ML-Metaheuristic Framework for Texture Optimization in Hybrid Thrust Bearings

Hydrostatic thrust bearings are essential machine elements for supporting heavy axial loads in a wide range of industrial applications. This study investigates a surface texturing approach to improve the tribological behaviour of these bearings, with the aim of reducing friction-induced power losses of the hybrid thrust bearing. The governing Reynolds equation for the hybrid thrust bearing configuration is solved numerically using the Finite Element Method (FEM), enabling accurate modelling of the pressure distribution across the pad surface. Surface texturing is introduced on the thrust pad in the form of micro-radial grooves, a pattern chosen for its potential to modify lubricant flow behaviour and enhance film formation. Since textured surfaces are prone to cavitation, a mass-conservation-based algorithm is integrated into the numerical framework to accurately capture rupture and reformation of the lubricant film within the grooved regions. Comparative simulations show that the textured thrust pad, combined with the hydrodynamic lubricant film, yields a marked reduction in friction power loss relative to a conventional, non-textured bearing design. To extend this analysis and identify optimal texturing configurations, a decision tree-based machine learning model is applied to predict key performance parameters. The findings confirm that machine learning-assisted optimization enhances the overall tribological performance of hydrostatic thrust bearings, establishing a data-driven, computationally efficient framework for the design of surface-textured bearing systems.