacking算法的研究及改进[D]. 广州: 华南理工大学, 2018.
	XU Huili. The study and improvement of stacking[D]. Guangzhou: South China University of Technology, 2018.
[20]	林军. 基于特征加权Stacking算法的教学评价情感分类模型研究[D]. 广州: 华南理工大学, 2020.
	LIN Jun. Research on semantic classification model of teaching evaluation based on feature weighted stacking algorithm[D]. Guangzhou: South China University of Technology, 2020.
[21]	DEMİR N, DALKILIÇ G. Modified stacking ensemble approach to detect network intrusion[J]. Turkish Journal of Electrical Engineering & Computer Sciences, 2018, 26: 418-433.
[22]	邓威, 郭钇秀, 李勇, 等. 基于特征选择和Stacking集成学习的配电网网损预测[J]. 电力系统保护与控制, 2020, 48(15): 108-115.
	DENG Wei, GUO Yixiu, LI Yong, et al. Power losses prediction based on feature selection and Stacking integrated learning[J]. Power System Protection and Control, 2020, 48(15): 108-115.
[23]	Tennet. System and transmission data[EB/OL]. (2017-01-31)[2022-01-23]. https://www.tennet.org/english.
[24]	Stefan Pfenninger. Renewables ninja[EB/OL]. (2019-04-23)[2022-01-23]. https://www.renewables.ninja/.
[25]	沈兆轩, 袁三男. 利用卷积神经网络支持向量回归机的地区负荷聚类集成预测[J]. 电网技术, 2020, 44(6): 2237-2244.
	SHEN Zhaoxuan, YUAN Sannan. Regional load clustering integration forecasting based on convolutional neural network support vector regression machine[J]. Power System Technology, 2020, 44(6):2237-2244.

基于特征加权Stacking集成学习的净负荷预测方法

Net Load Forecasting Method Based on Feature-Weighted Stacking Ensemble Learning

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 17

参考文献 25

相关文章 15

编辑推荐

Metrics

本文评价

[1]	刁利, 李光, 宋雪莹, 德格吉日夫, 谭忠富. 新能源分比例渗透下基于博弈的多主体效益综合评估方法[J]. 电力建设, 2022, 43(6): 43-55.
[2]	张兴平, 何澍, 王泽嘉, 张浩楠, 张又中. 不同新能源渗透率下燃煤机组行为策略分析[J]. 电力建设, 2022, 43(5): 9-17.
[3]	刘飞, 赵澄颢, 王世斌, 田旭, 刘联涛, 朱晓荣. 新能源直流汇集分群综合协调控制策略[J]. 电力建设, 2022, 43(5): 127-136.
[4]	杨立滨, 张磊, 刘艳章, 李正曦, 宗鸣. 基于分布式框架的新能源场站并网性能评估[J]. 电力建设, 2022, 43(5): 137-144.
[5]	康思伟, 董文凯, 郭诗然, 孙红军, 谢小荣. 基于虚拟同步机控制的新能源发电并网系统小干扰稳定临界短路比[J]. 电力建设, 2022, 43(3): 131-140.
[6]	李建宜, 李鹏, 徐晓春, 施儒昱, 曾平良, 夏辉. 基于综合概率模型与深度学习的智能电网功率-电压映射方法[J]. 电力建设, 2022, 43(2): 37-44.
[7]	刘一民, 王书扬, 李彬, 王兴国, 郑少明, 董鹏. 逆变型新能源场站柔性直流送出系统交流线路差动保护灵敏性优化方案[J]. 电力建设, 2022, 43(1): 63-69.
[8]	余洋, 贾浩, 陈启维, 范亚洲, 米增强, 袁玉宝, 常生强. 基于改进模型预测控制的电-气系统新能源功率波动平滑策略[J]. 电力建设, 2021, 42(9): 65-73.
[9]	任景, 薛晨, 马晓伟, 崔伟, 孟鑫羽, 杨迎, 王剑晓. 考虑能量存取的自备电厂去中心化交易机制探究[J]. 电力建设, 2021, 42(6): 1-8.
[10]	李晓萌, 张忠, 赵华, 饶宇飞, 张振安, 朱全胜, 吕泉. 能源互联网背景下电网备用问题探索[J]. 电力建设, 2021, 42(5): 57-68.
[11]	张涛, 胡娱欧, 张晶, 韩亮, 王佩雯, 何淼, 陈雨果. 适应高比例新能源发展的电力市场容量保障机制分析[J]. 电力建设, 2021, 42(3): 117-125.
[12]	陈人杰, 李华取, 彭晓涛, 杨军, 董旭柱, 刘首文, 李祥杰. 基于改进TOPSIS的新能源大数据服务项目评价研究[J]. 电力建设, 2021, 42(3): 126-134.
[13]	何廷一, 李胜男, 陈亦平, 吴水军, 沐润志, 和鹏, 孟贤, 何鑫, 杨博, 曹璞璘. 高比例新能源电网多源最优协同调频策略[J]. 电力建设, 2021, 42(10): 51-59.
[14]	崔红芬, 王德顺, 杨波, 吴福保, 朱天天, 赵晶晶. 支撑电网黑启动的风光储新能源电站协调控制策略[J]. 电力建设, 2020, 41(9): 50-57.
[15]	赵波, 章德, 胡柳君, 马俊杰, 李勇, 王炜宇. 基于机电扰动传播特性的电网惯量分布辨识方法[J]. 电力建设, 2020, 41(8): 25-31.