An Efficient Finite Element-Based Air Bearing Simulator for Pivoted Slider Bearings using Bi-Conjugate Gradient Algorithms

A computationally efficient air bearing simulator-for pivoted slider bearings has been developed that is based on finite element theory and uses bi-conjugate gradient algorithms in conjunction with a sparse matrix storage scheme. The air bearing simulator involves the simultaneous solution of the Reynolds equation and the slider equilibrium equations. The highly nonlinear problem requires the repeated solution of a nonsymmetric system of equations, typically with a large number of unknowns depending on the complexity of the slider geometry. Iterative solvers, such, as the bi-conjugate gradient algorithms used for this study, require significantly less core memory as compared to direct solvers and reduce the solution time if combined with a suitable preconditions. Of the bi-conjugate gradient algorithm/preconditioner combinations implemented, the Bi-CGSTAB algorithm combined with an ILU preconditioner provided the best performance in terms of smooth convergence and computational efficiency. Presented as a Society of Tribologists and Lubrication Engineers paper at the STLE/ASME Tribology Conference in Kissimmee, Florida, October 8-11, 1995