摘要
光伏电站故障频发,影响发电效率。而相对于正常运行数据,电站故障数据较少,导致故障检测精度不高。针对这个问题,提出了一种基于AFCM(alter-native fuzzy C-means)-SMOTE(synthetic minority over-sampling technique)算法与随机森林算法相结合的光伏电站故障诊断方法。用AFCM-SMOTE算法对故障样本进行处理,生成“人造”样本,用“人造”样本训练随机森林算法,最终实现对光伏电站故障的检测。实验结果表明,AFCM-SMOTE算法很好地解决了随机森林在光伏故障检测应用中因为故障样本数据少导致分类不精确的问题,提高了故障诊断的准确性。
Photovoltaic power plant failures frequently occur,affecting power generation efficiency.Compared with the normal operation data,there is less fault data in the power station,resulting in low fault detection accuracy.To solve this problem,a photovoltaic power plant fault diagnosis method based on the combination of AFCM(alternative fuzzy C-means)-SMOTE(synthetic minority over-sampling technique)algorithm and random forest algorithm is proposed.The AFCM-SMOTE algorithm is used to process the fault samples to generate“artificial”samples,and use the“artificial”samples to train the random forest algorithm,and finally realize the detection of photovoltaic power station faults.The experimental results show that the AFCM-SMOTE algorithm solves the problem of inaccurate classification of random deep forest in photovoltaic fault detection applications which because of the lack of fault sample data,and improves the accuracy of fault diagnosis.
作者
张治
马辉
王林
ZHANG Zhi;MAHui;WANG Lin(State Power Investment Group Photovoltaic Industry Innovation Center,Xining Qinghai 810000,China;School of Automation and Information Engineering,Xi’an University of Technology,Xi’an Shannxi 710048,China)
出处
《电源技术》
CAS
北大核心
2021年第11期1495-1499,共5页
Chinese Journal of Power Sources
基金
陕西省科技计划重点项目资助(2017ZDCXLGY-05-03)。
关键词
光伏电站
故障检测
AFCM
改进随机森林算法
photovoltaic power plant
fault detection
AFCM-SMOTE
improved random forest algorithm