摘要
为提升区域性大负荷场景下的负荷预测精度,同时满足小型区域性场景短期配电网的运维保护需求,设计一种融合指数平滑方法和梯度升压的短期负荷预测算法。该算法采用指数平滑方法对历史负荷数据进行预处理,减少了负荷随机波动的影响;进而构建梯度提升机制,利用梯度升压算法对预处理后的数据进行特征学习,增强了对非线性关系和高维数据的处理能力。同时,该算法引入了各类控制因素,实现了对短期配电网负荷的精准预测。采集某高校的真实用电数据作为样本数据集,进行短期预测数值实验,并与同类负荷预测算法进行横向对比。结果表明,所提算法的负荷预测精度为99.1%,预测准确率可达99.3%,有效提升了预测的准确性和可靠性,能够为区域内配电网的平稳运行提供有力的数据支持。
In order to improve the accuracy of load forecasting in regional high load scenarios,while meeting the operational maintenance and protection needs of short-term distribution networks in small regional scenarios,and achieve operation and maintenance protection of short-term distribution networks in small regional scenarios,a short-term load forecasting algorithm that integrates exponential smoothing method and gradient boosting is designed.In this algorithm,the exponential smoothing is used to preprocess historical load data to reduce the impact of random load fluctuations,and a gradient boosting mechanism is established.The gradient boosting algorithm is used to learn features from the preprocessed data,enhancing its ability to handle nonlinear relationships and high-dimensional data.At the same time,various control factors are introduced to achieve accurate prediction of short-term distribution network loads.The real electricity consumption data from a certain university is collected as a sample dataset for short-term forecasting numerical experiments,and compared horizontally with similar load forecasting algorithms.The results show that the proposed algorithm has a load forecasting accuracy of 99.1%and a prediction accuracy of up to 99.3%,effectively improving the accuracy and reliability of the prediction,and providing data support for the smooth operation of the distribution network in the region.
作者
王哲
王成福
WANG Zhe;WANG Chengfu(Shandong University,Jinan 250100,China)
出处
《现代电子技术》
北大核心
2026年第4期135-140,共6页
Modern Electronics Technique
基金
山东省重点研发计划(重大科技创新工程)项目(2019JZZY010903)。
关键词
短期负荷预测
指数平滑方法
梯度升压算法
区域性配电网
负荷预测精度
控制因素
short term load forecasting
exponential smoothing method
gradient boosting algorithm
regional distribution network
load prediction accuracy
control factor
