摘要
传统异常数据检测算法在面对海量日同期线损时,检测准确率低且稳定性弱,无法有效的检测出异常问题数据。为提高新能源配电网中日同期数据异常检测的有效性,本基于“源-荷-储”新能源智能配电网系统模型,提一种出将日同期线损差序对比和异常极值判断相结合的DAO日同期线损异常值检测算法。算法首先通过系统建模估算配电网各设备功耗模型,提高同期线损理论计算的准确性;然后通过设置时间尺度,完成IMS同期线损数据的预处理,提升模型构建的时效性;接着以日同期线损理论值与测量值的差序数据为基础,构建AVP差值评估模型,完成序列数据异常判断;最后通过设置序列区域半径与邻异常极点阈值,完成异常数据位置判断。数据异常检测仿真结果显示,较其它五类基线算法相比,DAO算法在测试集中的P、R和F1参数分别平均提高6.24%、7.35%和6.83%,表明DAO算法的检测准确率更高,稳定性更强。综上所述,DAO日同期线损异常数据检测算法通过异常判断与位置极值判断两步走,有效的提高了算法检测的精确性与鲁棒性,在计算机仿真与配电网运维领域中具有重要的研究意义。
The traditional abnormal data detection algorithm has low detection accuracy and weak stability in the face of massive line loss in the same period of the day,and can not effectively detect abnormal data.In order to improve the effectiveness of anomaly detection in new energy distribution network,based on the"source-load-storage"new energy smart distribution network system model,this paper proposes a DAO line loss anomaly detection algorithm in the same period,which combines the comparison of line loss difference sequence in the same period and the judgment of abnormal extreme value.Firstly,the algorithm estimates the power consumption model of each equipment in the distribution network through system modeling to improve the accuracy of theoretical calculation of line loss in the same period,and then completes the preprocessing of IMS line loss data in the same period by setting the time scale to improve the timeliness of model construction.Then,based on the difference sequence data between the theoretical value and the measured value of the line loss in the same period of a day,an AVP difference value evaluation model is constructed to complete the abnormal judgment of the sequence data.Finally,the abnormal data-position judgment is completed by setting a sequence region radius and an adjacent abnormal pole threshold.The simulation results of the data anomaly detection experiment show that compared with the other five kinds of baseline algorithms,the average P,R and F,parameters of the DA0 algorithm in the test set are increased by 6.24%,7.35%and 6.83%,respectively,which indicates that DAO algorithm has higher detection accuracy and stronger stability.To sum up,the DAO line loss anomaly data detection algorithm effectively improves the accuracy and robustness of the algorithm detection through two steps of anomaly judgment and location extremum judgment,which has important research significance in the field of computer simulation and distribution network operation and maintenance.
作者
廖耀华
王恩
李波
王毅
LIAO Yao-hua;WANG En;LI Bo;WANG Yi(Electric Power Science Research Institute of Yunnan Power Grid Co.,Ltd.,Kunming Yunnan 650217,China;Key Laboratory of Green Energy and Digital Power Measurement and Control in Yunnan Province,Kunming Yunnan 650217,China;School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《计算机仿真》
2024年第10期95-100,共6页
Computer Simulation
基金
中国南方电网有限责任公司科技项目(YNKJXM20220166)。
关键词
日同期线损
异常数据检测
差值评估
Contemporaneous line loss
Abnormal data detection
Difference evaluation
作者简介
廖耀华(1990-4),男(汉族),湖南衡阳人,硕士,工程师,研究方向:高电压计量、智能量测、电力计量故障检测与处理等;王恩(1973-12),男(汉族),云南昆明人,高级工程师,研究方向:热工计量;李波(1982-7),男(汉族),四川内江人,硕士,教授级高级工程师,研究方向:电力计量、智能量测;王毅(1981-11),男(汉族),重庆人,博士,副教授,研究方向:电力物联网、智能量测。
