摘要
为了提高宽频带振荡分类的准确率,提出一种基于优化深度置信网络(deep belief networks,DBN)和极限学习机(extreme learning machine,ELM)模型的宽频振荡监测方法。首先,采用混沛粒子群算法优化的DBN对宽频振荡信号进行特征提取,完成对输入数据的降维处理。然后,将特征矩阵输入到天牛须搜索算法优化后的ELM网络,进行振荡分类。最后,采用优化的DBN-ELM模型对系统仿真的宽频振荡信号进行监测和分类。结果表明,该方法对具有时变和非线性特性的宽频振荡具有更高的分类准确率和更短的分类时间。
In order to improve the accuracy of wide-band oscillation classification,a wide-band oscillation monitoring method based on optimized deep belief networks and extreme learning machine(DBN-ELM)model is proposed in this paper.Firstly,the DBN optimized by chaos particle swarm optimization is used to extract the characteristics of wide-band oscillation signals and complete the dimension reduction of the input data.Then,the characteristic matrix is input into the ELM network optimized by beetle antennae search algorithm for oscillation classification.Finally,the optimized DBN-ELM model is established to monitor and classify the wide-band oscillation signals simulated by the system.The results show that the proposed method has higher classification accuracy and shorter classification time for wide-band oscillations with time-varying and nonlinear characteristics.
作者
赵妍
申聪
唐文石
聂永辉
ZHAO Yan;SHEN Cong;TANG Wenshi;NIE Yonghui(Northeast Electric PowerUniversity,Jilin 132012,China)
出处
《吉林电力》
2023年第1期21-25,共5页
Jilin Electric Power
基金
大规模新能源接入下随机互联电力系统小扰动稳定性分析与稳定控制(61973072)。
关键词
宽频振荡监测
特征提取
深度置信网络
混沌粒子群优化算法
极限学习机
monitoring of wide-band oscillation
feature extraction
deep belief networks
chaos particle swarm optimization
extreme learning machine
作者简介
赵妍(1974),女,博士,副教授,研究方向为电力系统稳定分析和控制,人工智能在电力系统中的应用。
