基于优化RBF神经网络在电能质量扰动分类中的应用
Application of Optimized RBF Neural Network in Power Quality Disturbance Classification
姚宇 1方忠强 1张坤 1胡慧江 1刘宏伟2
作者信息
- 1. 华设设计集团股份有限公司,南京 江苏 210014
- 2. 长安大学能源与电气工程学院,西安 陕西 710018
- 折叠
摘要
为保证建筑中智能电子产品安全运行,解决实际工程中电能质量扰动识别分类准确率低、抗噪性差等问题,提出一种基于优化RBF神经网络识别电能质量扰动的方法.首先,将20 种电能质量扰动信号通过S 变换进行时频域分析,提取出的扰动时频域特征数据划分为测试集和训练集;然后,构建径向基函数(RBF)神经网络电能质量扰动分类模型;其次,引入蜣螂优化算法(DBO)对RBF神经网络参数进行参数优化;最后,将划分好的训练集和测试集输入到优化后的神经网络中进行扰动分类.仿真及工程实验表明,提出的方法对于电能质量扰动识别准确率高,抗噪性及泛化能力强.
Abstract
To ensure the safe operation of intelligent electronic products in buildings,a method based on an optimized RBF neural network for identifying power quality disturbances is proposed,which addresses issues such as low accuracy in power quality disturbance classification and poor noise resistance in practical engineering.Firstly,20 types of power quality disturbance signals were subjected to time-frequency domain analysis through S-transform,and the extracted disturbance time-frequency domain feature data was divided into a test set and a training set;Then,a radial basis function(RBF)neural network power quality disturbance classification model;Secondly,the Dung Beetle Optimizer(DBO)algorithm is introduced to optimize the parameters of the RBF neural network;Finally,input the divided training and testing sets into the optimized neural network for disturbance classification.Simulation and engineering experiments have shown that the proposed method has high accuracy in identifying power quality disturbances,strong noise resistance,and generalization ability.
关键词
电能质量/扰动分类/径向基函数/神经网络/蜣螂优化算法/S变换/特征提取Key words
electricity quality/disturbance classification/RBF neural network/DBO/S transformation/feature extraction引用本文复制引用
出版年
2024
NETL
NSTL
万方数据