Font Size: 

Reliable lubrication is essential to the life of bearing. An approach to predict the reliability of bearing elastohydrodynamic lubrication (EHL) under a random input torque is proposed in this article. Firstly, a nonlinear stochastic dynamic model that considers clearance and random loads is established to obtain the probability distribution of the contact force between the rollers and raceways. An EHL model is employed to obtain the relationship curve of the contact force and the minimum film thickness at different position of the rotation period. Then, the probability distribution of the minimum film thickness, which is the input variable of the reliability model, can subsequently be obtained by combining the probability distribution of the contact force and the contact force-minimum film thickness curve. Finally, the reliability of the lubrication can be evaluated by the generalized stress-strength interference theory. To validate the method, a roller bearing is analyzed, and the position of high failure probability of full-film lubrication is obtained. The wear condition of the bearing obtained from the failure probability is consistent with observations.