One-dimensional models predict the performance of centrifugal compressor in short time, being a helpful design tool in the early design stages. They assume uniform flow through the compressor. Conservation of mass, momentum and energy and some empirical loss correlations are applied to estimate the real outputs.
In this thesis, this one-dimensional approach is applied to model a turbocharger compressor. Two different models are implemented. They consist of an impeller, a vaneless diffuser and a volute. The model stage outputs, pressure ratio and efficiency, are compared with experimental data. Then, both models are further investigated by comparing their losses prediction with validated Reynolds-Averaged Navier-Stokes (RANS) data.
The implemented models are taken from literature. They use the same vaneless diffuser and volute approach, but different impeller loss sets. The next impeller losses are studied: incidence, skin friction, choking, jet-wake mixing, blade loading, hub to shroud, tip clearance, shock and distortion losses. The vaneless diffuser outlet is calculated using a one-dimensional numerical solution to the underlying differential equations. For the volute, a set of empirical losses is used. The losses from the CFD are also measured by entropy rise calculations. Due to the complexity of this model, not all the losses can be independently extracted. Incidence, choking, skin friction, blade loading and jet-wake mixing losses are measured along the impeller. Besides, vaneless diffuser and volute losses are also obtained.
Results show relative total pressure ratio errors less than 5% in 49 points in a total of 77 predicted operation conditions. 69 points are estimated with a relative error less than 10%. CFD still gives better predictions, especially at low tip speeds. However, at high tip speeds one-dimensional gives similar accuracy. The one-dimensional and CFD losses comparison shows largest differences in the vaneless diffuser and volute models. Some strengths and weaknesses of the impeller losses are revealed, being possible future improvements.