Tandem cylinder flow and noise predictions using a hybrid RANS/LES approach

The performance of a novel hybrid RANS/LES methodology for accurate flow and noise predictions of the NASA Tandem Cylinder Experiment is investigated. The proposed approach, the modified Flow Simulation Methodology (FSM), is based on scaling the turbulence viscosity and the turbulence kinetic energy dissipation rate with a damping function. This damping function consists of three individual components, a function based on the Kolmogorov length-scale ensuring correct behaviour in the direct numerical simulation (DNS) limit, a function ensuring that FSM provides the correct damping in large-eddy simulation (LES) mode, and a shielding function that forces the switch from Reynolds-Averaged Navier-Stokes (RANS) to LES to occur outside the boundary layer. The FSM is proposed for the kω-SST two-equation model (FSM-SST) and for an Explicit-Algebraic-Stress-Model (FSM-EASM), which is better suited to resolve anisotropy and non-equilibrium of the unresolved scales and the strain and rotation-rate dependent coefficients introduce a dynamic response of the model to the resolved flow field. Simulations are performed on a relatively coarse grid and the FSM data are compared with results obtained from the Scale-Adaptive-Simulation (SAS) and IDDES approaches. Acoustic predictions are obtained using an acoustic analogy approach based on Curle's theory. The FSM-SST approach was found to predict the hydrodynamic field in very good agreement with reference data, whereas the FSM-EASM did not improve the predictions. The acoustic spectra predicted show good agreement with experimental results at various microphone positions, with some deficiencies in capturing the broadband noise levels at high Strouhal numbers.