|
|
|
| Fig.7: Phase-averaged streamlines (phase 01, 09 and 17) from experiments (left) and RANS (right) |
Fig.8: Mean velocity and total energy distribution on the centre plane of the square cylinder |
| Results are confined to the flow past a circular cylinder at Re=500 and a slender 2D square cylinder at Re=22 000 , where a large data base of experimental and LES resutls is available. Moreover, an effort is made to elucidate the impact of different inflow and boundary conditions on the results. The evalutaion of the models predictive accuracy is established through both, integral parameters (e.g. Cl, Cd, St) and details of the phase-averaged mean velocity and turbulence field. Unlike, some earlier investigations reults are exclusively reported for low-Re boundary conditions. |
| ||||||||||||||||||||||||
The above sketched simulations using a linear Boussinesq-viscosity approach offer a reasonable predictive accuracy for the circular cylinder simulations. Significant failures occur, however, when attention is drawn to the square-cylinder example. A major difference between these two testcases is the shape of the obstacle. The blunt square cylinder is associated with an enhanced level of irrotational strain upstream of the obstacle, which yields a severe overprediction of turbulence intensity in the vicinity of the stagnation point. The augmented level of turbulence kinetic energy is convected over the obstacle and supresses the shedding mechanism aft of the cylinder.
|
|
|
|
| Fig.7: Phase-averaged streamlines (phase 01, 09 and 17) from experiments (left) and RANS (right) |
Fig.8: Mean velocity and total energy distribution on the centre plane of the square cylinder |
|
1 2 3 4 5 6 |
|