Optimization of an Air Bearing Device Using Surrogate Modeling and Simplified Conjugate Gradient Method

Abstract

This paper aims to apply Computational Fluid Dynamics (CFD) combined with an optimization method to find the optimum performances of a flat externally pressurized air bearing system After briefly introducing the topic and the air bearing technologies this paper presents the numerical model that have been developed with the commercial package ANSYS The paper then focuses on the optimization method that has been used in this study and after briefly reminds the bases of optimization and CFD-based optimization the Radial Basis Function (RBF) surrogate model is introduced This surrogate model is used to reduce the number of CFD computation required to find the optimal design of the air bearing The Simplified Conjugate Gradient Method (SCGM) is used to find the estimated minimum response of the objective function This paper therefore describes the equations used by SCGM and details the Python program that has been developed to ensure the data transmission between the optimization module and the CFD solver The analyses of the optimization results carried out in the last part of this paper allow to compare the optimized design and the base-line case The design found thanks to the combination of the optimization method and CFD allows to increase of 43 13 % the load capacity of the bearing

Date of Award	2016 Jan 27
Original language	English
Supervisor	Chin-Hsiang Cheng (Supervisor)