Accurate estimation of slip factor is of paramount importance to ensure centrifugal compressor work input is adequately predicted during the preliminary design process. However, variations in the flow field at impeller exit in both the pitchwise and spanwise directions complicate the evaluation procedure considerably. With the increasing implementation of engine downsizing technologies in the automotive sector, achieving a wide operating range has become a factor of prime importance for centrifugal compressors used in automotive turbocharging applications. As a result of the design features required to achieve this aim, modern impeller geometries have been shown to exhibit an approximately parabolic variation in slip factor across their respective operating maps. By comparison, traditional slip correlations typically exhibit a constant, or at best monotonic, relationship between slip factor and impeller exit flow coefficient. It is this lack of modelling fidelity which the current work seeks to address.
In order to tackle these shortcomings, it is proposed that the impeller exit flow should be considered as being made up of three distinct regions; a region of recirculation next to the shroud providing aerodynamic blockage to the stage active flow, and a pitchwise subdivision of the active flow region into jet and wake components. It is illustrated that this hybrid approach in considering both spanwise and pitchwise stratification of the flow permits a better representation of slip factor to be achieved across the operating map.
The factors influencing the relative extent of each of these three distinct regions of flow are numerous, requiring detailed investigations to successfully understand their sources and to characterise their extent. A combination of 3-D Computational Fluid Dynamics (CFD) data and gas stand test data for six automotive turbocharger compressor stages was employed to achieve this aim. Through application of the extensive interstage static pressure data gathered during gas stand testing at Queen’s University Belfast, the results from the 3-D CFD models were validated, thus permitting a more in-depth evaluation of the flow field in terms of locations and parameters that could not easily be measured under gas stand test conditions.
Building on previous knowledge gained about the variation in shroud side recirculation with geometry and operating condition, the characteristic jet/wake flow structure emanating from the active flow region of the impeller was represented in terms of area and mass flow components. This knowledge allowed individual slip factor values for the jet and wake to be calculated and combined to give an accurate passage average value which exhibited the distinctive nonlinear variation in slip across the operating map which is frequently absent from existing modelling methods. Fundamental considerations of the flow phenomena in each region provided explanation of the results, and permitted a modelling approach to be derived to replicate the trends observed in both the experimental data and the CFD simulations.