基于GWO_SVM的非侵入式负荷识别研究  

Research on Non-Invasive Load Recognition Based on GWO_SVM

作　　者：肖月 杨超[1] 王飘 

机构地区：[1]贵州大学电气工程学院,贵州 贵阳 [2]多彩贵州航空有限公司,贵州 贵阳

出　　处：《建模与仿真》2024年第1期932-940,共9页Modeling and Simulation

摘　　要：随着全球能源需求的持续增长和资源的日益紧张,非侵入式负荷监测(Non-Intrusive Load Monitor-ing, NILM)技术在实现资源节约和能源升级中扮演着至关重要的角色。本文针对NILM研究中存在的负荷特征较单一以及负荷识别准确率不高的问题,通过将有功功率、无功功率与电流五次谐波引入作为识别特征,提出了基于灰狼优化器算法(grey wolf optimizer, GWO)优化支持向量机(support vector machine, SVM)的模型,经过在公开数据集REDD上进行实验验证,该方法在负荷识别上具有98.96%的准确率,通过与不同算法在同一数据集上进行负荷识别的准确率进行对比,验证了该文算法在在准确率上有明显提升,证明了该文算法对于提升负荷识别的准确率具有优越性。With the continuous growth of global energy demand and the increasing scarcity of resources, Non-intrusive load monitoring (NILM) technology plays a crucial role in achieving resource conser-vation and energy upgrading. This article addresses the issues of single load characteristics and low accuracy in load recognition in NILM research. By introducing active power, reactive power, and fifth harmonic current as recognition features, a grey wolf optimizer (GWO) based model for opti-mizing support vector machine (SVM) is proposed. The model is validated through experiments on the public dataset REDD. This method has an accuracy of 98.96% in load identification. By compar-ing the accuracy of load identification with different algorithms on the same dataset, it was verified that the algorithm proposed in this paper has a significant improvement in accuracy, demonstrat-ing its superiority in improving the accuracy of load identification.

关 键 词：非侵入式负荷监测 灰狼优化器 支持向量机 负荷监测 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]