Grunnleggende konsepter
The choice of appropriate spanwise extent is critical for accurately predicting the stall and post-stall behavior of airfoils using scale-resolving simulations.
Sammendrag
The study investigates the influence of the spanwise extent on the accuracy of scale-resolving simulations (specifically Implicit Large Eddy Simulations) in predicting the stall and post-stall behavior of the NREL S826 airfoil at a chord Reynolds number of 105.
Key highlights:
- For small angles of attack near stall, the simulations with typical spanwise extents of around 0.1c (where c is the airfoil chord) showed reasonable agreement with experimental data.
- However, for larger angles of attack in the post-stall regime, the computed lift and drag coefficients were found to be highly dependent on the chosen spanwise extent.
- Simulations with smaller spanwise extents (e.g., 0.1395c) exhibited a persistent vortex on the suction side, leading to an overestimation of the lift coefficient compared to the experiments.
- Larger spanwise extents (around 1c) were required to accurately capture the breakdown of this vortex and the associated pressure distribution, resulting in better agreement with the experimental data.
- The analysis of the spanwise autocorrelations of the velocity fluctuations suggested that a spanwise integral length scale of around 0.1 times the spanwise extent can be used as a criterion to determine the sufficiency of the spanwise resolution.
The study highlights the importance of carefully considering the spanwise extent in scale-resolving simulations of airfoil stall and post-stall behavior, as the choice can significantly impact the accuracy of the predicted aerodynamic forces.
Statistikk
The lift and drag coefficients (CL and CD) for the NREL S826 airfoil at a chord Reynolds number of 105 and Mach number of 0.15 are reported for various angles of attack and spanwise extents.
Sitater
"The dramatic influence of the spanwise size even leads to a non stalled airfoil for lz ≤ 0.1395c, only small differences appear for the simulation with lz = 1c and lz = 2c making us confident that, at least for this case, lz = 1c is sufficient to get a converged lift independent of span."
"Empirically, we find that a value of ILS/lz ≈ 0.1 serves as a suitable criterion for defining a sufficient spanwise size, corroborated by the aforementioned analysis."