Towards robust design optimization of automotive turbocharger rotor-bearing systems

In the competitive automotive market, the performance of turbochargers is constantly being pushed towards their theoretical optimum. One of the key components of the turbocharger is the rotor-bearing system, which determines the friction losses and noise output and furthermore affects the overall turbocharger efficiency, reliability and cost. In order to fulfil the demands of the automotive market, developing methods to optimize the rotor-bearing system is the focus of this study, where particular attention is paid to taking into account the product-to-product variations that are inevitable in cost-effective mass-produced parts, as well as the variations in turbocharger operating conditions.

First, a model of the rotor-bearing system was developed to predict the rotordynamic response over the operating range. The model is constructed in a step-by-step fashion, starting with a simple test case: a Laval rotor supported by plain journal bearings. As the behavior of the rotor-bearing system varies over its rotation speed range, run-up simulations were performed by a time-transient multi-physical model. In this model, several sub-models are coupled: a rotordynamic sub-model, a thermo-hydrodynamic submodel and a thermal network model.

Once a satisfactory correlation was found between numerical simulation results and measurement results, the test case progressed to a Laval rotor with floating ring bearings instead of plain journal bearings. Correspondingly, the bearing model was extended to include the dynamics of the floating ring and its two oil films. The resulting run-ups showed a response consisting of a critical speed, an oil whirl and an oil whip.

Analysis of a turbocharger rotor-bearing system was subsequently performed, showing a more complex response, consisting of multiple critical speeds and the co-existence of sub-synchronous whirling modes. The effect of the rotor-bearing operating conditions, unbalance configuration, the thrust bearing and the bearing cylindricity were investigated. Most of the trends are correctly predicted by the model, however the correlation between measurement results and simulation results was clearly inferior to the case of the Laval rotor, most likely due to the uncertainties in the actual turbocharger geometry and the actual unbalance distribution.

Lastly, an optimization of a Laval rotor-bearing system was performed. The resulting robust optimum design ensures optimum rotor-bearing performance, even at the most severe operating conditions and even if all manufacturing tolerances represent the worst case scenario. Particularly the uncertainties in rotor unbalance and oil supply temperature were found to have a significant influence on the optimum design.