基于麻雀搜寻优化算法的代理购电用户用电量多维度协同校核  

Multi-dimensional Collaborative Checking of Power Consumption of Purchasing Agent Users Based on ISSA Optimization Algorithm

作　　者：周颖 乔婧 陈宋宋 赵伟博 丁一 武亚杰 田宇 ZHOU Ying;QIAO Jing;CHEN Songsong;ZHAO Weibo;DING Yi;WU Yajie;TIAN Yu(China Electric Power Research Institute,Haidian District,Beijing 100192,China;School of Economics and Management,North China Electric Power University,Changping District,Beijing 102206,China;State Grid Tianjin Electric Power Research Institute,Hexi District,Tianjin 300384,China)

机构地区：[1]中国电力科学研究院有限公司,北京市海淀区100192 [2]华北电力大学经济与管理学院,北京市昌平区102206 [3]国网天津市电力公司电力科学研究院,天津市河西区300384

出　　处：《电网技术》2025年第2期604-612,I0064-I0067,共13页Power System Technology

基　　金：国家电网公司总部科技项目“支撑代理购电新业务的用户分群电力电量精准预测关键技术研究与应用”(5400-202312226A-1-1-ZN)。

摘　　要：随着代理购电业务稳步推进,用电量预测在智能电网运行中发挥着至关重要的作用。现阶段研究大多侧重于通过算法来提高预测结果的准确度和可靠性,而这些方法缺乏对电力系统多维因素的全面考量和精确校核。因此,多维度且全面地对代理购电用户用电量进行预测是代理购电业务中面临的问题之一。对此,该文提出了计及多维度协同的用户用电量预测结果校核方法。首先,该文采用了偏差概率分布模型分析各个维度(区域、行业、电压等级)的有效偏差分布,进行各维度有效偏差识别;其次,以误差最小为目标采用改进麻雀搜索算法(improved sparrow search algorithm,ISSA)优化算法进行多维度权重优化配比,构建预测值和权重值组合加权的多维度协同校核模型;最后选取误差指标对多维度校核后的预测值进行误差指标评估。结合某省的代理购电用户用电量对上述算法进行了验证,结果表明,基于ISSA优化算法的多维度协同校核方法在平均绝对误差指标下较行业维度、区域维度及电压等级维度分别降低了51.9%、23.4%和19.1%,均方根误差指标下较行业维度、区域维度及电压等级维度分别降低了40.0%、15.0%和8.6%,具有良好的泛化性。With the steady progress of the proxy electricity purchase business,electricity consumption prediction plays a crucial role in the operation of smart grids.Currently,most research focuses on improving the accuracy and reliability of prediction results through algorithms.Still,these methods need more comprehensive consideration and accurate verification of multidimensional factors in the power system.Therefore,predicting the electricity consumption of proxy electricity purchasing users from multiple dimensions comprehensively is one of the problems faced in the proxy electricity purchasing business.This article proposes a verification method for user electricity consumption prediction results considering multi-dimensional collaboration.Firstly,this article adopts a deviation probability distribution model to analyze the effective deviation distribution of each dimension(region,industry,voltage level)and identify the effective deviations of each dimension.Secondly,to minimize error,the ISSA optimization algorithm is used to optimize the ratio of multidimensional weights,and a multifaceted collaborative verification model is constructed by combining predicted values and weight values for weighting.Finally,error indicators are selected to evaluate the expected values after multi-dimensional verification.This article validates the above algorithm by combining the electricity consumption of proxy electricity-purchasing users in a particular province.The results show that the multi-dimensional collaborative verification method based on the ISSA optimization algorithm reduces the average absolute error index by 51.9%,23.4%,and 19.1%compared to the industry,regional,and voltage level dimensions,respectively.The root mean square error index showed a reduction of 40.0%,15.0%,and 8.6%compared to the industry,regional,and voltage level dimensions,respectively,indicating good generalization ability.

关 键 词：代理购电 误差校核 ISSA优化算法 组合权重 均方根误差 

分 类 号：TM721[电气工程—电力系统及自动化]