Turbulence Models and Heat Transfer in Nozzle Flows

Abstract

Introduction S IGNIFICANT resources have been spent on creating accurate and robust turbulence models for flows with adverse pressure gradients. Capturing correct boundary-layer profiles under adverse pressure gradients is of particular importance to the accurate modeling of separated flows. In favorable pressure-gradient flows, smallscale features of boundary layers generally do not couple to largerscale flow features such as flow separations. As a result, problems that can occur in regions of favorable pressure gradients are more difficult to spot in standard benchmarks. However, accurate modeling of such flows will be required for reliable predictions of surface heat fluxes or shear stresses. Here we investigate the performance of Menter’s two equation shear-stress-transport (SST) turbulence model1 for heat-transfer computation under strongly favorable pressure gradients found in choked nozzle flows (such as required in rocket nozzle analysis and design). We have observed anomalous behavior of the SST turbulence model for strongly favorable pressure-gradient flows. We shall demonstrate the anomaly using an experimental choked nozzle problem published in the open literature.2 The anomaly in the computed heat transfer appears to be caused by the shear-stress-transport correction term of Menter’s model. By design, this term should be active in near-wall regions under adverse pressure gradients. However, it also appears to incorrectly become active under strongly favorable pressure-gradient conditions.