Quantification of sub-synchronous vibrations for turbocharger rotors with full-floating ring bearings

Abstract

The correct capture and understanding of the bearing- induced rotor vibrations is nowadays a rather compulsory task, which should accompany the modeling and simulation work flow of high-speed rotor systems, such as turbochargers. The oil-film concentrated in the rotor’s journal bearings is the root cause of the system’s occurring nonlinear effects known as sub-synchronous vibrations, the behavior of which depends on both the system’s geometric and dynamic configuration. In this paper, a methodology is applied for the case of a turbocharger with full-floating ring bearings that allows the quantification of the sub-synchronous vibrations during run-up simulations. It is conducted by considering both the wheel shaft and shaft-bearing geometry as a set of input parameters, the variation of which contributes in quantifying their influence upon the sub-synchronous evolution with respect to amplitude and duration criteria. Motivated by linear multivariate regression algorithms and data mining techniques, i.e., correlation coefficients and global sensitivity methods, the influence of each design parameter on the sub-synchronous formation is analyzed. Furthermore, with the help of the non-supervised neural network methods, design configurations are indicated that could be set as a compromise in terms of feasibility and low-cost production.