Numerical Investigation of Boundary Layers in Wet Steam Nozzles

Abstract

Condensing nozzle flows have been used extensively to validate wet steam models. Many test cases are available in the literature, and in the past, a range of numerical studies have dealt with this challenging task. It is usually assumed that the nozzles provide a one- or two-dimensional flow with a fully turbulent boundary layer (BL). The present paper reviews these assumptions and investigates numerically the influence of boundary layers on dry and wet steam nozzle expansions. For the narrow nozzle of Moses and Stein, it is shown that the pressure distribution is significantly affected by the additional blockage due to the side wall boundary layer. Comparison of laminar and turbulent flow predictions for this nozzles suggests that laminar–turbulent transition only occurs after the throat. Other examples are the Binnie and Green nozzle and the Moore et al. nozzles for which it is known that sudden changes in wall curvature produce expansion and compression waves that interact with the boundary layers. The differences between two- and three-dimensional calculations for these cases and the influence of laminar and turbulent boundary layers are discussed. The present results reveal that boundary layer effects can have a considerable impact on the mean nozzle flow and thus on the validation process of condensation models. In order to verify the accuracy of turbulence modeling, a test case that is not widely known internationally is included within the present study. This experimental work is remarkable because it includes boundary layer data as well as the usual pressure measurements along the nozzle centerline. Predicted and measured boundary layer profiles are compared, and the effect of different turbulence models is discussed. Most of the numerical results are obtained with the in-house wet steam Reynolds-averaged Navier–Stokes (RANS) solver, Steamblock, but for the purpose of comparison, the commercial program ansys cfx is also used, providing a wider range of standard RANS-based turbulence models.