Simulation of Single-Choked Supersonic Ejectors. Part 2: Validation Study

A validation study has been done to provide evidence on the reliability of Reynolds-averaged simulations for supersonic ejectors operating in the single-choked condition. The existing works show that the simulation accuracy for a given turbulence model strongly depends on ejector geometry and the working conditions. In addition, the performance of different turbulence models is usually presented by comparing the results without discussing the origins of their failures. Part 1 of the present work highlights the importance of properly modeling the turbulence response to the varying shear intensity that characterizes the ejector flow-field. The key elements related to this point have been analyzed for commonly used eddy viscosity models and for a new K−ε model. The models described in Part 1 are tested here on two complementary test cases. A free jet with detailed turbulence measurements in the initial region is used to provide information on the models’ behavior, data which could hardly be obtained directly for ejectors. The simulation of a supersonic ejector is then presented, comparing the results with the measured velocity and pressure distributions. The assessed validity of the new eddy viscosity model is exploited to give insightful considerations on ejector flow-field development.