| 673 | 8 | 452 |
| 下载次数 | 被引频次 | 阅读次数 |
为解决配变负荷日峰值预测精度不高、重过载预警误差大的问题，提出一种配变日峰值负荷预测及重过载预警方法。首先，基于时间卷积网络对配变日负荷进行预测；然后，通过贝叶斯优化极限梯度提升模型对配变日负荷峰值出现时刻及峰值区间幅值进行独立预测；最后，使用峰值预测补正日负荷预测结果并转化为预警等级，实现配变重过载预警。采用湖南某地区配电台区数据实例验证，结果表明，所提方法可实现配变日峰值负荷精确预测及准确预警重过载运行风险。
Abstract：To solve the problems of low accuracy of daily peak load forecasting and large error of heavy and overload warning for a distribution transformer,a method of daily peak load forecasting and heavy and overload warning for distribution transformer is proposed. First,the daily load of the distribution transformer is predicted based on temporal convolutional network. Then,the Bayesian optimization extreme gradient boosting model is used to independently predict the peak time and peak interval amplitude of daily load for distribution transformer. Finally,the peak prediction is used to correct the daily load forecasting result and convert it into a warning level,thus realizing the heavy and overload warning of the distribution transformer. The data of distribution station area in one region of Hunan Province is taken as an example for verification,and results show that the proposed method can accurately predict the daily peak load of distribution transformer and accurately warn the risk of heavy and overload operation.
[1]李玉志，刘晓亮，邢方方，等（Li Yuzhi,Liu Xiaoliang,Xing Fangfang,et al）.基于Bi-LSTM和特征关联性分析的日尖峰负荷预测（Daily peak load prediction based on correlation analysis and bi-directional long short-term memory network）[J].电网技术（Power System Technology）,2021,45(7):2719-2730.
[2]王文秀，田世明，王泽忠，等（Wang Wenxiu,Tian Shiming,Wang Zezhong,et al）.一种基于贝叶斯网络的电力负荷峰值预测方法（A power peak load forecasting method based on Bayesian network）[J].供用电（Distribution&Utilization）,2019,36(7):57-64.
[3]姜建，刘海琼，李衡，等（Jiang Jian,Liu Haiqiong,Li Heng,et al）.基于XGBoost的配电网线路峰值负荷预测方法（Peak load forecasting method of distribution network lines based on XGBoost）[J].电力系统保护与控制（Power System Protection and Control）,2021,49(16):119-127.
[4]马浩，贺朝会，王立斌，等（Ma Hao,He Chaohui,Wang Libin,et al）.基于用电信息采集大数据的公变负载状态评估（Load assessment of public transformer based on power consumption information acquisition and big data）[J].电测与仪表（Electrical Measurement&Instrumentation）,2020,57(14):99-105.
[5]金娜（Jin Na）.电力变压器运行状态监测系统研究与设计（Research and Design of Power Transformer Operation Monitoring System）[D].镇江：江苏大学（Zhenjiang:Jiangsu University）,2019.
[6]李富鹏，沈秋英，王森，等（Li Fupeng,Shen Qiuying,Wang Sen,et al）.基于大数据和多因素组合分析的单元制配电网精细化负荷预测（Refined load forecasting method for unit distribution network based on big data and multiple factors）[J].智慧电力（Smart Power）,2020,48(1):55-62.
[7]龚雪娇，朱瑞金，唐波（Gong Xuejiao,Zhu Ruijin,Tang Bo）.基于贝叶斯优化XGBoost的短期峰值负荷预测（Short-term peak load forecasting based on Bayesian optimization XGBoost）[J].电力工程技术（Electric Power Engineering Technology）,2020,39(6):76-81.
[8] Rallapalli S R,Ghosh S. Forecasting monthly peak demand of electricity in India-A critique[J]. Energy Policy,2012,45:516-520.
[9]邓带雨，李坚，张真源，等（Deng Daiyu,Li Jian,Zhang Zhenyuan,et al）.基于EEMD-GRU-MLR的短期电力负荷预测（Short-term electric load forecasting based on EEMD-GRU-MLR）[J].电网技术（Power System Technology）,2020,44(2):593-602.
[10]唱友义，孙赫阳，顾泰宇，等（Chang Youyi,Sun Heyang,Gu Taiyu,et al）.采用历史数据扩充方法的风力发电量月度预测（Monthly forecast of wind power generation using historical data expansion method）[J].电网技术（Power System Technology）,2021,45(3):1059-1068.
[11]张铁岩，孙天贺（Zhang Tieyan,Sun Tianhe）.计及季节与趋势因素的综合能源系统负荷预测（Loading forecast for integrated energy system considering season and trend factors）[J].沈阳工业大学学报（Journal of Shenyang University of Technology）,2020,42(5):481-487.
[12]蒋刚（Jiang Gang）.基于模糊支持向量核回归方法的短期峰值负荷预测（Short-term peak load fore casting based on fuzzy support bector kernel regression method）[J].控制理论与应用（Control Theory&Applications）,2007,24(6):986-990.
[13] Li Ming,Zhou Qin,Yang Zhihui. Distribution transformer mid-term heavy load and overload pre-warning based on logistic regression[C]//IEEE Eindhoven PowerTech. Eindhoven,Netherlands,2015.
[14] Ngo V,Wu Wenchuan,Zhang Boming,et al. Ultra-shortterm load forecasting using robust exponentially weighted method in distribution networks[J]. Journal of Information,Control and Management Systems,2015(9):301-308.
[15]沈渊彬，刘庆珍（Shen Yuanbin,Liu Qingzhen）.基于卡尔曼滤波-SVR时刻峰值的短期负荷预测（Short-term peak load forecasting of every time forecasting based on Kalman filtering-SVR）[J].电气开关（Electric Switchgear）,2016,54(2):35-38.
[16]张国宾，王晓蓉，邓春宇（Zhang Guobin,Wang Xiaorong,Deng Chunyu）.基于关联分析与机器学习的配网台区重过载预测方法（Prediction method in distribution transformer heavy and overload supply areas with relevance analyze and machine learning）[J].大数据（Big Data Research）,2018,4(1):105-116.
[17]吕志星，张虓，王沈征，等（LüZhixing,Zhang Xiao,Wang Shenzheng,et al）.基于K-Means和CNN的用户短期电力负荷预测（Hybrid method for short-term load forecasting based on K-means and convolutional neural network）[J].计算机系统应用（Computer Systems&Applications）,2020,29(3):161-166.
[18]胡全贵，谢可，任玲玲，等（Hu Quangui,Xie Ke,Ren Lingling,et al）.人工智能在电力行业中的应用分析（Research on application of artificial intelligence in power industry）[J].电力信息与通信技术（Electric Power Information and Communication Technology）,2021,19(1):73-80.
[19] Zheng Jian,Xu Cencen,Zhang Ziang,et al. Electric load forecasting in smart grids using long-short-term-memory based recurrent neural network[C]//51st Annual Conference on Information Sciences and Systems. Baltimore,USA,2017.
[20] Sun Xiaorong,Luh P B,Cheung K W,et al. Efficient approach to short-term load forecasting[J]. IEEE Transactions on Power Systems,2016(3):301-307.
[21]杨秀，胡钟毓，田英杰，等（Yang Xiu,Hu Zhongyu,Tian Yingjie,et al）.基于关注指标和深度学习的台区配变重过载预警方法研究（Heavy overload early warning method of distribution transformer based on attention indicators and deep learning）[J].智慧电力（Smart Power）,2021,49(4):66-74.
[22]杨龙，吴红斌，丁明，等（Yang Long,Wu Hongbin,Ding Ming,et al）.新能源电网中考虑特征选择的Bi-LSTM网络短期负荷预测（Short-term load forecasting in renewable energy grid based on bi-directional long short-term memory network considering feature selection）[J].电力系统自动化（Automation of Electric Power Systems）,2021,45(3):166-173.
[23] Bichpuriya Y K,Soman S A,Subramanyam A. Non-parametric probability density forecast of an hourly peak load during a month[C]//Power Systems Computation Conference. Wroclaw,Poland,2014.
[24] Grinsztajn L,Edouard O,Varoquaux G. Why do treebased models still outperform deep learning on tabular data?[EB/OL]. https：//arXiv.org/abs/2207. 08815,2022-07-18.
[25]梁露，刘远龙，刘韶华，等（Liang Lu,Liu Yuanlong,Liu Shaohua,et al）.基于ECA-TCN的电力系统短期负荷预测研究（Research on short-term load forecasting of power system based on ECA-TCN）[J].电力系统及其自动化学报（Proceedings of the CSU-EPSA）,2022,34(11):52-57.
[26]周王峰，李勇，郭钇秀，等（Zhou Wangfeng,Li Yong,Guo Yixiu,et al）.基于Hausdorff形状分类的配电网日最大负荷预测（Forecast of daily maximum load of distribution network based on Hausdorff shape classification）[J].湖南电力（Hunan Electric Power）,2021,41(1):1-5,10.
基本信息:
DOI：10.19635/j.cnki.csu-epsa.001295
中图分类号:TM421
引用信息:
[1]邓威,梅玉杰,李勇,等.基于BOXGBoost的配变日峰值负荷预测及重过载预警方法[J].电力系统及其自动化学报,2024,36(07):49-58.DOI:10.19635/j.cnki.csu-epsa.001295.
基金信息:
国家自然科学基金资助项目（U22B20104）; 国网湖南省电力有限公司科技项目（5216A521001F）