加权数据级融合在滚动轴承故障诊断的应用

Application of Weighted Data Level Fusion in Rolling Bearing Fault Diagnosis

在滚动轴承故障诊断中,为了使采集的振动信号有效利用,提高诊断的准确率,采用了一种根据相关函数确定权重来进行数据级融合的方法.该方法可以在无法确定传感器任何先验知识的情况下,可依据各传感器所测数据相关程度的改变,从而可以适时地调整参与融合的各传感器权值.仿真试验证明,该方法动态适应性好,抗干扰能力强,融合结果在精度、容错性方面均优于传统的平均值加权法,真实测得的数据经过试验验证后确定该方法在实际操作中切实可行,结果准确性高.

In order to improve the accuracy of diagnosis, this paper adopts a method of data-level fusion based on the correlation function to determine the weight during the rolling bearing fault diagnosis. Without knowing any prior knowledge of the sensor, the method can be adjusted according to the degree of correlation of the measured data of each sensor so that the sensor weights involved in the fusion can be adjusted in time. The experimental results show that this method has good dynamic adaptability and strong anti-interference ability, and the fusion result of this method is superior to the traditional average weighting method in terms of accuracy and fault tolerance, which is practicable in practical operation.