Genetic Algorithms Combined with the Finite Elements Method as an Efficient Methodology for the Design of Tapered Roller Bearings

Abstract

This research presents an efficient hybrid approach based on soft computing techniques and the finite element method for the design of mechanical systems. The use of non-linear finite element models to design mechanical systems provides solutions that are consistent with experimental results; but this use is often limited in practice by a high computational cost. In order to reduce this cost, we propose a linear finite element model that replaces the non-linear elements of the mechanical system with beam and plate elements of equivalent stiffness that are adjusted by means of genetic algorithms. Thus, the adjusted linear model behaves in the same way as the non-linear model, but with a much lower computational cost, which would allow to redesign any mechanical system in a more efficient and faster way. A case study demonstrates the validity of this methodology as applied to the design of a tapered roller bearing.