A PANS implementation of k-omega SST turbulence model

Abstract

In this thesis, a Partially Averaged Navier Stokes (PANS) turbulence model is implemented for the standard k-omega Shear Stress model (SST). The source code for the PANS implementation of the k-omega SST turbulence model was developed using the C++ programming language. The dynamic PANS turbulence model consists of the active filter function fk, responsible for the transition from the Reynolds Averaged Navier Stokes (RANS) region to the Large Eddy Simulation (LES) region provided with a smoother transition for refined mesh when compared to coarser mesh. The developed turbulence model was first validated by studying the flow around a cylinder. After validation, the PANS turbulence model was used to investigate the flow quantities around BB2 Generic Submarine hull form. The outcome of this study is that the PANS model simulations predicted the flow around the circular cylinder better than the k-omega SST model. The validation was performed for different cylinder grid resolutions for comparison. CFD simulations were then carried out for the submarine hull model at straight flight and 10-degree yaw. The PANS model had better flow prediction along the surface of the submarine when compared to the standard SST model for a straight flight, whereas for the 10-degree yaw, both PANS and SST turbulence model had similar flow predictions.