摘要
随着新能源电动汽车的不断普及,充电桩的使用需求越来越高。但在充电桩日常运维时,往往由于硬件设计以及技术手段无法满足实际的运维需求,导致充电桩故障检测不够精确与及时的问题。针对这一问题,提出一种以随机森林算法为基础依据,对充电桩故障进行自动化检测的系统,以保证实时检测充电桩的运行数据信息,及时处理充电桩故障。首先介绍系统硬件的整体构造,而后结合随机森林算法搭建充电桩充电故障自动化检测模型,经过实验验证,检测系统进行充电桩故障检测准确率高达97%。
With the continuous popularization of new energy electric vehicles,the demand for charging piles is getting higher and higher.However,during the daily operation and maintenance of charging piles,the hardware design and technical means often cannot meet the actual operation and maintenance needs,resulting in the problem that the fault detection of charging piles is not accurate and timely.In order to solve this problem,a system based on random forest algorithm was proposed to automatically detect charging pile faults,so as to ensure real-time detection of charging pile operation data information and timely processing of charging pile faults.Firstly,the overall structure of the system hardware is introduced,and then the automatic detection model of charging pile charging fault is built by combining the random forest algorithm,and the accuracy of the detection system for charging pile fault detection is as high as 97% after experimental verification.
作者
郭静
郭雅娟
姜海涛
周超
王梓莹
GUO Jing;GUO Yajuan;JIANG Haitao;ZHOU Chao;WANG Ziying(Electric Power Research Institute,State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 211100,China)
出处
《自动化与仪器仪表》
2023年第12期252-256,共5页
Automation & Instrumentation
基金
国网江苏省电力有限公司科技项目(J2022027)新型电力系统中第三方接入与交互安全防护关键技术研究。
关键词
随机森林算法
基尼系数
感受性曲线
充电桩监测装置
模型优化
深度学习
random forest algorithm
gini coefficient
susceptibility curve
charging pile monitoring device
model optimization
deep learning
