摘要
为解决断路器储能系统的运行状态在线监测问题,基于储能子机构动作时的振动信号频率特性,提出基于振动信号特征模态与CAMMC-LSTM模型的预测方法,实现断路器储能系统的剩余寿命预测。采用研发的断路器机械特性实验平台获取退化样本数据,通过VMD获得振动信号有效模态与电流信号一起构成特征模态。依据通道注意力机制和多尺度卷积原理提出CAMMC结构,结合长短时记忆网络构建CAMMC-LSTM模型。试验结果表明,特征模态增强模型输入对运行状态的表征能力,通道注意力机制与多尺度卷积改进传统长短时记忆网络对特征的提取性能,可提高剩余寿命定量预测的精度。
In order to solve the problem of on-line monitoring of circuit breaker energy storage system,a prediction method based on vibration signal characteristic mode and CAMMC-LSTM model was proposed to predict the remaining life of circuit breaker energy storage system.The degraded sample data is obtained by using the developed circuit breaker mechanical characteristic experimental platform,and the effective mode of vibration signal and current signal together constitute the characteristic mode through VMD.Based on the channel attention mechanism and the multi-scale convolution principle,the CAMMC structure is proposed,and the CAMMC-LSTM model is constructed by combining the short-duration memory network.The experimental results show that the characteristic mode enhances the representation ability of model input to the running state,channel attention mechanism and multi-scale convolution improve the feature extraction performance of traditional long and short time memory network,and can improve the accuracy of quantitative prediction of residual life.
作者
孙曙光
纪卫震
陈静
黄光临
王景芹
SUN Shuguang;JI Weizhen;CHEN Jing;HUANG Guanglin;WANG Jingqin(School of Artificial Intelligence,Hebei University of Technology,Tianjin 300130,China;Wenzhou Juxing Technology Co.,Ltd.,Wenzhou 325062,China;State Key Laboratory of Reliability and Intelligence of Electrical Equipment,Hebei University of Technology,Tianjin 300130,China)
出处
《中国测试》
CAS
北大核心
2024年第10期12-22,共11页
China Measurement & Test
基金
河北省自然科学基金项目(E2021202136)。
关键词
断路器
振动信号
深度学习
寿命预测
circuit breaker
vibration signal
deep learning
life prediction
