Aerodynamic analysis of conical diffusers operating with air and supercritical carbon dioxide

Abstract

The design of efficient supercritical carbon dioxide (S-CO2) turbomachinery to be used for power generation (and also for CO2 capture facilities) has gained interest in recent years due to the compactness and good performance of the S-CO2 recuperative Brayton cycle in nuclear, waste heat and solar applications. Presently, in addition to a large amount of theoretical work focused on the analysis and optimisation of the system, there are even some prototypes of centrifugal compressors running on experimental facilities like, for instance, the pilot plant at SANDIA National Laboratories engineered by Barber Nichols Inc. Nevertheless, the performance of this experimental unit is far from an equivalent air/gas turbomachinery, say 80% compressor efficiency, mainly due to a lack of knowledge about the particular behaviour of this working fluid. The need to research these aspects of turbomachinery design has already been identified by the scientific and industrial communities.This work aims to provide more information about diffusion of S-CO2 flows in conical ducts based on the experimental work on air diffusers carried out from the sixties to the eighties (in particular the test conducted by Dolan and Runstadler in 1973). Following a similar approach but by means of numerical analysis (CFD) rather than tests, the work presented here provides a comparison of the expected performance when S-CO2 is used. It is observed that this new working fluid is likely to enhance the pressure rise capability while, at the same time, contribute to reducing the total pressure losses with respect to air.The work commences with a brief introduction to the fundamentals of conical diffusers followed by a discussion on the methodology, mesh size and convergence criteria. Then the most relevant geometric and aerodynamic parameters are assessed prior to providing some conclusions about the expected impact of using S-CO2 on a radial compressor.