Turbulent flow modelling and prediction of blade profile losses in radial outflow turbine

Abstract

Motivated by the high potential radial outflow turbines offer in utilizing waste heat, against relatively limited studies on radial outflow turbine aerodynamics, this thesis has the objective to examine the predictive prospects for determining the ROT primary losses, by using axial turbine loss prediction correlations and various turbulence models. Selected loss prediction models and different turbulence models’ predictive capability and applicability was examined against experimental data obtained from a radial outflow turbine cascade rig, designed and built at LUT University.

Results indicate that Traupel’s loss prediction method can be applied to radial outflow turbines and is the most complete model out of all examined. The method however, overpredicted the loss compared to the modelling results, implying the need of further research for ensuring compliance between the actual and modelled losses. The turbulence models derived for specific dissipation rate, indicated higher agreement with measured data than dissipation rate-based models. Differences between the two groups of models were observed in the wake prediction and blade loading. Out of the two equation turbulence models examined, SST gave most satisfactory prediction of the flow field, proving to be well applicable to modelling radial outflow turbine profile losses. Reynolds stress models did not justify the computational effort given, as the results did not differ significantly compared to the two equation models from the same family. Spalart-Allmaras model, predicted the primary losses in high agreement with Traupel’s correlations result and indicated high compliance in results with the specific dissipation-based models and SST, implying high potential for obtaining reliable results, at a relatively low computational cost.