Gear wear often results in both tooth profile changes caused by abrasive wear, and fatigue pitting. Being able to accurately monitor and predict the profile change (i.e., the wear depth in the direction normal to the gear surface) and surface pitting propagation can bring enormous benefits to industrial practice. However, there is a lack of efficient, reliable, and effective tools to do so. To address this, this paper proposes a gear wear monitoring and prediction approach through the integration of: (i) a dynamic model, to simulate the dynamic responses of the gear system; (ii) two tribological models, to estimate wear depth (in the direction normal to the gear surface) and pitting density (on the gear surface); and (iii), model updating, by comparing simulated and measured vibration signals.More specifically, a 21-degree-of-freedom dynamic model is used to simulate a spur gearbox setup and produce simulated vibrations and contact forces between the meshing gear teeth. Using the contact pressure (calculated from the force) as an input, the wear depth and pitting density are then predicted by the tribological models and used to modify the gear geometry profile and contact area in the dynamic model. The developed approach allows the dynamic model and the wear models to communicate so that both the gear tooth profile change and pitting density can be simulated continuously. To guarantee accurate prediction results from the models, novel approaches are developed to update the wear coefficients in the tribological models by comparing simulated and measured vibrations. The paper demonstrates the ability and effectiveness of the proposed vibration-based methodology in monitoring and predicting gear wear, specifically the tooth profile change and surface pitting propagation, using measurements from both a lubricated test, dominated by surface pitting propagation with mild tooth profile change, and a dry test dominated by tooth profile change.
Read full abstract