基于随机森林算法的电压暂降特征量预测研究  

Predicting Characteristic Quantities of Voltage Sag Based on Random Forest Algorithm

作　　者：何觅 杨发宇 苟源芳 蒋羽鹏 HE Mi;YANG Fayu;GOU Yuanfang;JIANG Yupeng(Kunming Power Supply Bureau of Yunnan Power Grid Co.,Ltd.,Kunming 650500,China;School of Mechanical and Electrical Engineering,Kunming University of Technology,Kunming 650500,China)

机构地区：[1]云南电网有限责任公司昆明供电局,云南昆明650500 [2]昆明理工大学机电工程学院,云南昆明650500

出　　处：《电工技术》2023年第20期82-85,共4页Electric Engineering

基　　金：云南电网公司科技项目(编号YNKJXM20220091)。

摘　　要：电压暂降是发生概率较高、对电力用户影响较严重的一类电能质量问题,准确地对敏感用户开展特征量指标评估十分重要。基于此,提出了一种基于随机森林算法的用户侧暂降类型的特征量评估方法,首先根据电网侧和用户侧的暂降数据建立随机森林回归预测模型,然后对模型进行参数设置优化,最后对用户侧电压暂降特征量数据进行预测。利用云南某地区配电网实测数据进行的与多项式预测算法的对比分析表明了随机森林算法用于电压暂降特征量预测的优越性。Voltage sag is a kind of power quality problem with high probability of occurrence and serious impact on power users.It is very important to accurately evaluate indices of characteristic quantities of voltage sag with respect to voltage-sensitive users.This paper proposes a method which is based on random forest algorithm and which is used to predict characteristic quantities reflecting type of user-side sag.According to grid-side and user-side sag data,the random forest regression prediction model is first established,of which the parameters are then optimized,and through which the characteristic quantities of user-side voltage sag are finally predicted.The analysis in comparison with polynomial prediction algorithm using the measured data of distribution network in a certain region in Yunnan Province confirms the superiority of the proposed random forest algorithm.

关 键 词：随机森林 电压暂降 特征量 预测 

分 类 号：TM74[电气工程—电力系统及自动化]