Prediction and measurement of the rotordynamic response of an automotive turbocharger with floating ring bearings

Abstract

The present investigation advances a rotordynamic analysis of an automotive turbocharger with floating ring bearings. Each fluid film of the floating ring bearing is analyzed using a finite element model. The model includes the effects of lubricant heating and bearing clearance changes due to bearing power consumption. The floating ring bearing analysis provides both floating ring speeds and bearing force coefficients for use in a linear rotordynamic model. The linear rotordynamic evaluation yields predictions of system stability as well as response to imbalance. The linear analysis employs a 27-station finite element model to capture the flexibility of the shaft. The analysis evidences several modes that become unstable at very low speeds, including cylindrical and conical modes for both inner and outer fluid films. Typically, the frequencies of the unstable modes are close to half the ring speed or half the ring plus rotor speeds. These results are supported by a non-linear transient analysis of the rotor-bearing system. The transient predicts the amplitude of the limit cycle motion for the unstable modes. Non-linear transient predictions are generated for a range of lubricant supply conditions with varied supply temperature and pressure. Ring speed and shaft motion predictions are compared to measurements taken on an automotive turbocharger test apparatus. Eddy-current proximity probes measure the shaft motion at both ends of the shaft, while fiber optic probes measure the ring speeds. The non-linear transient predicts the frequency content of the measured response with fair accuracy, but it usually over-predicts the amplitude of the response. In particular, the analysis fails to predict the effect of increasing lubricant supply pressure. In general, increasing the lubricant supply pressure reduces shaft motion amplitude. This investigation also attempts to identify through shaft motion measurements the effects of aerodynamic loading of the compressor, if any. Shaft motion measurements taken with the compressor housing re-oriented (in order to re-direct the aerodynamic load) showed that the aerodynamic load impacted the static displacement of the rotor, but the dynamic response showed little deviation from the original measurements.