基于多特征融合和深度信念网络的断路器元件故障识别研究
Research on fault identification of circuit breaker components based on multi feature fusion and deep belief network
孔维林 1曾冀 1刘雪涛 1张亮 1田万玺1
作者信息
- 1. 天津天传电控设备检测有限公司,天津 300300;天津电气科学研究院有限公司,天津 300180
- 折叠
摘要
为了标准化断路器元件故障识别与诊断流程,提高断路器运行的稳定性,引进多特征融合技术和深度信念网络,研究设计断路器元件的故障识别方法.同步布置传感器,建立传感器与终端计算机之间的通道,进行断路器元件运行信号采集与多特征深度融合;为排除其他因素造成的振动信号,实现对断路器元件故障的准确识别,引进深度信念网络,进行断路器元件振动信号的重构;将重构后的微调整信号作为核心,通过对其时域与频域的分析,进行断路器元件故障的识别与分类.对比实验结果证明:设计的方法可以实现对断路器元件故障的诊断与故障类型的精准识别,应用效果良好.
Abstract
In order to standardize the fault identification and diagnosis of circuit breaker components,improve the stability of circuit breaker operation,multiple feature fusion technology and deep belief network are introduced to carry out the design and research of circuit breaker component fault identification methods.Simultaneously arrange sensors,establish channels between sensors and terminal computers,collect operating signals of circuit breaker components and deeply fuse multiple features;To eliminate vibration signals caused by other factors and achieve accurate identification of circuit breaker component faults,a deep belief network is introduced to reconstruct the vibration signals of circuit breaker components;Using the reconstructed micro adjustment signal as the core,identify and classify circuit breaker component faults through analysis of its time and frequency domains.The comparative experimental results show that the designed method can achieve accurate diagnosis of circuit breaker component faults and identification of fault types,and the application effect is good.
关键词
多特征融合/信号重构/故障识别/断路器/深度信念网络Key words
multi feature fusion/signal reconstruction/fault identification/circuit breaker/deep belief network引用本文复制引用
出版年
2025
NETL
NSTL
万方数据