Parallel Iterative Methods for Modeling Foil Air Bearings
| Autor: | Shih Hung Chang, 張世宏 |
|---|---|
| Rok vydání: | 2012 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 100 The applications of air-lubricated bearings have several advantages over liquid-lubricated bearings, such as low frictional resistance, high rotating speed and easily accessed lubricant. Air bearings can also be operated in extreme low or high temperature environments to meet some special demands. But for some forms of air bearings, such as plain journal air bearings, the stability in high speed operation is a major concern. Compared with externally pressurized air bearings, foil air bearings can have higher load capacity and stiffness as well as increased stability. However, the stability gain comes with a slightly increase of bearing weight, volume and cost. In a performance analysis of foil air bearings, the solution procedure usually requires a computation which consists of a multi-level-deep nested loop. And the numerical solution of the Reynolds equation is the bottleneck of the computation. When iterative methods are used to solve the linearized Reynolds equation, the stopping criterion in the iterative procedure plays an important role on solution efficiency as well as accuracy. In this study, the proposed stopping criterion is based on the truncation error analysis. It can be seen that a minimum number of iterative steps without sacrificing solution accuracy can be obtained. Several iterative methods are compared for their solution efficiency in solving the Reynolds equation. The methods tested are the successive over-relaxation method, a preconditioned conjugate gradient method, and a multi-grid method. In the elastohydrodynamic simulation of foil air bearings, an iteration process is required to have a converged film profile and pressure. In searching a converged film profile the numerical solution is usually fluctuated with possible solution divergence. In this study, the incorporating of numerical damping and hybrid root-finding process are introduced, which minimize the solution fluctuation and prevent the solution divergence. And parallel computing is one of the most eminent tools for reducing execution time in those cases. In this study, the OpenMP programming for multithreaded computing is used. In this study a fast solution procedure for foil air bearings modeling is presented. The procedure can be used as the foundation for optimization study of similar air bearings or other types of bearing having elastohydrodynamic lubrication behavior, which usually require extended computational time. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|