现代电子技术2025，Vol.48Issue(24)：10-18,9.DOI:10.16652/j.issn.1004-373x.2025.24.002
基于ICFOA-VMD和混合神经网络的交流串联电弧故障检测
AC series arc fault detection based on ICFOA-VMD and hybrid neural network
摘要
Abstract
There are various types of loads in low-voltage AC power distribution systems,and environmental interference factors are relatively strong,which makes the detection of series arc faults difficult,thereby seriously affecting electrical safety.In allusion to the problem of insufficient detection capability of AC series arc faults,a method is proposed that uses the improved catfish optimization algorithm(ICFOA)to optimize the parameters of variational mode decomposition(VMD),extracts the characteristic information of intrinsic mode functions(IMFs),and realizes series arc fault detection by means of deep learning model.A low-voltage AC series arc fault experimental platform is constructed and current data are collected under normal and arc fault conditions.The VMD is optimized by means of the ICFOA algorithm to obtain the optimal parameters,and the current signals are decomposed by means of ICFOA-VMD to obtain k IMFs components.Multi-dimensional features are extracted as inputs for the IMFs components to achieve the detection of arc faults by means of CNN-LSTM hybrid neural network model.The experimental results show that the proposed method is applicable to resistive,inductive and capacitive loads,has good universality,the average accuracy of detection for different loads reaches 99.72%,and can realize arc fault detection with high accuracy.关键词
电弧故障检测/变分模态分解/改进捕鱼优化算法/混合神经网络/特征提取/深度学习Key words
arc fault detection/variational modal decomposition/improved fishing optimization algorithm/hybrid neural network/feature extraction/deep learning分类
信息技术与安全科学引用本文复制引用
闻龙,刘松..基于ICFOA-VMD和混合神经网络的交流串联电弧故障检测[J].现代电子技术,2025,48(24):10-18,9.基金项目
国家重点研发计划资助项目(2021YFB4001700) （2021YFB4001700）
