DAMAGE MECHANICS BASED BEARING PROGNOSIS USING HUMS CONDITION INDICATORS

摘要

In this paper, our experience in developing damage mechanics based bearing prognostic capability using health and usage monitoring systems (HUMS) condition indicators (CIs) is reported. We consider the bearing system as a single-degree-of-freedom vibratory system. In such a system, the natural frequency and the acceleration amplitude at the natural frequency can be related to the system stiffness. Therefore, the natural frequency and the CIs at the natural frequency can also be related to the system stiffness. As the relationship between failure time and stiffness variation can be established from damage mechanics, a damage bearing life model can be established using HUMS CIs. A diversity-guided convergent particle swarm optimization (DCPSO), algorithm is used to determine parameters of the bearing life model.