基于MVMD-CapSA-DBN的工业多元负荷分类研究
- 分享到:
摘要：
针对多元电力负荷数据时间序列非平稳性、时序相关性和非线性等特性,为掌握电力负荷的变化规律和发展趋势,实现对电力负荷的科学管理,将工业多元负荷有功功率作为原始数据,提出MVMD-CapSA-DBN负荷分类模型.首先,利用改进停止准则的变分模态分解(Variational Mode Decomposition,VMD)将数据分解,得到较为平稳的多个数据分量;之后,提取各分量能量值作为特征;最后,将0~1标准化的数据作为特征向量,输入经卷尾猴搜索算法(Capuchin Search Algorithm,CapSA)优化参数后的深度置信网络(Deep Belief Nets,DBN)信号分类.实验证明,可实现对工业多元负荷数据的有效分类,整体准确率在88.89%左右,部分负荷分类准确率可达100%.
Aiming at the non-stationarity,time series correlation and nonlinearity of multiple power load data time series,in order to grasp the change law and development trend of power load and realize the scientific management of power load,this paper takes the active power of industrial multiple load as the original data.The MVMD-CapSA-DBN load classification model is proposed.First,the data is decomposed by Variational Mode Decomposition(VMD)with improved stopping criterion to obtain multiple data components that are relatively stable;then,the energy value of each component is extracted as a feature;Finally,the 0-1 normalized data is used as a feature vector and input into the Deep Belief Nets(DBN)signal classification after parameters are optimized by the Capuchin Search Algorithm(CapSA).Experiments show that the model in this paper can achieve effective classification of industrial multi-load data,the overall accuracy rate is about 88.89%,and the partial load classification accuracy rate can reach 100%.
作者：
周孟然 张易平 汪胜和 马金辉 高博 胡锋 朱梓伟 汪锟 刘宇
Zhou Mengran;Zhang Yiping;Wang Shenghe;Ma Jinhui;Gao Bo;Hu Feng;Zhu Ziwei;Wang Kun;Liu Yu(School of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan 232001,China;State Grid Anhui Electric Power Co.,Ltd.,Hefei 230001,China)
机构地区：
安徽理工大学电气与信息工程学院 国网安徽省电力有限公司
出处：
《betway官方app 学报:自然科学版》 CAS 北大核心 2023年第3期123-130,共8页
Journal of Henan Normal University(Natural Science Edition)
基金国家重点研发计划(2018YFC0604503) 安徽省自然科学基金能源互联网联合基金重点项目(2008085UD06) 安徽省科技重大专项(201903a07020013) 安徽高校自然科学研究重点项目(KJ2021A0470).
关键词：
电力负荷 负荷分类 变分模态分解 深度信念网络 卷尾猴搜索算法
power load load classification variational model decomposition Deep Belief Network Capuchin Search Algorithm
分类号：
TM714 [电气工程—电力系统及自动化]
