核电设备冷却水泵驱动端滚动轴承健康状态评价方法  

Research on bearing health state evaluation of component cooling pump based on channel attention model

作　　者：夏莹沛 董文辉 胡俊锋 段腾飞 陈林 吴琪文 Xia Yingpei;Dong Wenhui;Hu Junfeng;Duan Tengfei;Chen Lin;Wu Qiwen(Shandong Nuclear Power Co.,Ltd.,Shandong Haiyang,265199,China;Hangzhou AIMS Intelligent Technology Co.,Ltd.,Zhejiang Hangzhou,310051,China)

机构地区：[1]山东核电有限公司,山东海阳265199 [2]杭州安脉盛智能技术有限公司,浙江杭州310051

出　　处：《机械设计与制造工程》2025年第3期106-110,共5页Machine Design and Manufacturing Engineering

摘　　要：核电设备冷却水系统的稳定运行是核电稳定供应的基础,由于设备冷却水泵运行工况复杂多变,其驱动端滚动轴承常发生故障且劣化过程具有强非线性而难以预测。针对核电设备冷却水泵驱动端滚动轴承早期劣化过程中振动信号故障特征微弱而难以提取的问题,提出一种基于通道注意力机制的滚动轴承健康状态评估和故障诊断方法。通过自适应提取滚动轴承早期失效的微弱故障特征,在保证诊断效果的同时降低了前期信号处理的复杂性,实验结果表明,所提方法在不同型号的滚动轴承上也具有较好的泛化性。The component cooling water system of nuclear reactors is the basis for stable supply of electricity.Due to complex and changeable operating conditions of the component cooling pump,the bearings often have failures and its deterioration process has strong nonlinearity and unpredictability.Aiming at the problem that the fault characteristics of vibration signal in the early deterioration of component cooling pump bearings are weak and difficult to extract,this paper proposes a bearing health assessment method based on adaptive channel attention model.It can adaptively extract the weak characteristics of rolling bearings during early failure,which lowers the complexity of early signal processing while ensuring the diagnostic effect.Experimental results show that the proposed model has a good generalization ability on different types of rolling bearings.

关 键 词：设备冷却水泵 振动信号 通道注意力机制 故障诊断 

分 类 号：TH133.3[机械工程—机械制造及自动化]