Thermal effects on the performance of floating ring bearings for turbochargers

Abstract

Physical models and fast computational programs aim to improve the design and performance of turbocharger (TC) rotordynamics. Most commercial automotive TCs incorporate floating ring bearings (FRBs) owing to their low cost and reduced power losses. However, persistent subsynchronous motions afflict this type of rotor/bearing system, albeit reaching limit cycles that enable their continuous operation. FRBs comprise two fluid films in series and are prone to show one or two subsynchronous instabilities over extended speed ranges of operation. A flow model for prediction of FRB forced response is detailed here. The model incorporates a lumped-parameter thermal energy balance for estimation of the lubricant viscosity and thermal growth of the rotor, bearing and floating ring. The FRB model, fully integrated into a non-linear rotordynamics computational program, predicts the floating ring speed, journal and ring eccentricities, power loss and the rotordynamic force coefficients of the inner and outer films as a function of the load applied at a given rotor speed. Knowledge of the actual load conditions, static and dynamic, and the changes in operating clearance and effective lubricant viscosity are most important for accurate estimation of a TC dynamic forced response. Predictions for the exit lubricant temperature, power losses and floating ring speeds agree well with measurements obtained in an automotive turbocharger test rig. The rotordynamic stability characteristics of the test TC are also highlighted. NOTATIONa, b parameters in Vogel's equation A area for heat transfer (m 2 ) ¼ pDL c p lubricant specific heat (J/kg K) C bearing radial clearance (m) D bearing diameter (m) ¼ 2R e eccentricity vector ðe X , e Y Þ (m) FRB floating ring journal bearing F fluid film force ðF X , F Y Þ (N) h film thickness (m) H heat convection coefficient (W/m 2 K) I R ring mass moment of inertia (kg m 2 ) L bearing length (m) M mass (kg) } mechanical power (W) P hydrodynamic pressure (Pa) Pr Prandtl number ¼ c p m/k Q volumetric flowrate (m 3 /s) R radius (m) ¼ D/2 Re Reynolds number ¼ rORC/m TC turbocharger T temperature (8C) W external force ðW X , W Y Þ ( N) x, z circumferential and axial coordinate on the plane of the film, x ¼ Ry X, Y inertial coordinate system a thermal expansion coefficient (K À 1 ) g lubricant shear rate (m À 1 ) G torque (N m) DT temperature difference ðKÞ ¼ T À T S e dimensionless eccentricity ¼ e/C y angle (rad) ¼ x/R k lubricant thermal conductivity (W/m 2 K) m lubricant viscosity (N s/m 2 ) r lubricant density (kg/m 3 ) O rotational speed (rad/s) O ring speed ratio ¼ O R /O J The MS was

Thermal analysis of small high-speed floating-ring journal bearings

“…Early investigations of (semi) floating ring bearings (summarized in [10,11]) were limited to experiments at low shaft speeds used for verification of isothermal analyses only. Ring speeds predicted by isothermal flow models exhibit poor correlation with respect to test data especially at high shaft speeds [12].…”

“…San Andrés et al [10,11] developed a model incorporating a lumped-parameter thermal energy balance for the estimation of the lubricant viscosity and change of operating clearances in the fluid films. The model predicts the bearing performance based on an effective lubricant viscosity and assumes constant film and ring temperature throughout the flow domains.…”

Predictions for run-up procedures of automotive turbochargers with full-floating ring bearings including thermal effects and different bearing setups

“…This model is further applied and verified by measurements of Meyer [5]. An alternative model is presented by San Andres and Kerth [6] and validated with test data. However, for hot gas applications all parameters strongly depend on the heat disposed from a hot shaft [7].…”

Towards Accurate Prediction of Unbalance Response, Oil Whirl and Oil Whip of Flexible Rotors Supported by Hydrodynamic Bearings

“…Heat flow from the fluid film to its surroundings can be modeled by solving the energy equation in three directions, as done by Mahner et al [21], or alternatively by assuming an average temperature at the center of the height of the fluid film [22]. For this model, we assume the latter in order to keep the computational effort at a minimum during time-transient simulations of the rotor-bearing system.…”