UFR 1-05 Evaluation
Jump to navigation
Jump to search
Jet in a cross flow
Underlying Flow Regime 1-05 © copyright ERCOFTAC 2004
Evaluation
Conclusion and Comparison of CFD calculations with Experiments
- With eddy viscosity models, caution should be taken particularly with those results that depend on the shear stress
, which controls the lateral spread of the jet.
- As pointed out by Andreopoulos and Rodi (1984), the choice of turbulence model is more critical when low velocity ratios are considered (case of blade cooling) because in these cases even the region near the jet is affected by turbulence.
- The Baldwin Lomax model over the flat plate up to z/D=2.0 and the Oh and Schetz model at the curved jet region have worked well. However the Baldwin Lomax model may have not worked well at the immediate vicinity of the jet exit. This should be confirmed with a very well designed, undistorted, grid at that region.
- The following boundary conditions have produced good results with LES:
- Free slip on the lateral and top surfaces
- No slip on the bottom
- Laminar boundary layer velocity profile at inlet
- Zero gradient for all variables at outlet
- Instantaneous velocity profiles given by a temporal pipe flow simulation, at the pipe flow exit.
- For large eddy simulations, 1.34 x 106 control volumes have been necessary, 45% of them directly downstream of the jet exit. The first control volume next to the jet exit had dimensions Δx=0.004D and Δz=0.006D. Scaled up by the shear velocity of the turbulent pipe flow, we have Δx+=1.4 and Δz+=2.2.
- With the large eddy simulations, two main discrepancies have been found in the comparison of results with experiments. Velocities are consistently higher than measurements in the lower regions of the vertical profiles and to obtain comparable values the simulation velocity ratio must be smaller than that used in the experiment. Two possible reasons have been identified: the jet inflow conditions of the experiments differ from the CFD inflow conditions and the lower Reynolds number of the CFD calculations results in a fundamentally different cross flow boundary layer. These two possible explanations are discussed by Yuan et al. in detail.
- The work by Yagci and Kavsaoglu was based on a Reynolds number of 51400 with respect to the main stream velocity and the jet pipe diameter. This Reynolds number may be equivalent to the typical turbine Reynolds number based on chord. However, their conclusions should be seen with reservations for blade cooling, because the velocity ratio they have considered is high (R=4).
© copyright ERCOFTAC 2004
Contributors: Flavio Franco - ABB ALSTOM Power UK Ltd