摘要
文章针对工业电力系统负荷多变场景,基于电力负荷数据采集与分析,提出了电力负荷预测中的敏感特征识别方法,为电力负荷短期预测提供特征选择支持。首先,文章回顾了电力负荷预测对电网运行与规划的关键作用及预测技术的发展历程。随后,文章概述斯皮尔曼与皮尔逊相关系数的理论基础,在此基础上,通过特征选择与数据处理,构建了实证分析框架,基于两种相关系数实现敏感特征的量化评价与对比分析。研究发现,在电力负荷预测中,选取不同的特征或者采用不同的相关性评价方法对预测结果有着不同的影响,且斯皮尔曼和皮尔逊相关系数在特征识别中表现出各自的优势。本研究成果对提高负荷预测精度具有重要意义,为后续研究提供了新的视角和方法参考。
Focusing on industrial power systems with fluctuating loads,this paper proposes a method to identify sensitive features in power load forecasting using collected data,thus offering feature support for short-term forecasting.The critical role of power load forecasting in grid operation and planning as well as the development of forecasting technologies are firstly reviewed.After that,the fundamental basics of Spearman and Pearson correlation coefficients is summarized.An empirical framework analysis is then constructed through feature selection and data processing that can achieve quantitative evaluation and comparative analysis based on two types of correlation coefficients’sensitive features.It is shown that selecting different features or correlation evaluation methods can lead to distinguishing impacts on the forecasting results.Also,both Spearman and Pearson correlation coefficients have their respective advantages in the field of power load application.Therefore,the results of this paper hold significant implications for improving load forecasting accuracy,thus providing novel perspectives and methodological references for subsequent studies.
作者
王义洋
王广威
李嘉楠
柳鑫恩
孙振龙
WANG Yiyang;WANG Guangwei;LI Jianan;LIU Xinen;SUN Zhenlong(School of Electrical and Automation Engineering,Liaoning Institute of Science and Technology,Benxi Liaoning 117004,China)
出处
《辽宁科技学院学报》
2025年第3期22-26,共5页
Journal of Liaoning Institute of Science and Technology
基金
辽宁省教育厅2023年面上项目“基于数据驱动的电力负荷预测关键技术研究”(JYTMS20231798).
关键词
电力负荷预测
特征识别
斯皮尔曼相关系数
皮尔逊相关系数
敏感特征
Power load forecasting
Feature recognition
Spearman correlation coefficient
Pearson correlation coefficient
Sensitive features
