基于特征融合与LOF的滚动轴承性能退化早期故障预警技术

Early Fault Warning Technology for Rolling Bearing Performance Degradation Based on Feature Fusion and LOF

针对传统滚动轴承性能退化故障预警技术未能考虑转子故障干扰的问题,提出了一种基于高通滤波的轴承健康状态评估技术。首先,根据滚动轴承性能退化机理,在传统经典特征以及高通滤波频域特征的基础上,建立了基于特征融合与局部离群因子(LOF)的滚动轴承早期故障预警模型;然后,通过实验验证了所提方法。结果表明该方法在排除转子故障干扰的同时,能够有效识别与追踪轴承性能退化过程,且识别的早期故障发生时间比传统振动加速度有效值突变点提前1700 min。

Aiming at the problem that traditional rolling bearing performance degradation fault warning technology fails to consider rotor fault interference,a bearing health status assessment technology based on high-pass filter was proposed.Firstly,according to the performance degradation mechanism of rolling bearings,a rolling bearing early fault warning model based on feature fusion and local outlier factor(LOF)was established on the basis of traditional classical characteristics and high-pass filter frequency domain characteristics.Then,the proposed method was verified by experiments.The results showed that this method can effectively identify and track the bearing performance degradation process while eliminating rotor fault interference,and the identified early fault occurrence time was 1700 min earlier than the traditional effective mutation point of vibration acceleration.