一种滚动轴承振动信号自适应数据级融合方法  

An adaptive data-level fusion method for rolling bearing vibration signals

作　　者：郭俊锋[1] 樊怡明 Guo Junfeng;Fan Yiming(School of Mechanical and Electrical Engineering,Lanzhou University of Technology,Gansu Lanzhou,730050,China)

机构地区：[1]兰州理工大学机电工程学院,甘肃兰州730050

出　　处：《机械设计与制造工程》2022年第12期98-103,共6页Machine Design and Manufacturing Engineering

摘　　要：传感器信号的充分利用对于设备和零件的状态监测具有重大意义。为了通过采集的多源振动信号得到设备和零件的完备退化信息,提出一种自适应加权数据级融合方法。首先对振动信号进行预处理,然后以K最近邻算法的分类结果作为粒子群优化算法的适应度函数,通过不断迭代,寻找多源传感器融合的最佳权重。对多源传感器融合系统、多源传感器融合方法以及滚动轴承的故障诊断进行了研究,最后在滚动轴承的全寿命周期数据集上进行试验验证,证明该方法实现了多源传感器采集数据的有效利用,能够完备反映滚动轴承的故障特征,对振动信号的故障诊断和寿命预测具有长远意义。The full use of sensor signals is of great significance for the condition monitoring of equipment and parts.In order to obtain complete degradation information of equipment and parts from the collected multi-source vibration signals,an adaptive weighted data level fusion method is proposed.Firstly,the vibration signal is preprocessed,and then the classification result of K nearest neighbor algorithm is taken as the fitness function of particle swarm optimization algorithm.Through continuous iteration,the best weight of multi-source sensor fusion is found.The multi-source sensor fusion system,the multi-source sensor fusion method and the fault diagnosis of rolling bearings are studied.Finally,the experimental verification is carried out on the data set of the whole life cycle of rolling bearings.It is proved that the method realizes the effective utilization of the data collected by multi-source sensors,can fully reflect the fault characteristics of rolling bearings,and has long-term significance for the fault diagnosis and life prediction of vibration signals.

关 键 词：多源传感器 粒子群优化算法 K最近邻 数据级融合 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置] TN911.7[自动化与计算机技术—控制科学与工程]