Physics-Based Modeling of Bearing Based on Finite Element Modeling Technique

Abstract

The physical condition of rolling element ball bearings can be estimated using analytical and empirical methods or predicted grossly using artificial intelligence techniques such as expert systems, fuzzy logic, and neural network, etc. When the operational condition of a bearing is dynamic and there is a need to determine the actual stresses and tolerances of the bearing concisely, then it is wiser to develop a physics-based model of the bearing using finite element modeling (FEM) techniques. Applying such FEM techniques, one can virtually examine any of the possible mechanical characteristics of different types and sizes of ball bearings operating under different speeds and environmental conditions. Understanding such mechanical characteristics is crucial to accurate fault diagnosis of the bearings in practice. For example, such mechanical characteristics can be digitized or mathematically modeled in order to reduce the computational extent of the analysis as well as serve as a reference look-up table for better and faster fault diagnosis purposes than current practices. It can also be applied to determine the remaining useful life of the ball bearing more precisely. In this paper, we present our technical approach toward the development of a physics-based finite element model of rolling element bearings and provide some examples based on results of this research effort.