摘要
针对接地短路故障中识别率较低、抗干扰性较大的问题,该文提出短路故障自动化监测系统。通过集成经验模态对零序电流做出分解,提取相应的本证模态分量,构建特征向量。结合灰狼算法对门控神经网络优化,加入相应的softmax分类优化,构建接地短路故障的自动化监测系统。经过实际算例验证,改进系统能够更稳定地进行故障识别,具有更高的准确率以及抗干扰性能。
Aiming at the problems of low recognition rate and high anti-interference in grounding short-circuit fault,an automatic monitoring system for short-circuit fault is proposed.The zero sequence current is decomposed by integrating empirical modes,and the corresponding eigenmode components are extracted.Based on the grey wolf algorithm to optimize the gated neural network and add the corresponding softmax classification optimization,the automatic monitoring system of ground short circuit fault is constructed.The practical examples show that the improved system can identify faults more stably,with higher accuracy and anti-interference performance.
作者
施晓敏
徐飞
沈磊
SHI Xiaomin;XU Fei;SHEN Lei(Economic and Technological Research Institute of State Grid Anhui Electric Power Co.,Ltd.,Hefei 230022,China)
出处
《自动化与仪表》
2024年第4期83-88,共6页
Automation & Instrumentation
关键词
集成经验模态
神经网络
灰狼算法
故障检测
softmax层
integrated empirical modality
neural networks
grey wolf algorithm
fault detection
softmax layer
