A Critical Analysis on the Impact of External Losses on the Performance of a Centrifugal Compressor Stage – Experimental and Multi-Fidelity Numerical Assessment

While the increasing calculation resources are empowering the optimization of turbomachinery components’ design, the role of experiments is still key in developing new products since they do represent the only reliable way to validate simulations. This is particularly true during the “early validation” stage of the design process of new products, where novel solutions going beyond the state of the art need to be verified for the first time. To this end, the possibility of exploiting rigs able to easily test multiple configurations and provide a closer look to the involved physics is increasingly welcome. This is the case, for example, of the rig developed by the University of Florence in support to its industrial partner Baker Hughes. Within this context, a combined experimental and numerical study has been carried out to evaluate the impact of external losses (i.e., disk friction, recirculations and leakages) in defining the real performance of a model stage. While some engineering methods and correlations are present in the literature for this scope, their effective calibration and a more detailed characterization of these phenomena are key to close the gap between simulations and benchmarking experiments. Due to the number of simulations to be carried out in the design phase, in fact CFD calculations often neglect these losses and only focus on impeller/diffuser aerodynamics. If optimization is carried out within a known design space, this does not create any problem and the different solutions can be analyzed in a comparative way. On the other hand, in the early validation of a new design solution, possible changes in the impact of losses can lead to unfair comparison between experimental data and numerical predictions. In the present study, an extended analysis has been carried out using different numerical models and dedicated experiments. In the first part of the same, presented in this paper, we compare most credited engineering loss models with experiments and CFD simulations, referring to either the single vane passage, without any cavity or leakage modeled, or the full passage as tested at the rig. The combined use of experimental and numerical techniques led to a clear overview on the impact of each loss source and proved how numerical simulations can be effectively corrected a posteriori to account for these effects, becoming a more robust tool for the development of novel solutions.