电力信息与通信技术2024，Vol.22Issue(11)：43-51,9.DOI:10.16543/j.2095-641x.electric.power.ict.2024.11.06
基于组合机器学习的配电网区域故障等级预测方法
Regional Fault Level Prediction Method for Distribution Network Based on Combined Machine Learning
摘要
Abstract
To predict the fault risk level of regional distribution network accurately,this paper proposes a method based on combined machine learning models.After preprocessing the original data and reasonably dividing the fault levels,the relevant features of the distribution network faults are summarized and sorted.The random forest algorithm is used to calculate and rank the feature weights to obtain the optimal fault feature set.Subsequently,the CNN-MSVM algorithm is used to predict the fault risk level and optimize the model parameters.Through practical example analysis,it is confirmed that the proposed method has good prediction performance,and the predicted results can provide a basis for risk response in distribution networks with practical promotion and application value.关键词
配电网/随机森林/特征选择/故障预测Key words
distribution network/random forest/feature selection/fault prediction分类
信息技术与安全科学引用本文复制引用
杨晓雨,贾东梨,刘科研,何开元,詹惠瑜..基于组合机器学习的配电网区域故障等级预测方法[J].电力信息与通信技术,2024,22(11):43-51,9.基金项目
国家电网有限公司总部科技项目资助"面向高比例分布式电源接入的配电网数字孪生关键技术及应用"(5400-202255154A-1-1-ZN). （5400-202255154A-1-1-ZN）
