Integrating CFD and Experiment in Aerodynamics - CFD4Aircraft

More documents

Recommendations

Info

3 2. Surface flow visualisation. Oil flow visualisation gives an indication of surface streamlines, but only in a time averaged sense. Tufts also give an indication of surface streamlines and can reveal flow separation and reattachment, but are limited with the response time in unsteady flows and can also be intrusive. 3. Off surface flow visualisation (smoke/dye). This can provide useful information on shear layer structures and vortex breakdown, but extra care should be taken in interpreting the streakline patterns in unsteady flows. 4. Multi-hole velocity probes. These can measure three-components of mean velocity, but are intrusive and can cause premature breakdown. 5. Hot-wire anemometry. This can provide unsteady velocity components but can be intrusive. 6. LDV and PIV. These are non-intrusive point and field measurements respectively of velocity vectors in a plane. Seeding of vortical flow near the axis becomes problematic with increasing speed in air flows. 2.2 CFD Techniques It has been well documented that CFD has developed at a rapid pace over the past 30 years. With developments in algorithms and computers it is possible to simulate complex flows on real aircraft using cheap computers. A recent NATO technical organisation (RTO) working group has examined the predictive capability for vortical flows on generic delta wing configurations (AVT 80) [2]. 1. Euler simulations can predict vortex breakdown and vortical interactions when a sharp leading edge is used, fixing the separation point. No secondary separation can be predicted since this is due to boundary layer separation having the effect of shifting the primary vortex closer to the wing leading edge. In addition the strength of the leading edge vortex is strongly dependent on the grid used. However, for sharp leading edges this level of modelling is useful for evaluating qualitative behaviour at a low cost. 2. Unsteady RANS simulations can give good prediction for the secondary separation although the prediction of primary separation and vortex formation for rounded leading edge wings has
4 not received much attention in the literature. A major problem with URANS is the prediction of the levels of turbulence in the vortex itself which can strongly influence the development of breakdown. Ad-hoc treatments [3] can be used to limit production in regions of high vorticity but the turbulence levels after breakdown are still too high, making the simulation of the helical instability questionable. 3. Detached Eddy Simulation (DES) [4] has been used to overcome this problem by simulating the large scale turbulence in the vortex by Large Eddy Simulation (LES). In the wing boundary layer, where the cost of LES would be prohibitive at realistic Reynolds numbers, the RANS model is used. Some promising results for the prediction of vortex breakdown have been published, indicating the promise of the approach. The disadvantage is that the simulations are more costly in terms of the finer grids needed in the vortex and the small time steps that are required. In addition, the DES gives no improvement over URANS in terms of the vortex formation from rounded leading edges and predicting the influence of transition. 4. Finally, LES and Direct Numerical Simulation (DNS) [5] have been used at low Reynolds numbers to indicate fundamental physics. The cost of these calculations is prohibitive at flight Reynolds numbers because of the grid and temporal resolution required. CFD predictions have progressed to the point where a current RTO working group (AVT-113) is evaluating the predictions of the flow on the F-16XL aircraft through comparison with in-flight measurements. There are clearly a number of useful tools in the CFD bag with varying cost and predictive capability. 3 Delta Wing Phenomena 3.1 Shear Layer Instabilities The separated shear layers on a delta wing roll up periodically into discrete vortical substructures as visualised by Gad-el-Hak and Blackwelder [6]. This phenomenon was attributed to a Kelvin- Helmholtz type instability of the shear layer. The origin of these structures has been the subject of controversy as several researchers [7] [8] revealed the existence of stationary small-scale vortices
Page 1 and 2: Integrating CFD and Experiment in A
Page 3 and 4: Experimentalist’s requirements fo
Page 5 and 6: persistent discrepancy may be due t
Page 7 and 8: numbers significantly lower than th
Page 9 and 10: Heat flux measurements. Over the pa
Page 11 and 12: The measurements are made on select
Page 13 and 14: advantage that the signal is emitte
Page 15 and 16: 12. Mébarki, Y. and Mérienne, M.
Page 17 and 18: castellated nozzles entrained more
Page 19 and 20: Axisymmetric Regular Convergent Div
Page 21 and 22: Figure 5: Typical instantaneous PIV
Page 23 and 24: Figure 8: Streamwise Velocity Profi
Page 25 and 26: 1.25 CFD (SKW PRESTO) Experimental
Page 27 and 28: 5 Conclusions In this paper we have
Page 29: 2 which include boundary layer and
Page 33 and 34: 6 Although a great deal of effort h
Page 35 and 36: 8 Computational simulations can con
Page 37 and 38: 10 separation location over rounded
Page 39 and 40: 12 3.7 Vortex / flexible wing inter
Page 41 and 42: 14 [8] Mitchell, A.M. and Molton, P
Page 43 and 44: 16 [31] Gursul, I., Proposed Mechan
Page 45 and 46: 18 Figure 3: Spectrum of unsteady f
Page 47 and 48: 20 Figure 7: Flow visualisation of
Page 49 and 50: Figure 11: Upper surface pressure d
Page 51 and 52: While the information presented in
Page 53 and 54: Figure 3 : DFR against Flap angle f
Page 55 and 56: Figure 7 shows a series of plots of
Page 57 and 58: Figure 9 : Contours of Mach number,
Page 59 and 60: ¥:97§Y©R¥e©H !V( H ¨9#f§e!"
Page 61 and 62: é¡êŒëWì¤í?î(ïeð±ñWòL
Page 63 and 64: Ó3×ÓnØ*Ù(Ú*Û+Ü ÝÞ#ß)ÝG
Page 65 and 66: 6‹wB6h67 C,7D4 ')(r,+.st/9:!/L£8
Page 67 and 68: ACBDsEutCIvwlxV^ c Q!Lkd_Q_SFM PQ_^
Page 69 and 70: Ç¹ºF»¼@½:¾¿_À!Á »_ÂÃk
Page 71 and 72: yaw angle. Also, extreme value coef
Page 73 and 74: 0.9 Cs vs yaw angle Side force coef
Page 75 and 76: Fig. 5. Grid block structure around
Page 77 and 78: Cs comparison at V=0.6 m/s Fig. 10.
Page 79 and 80: Investigation of Flow Turning in a
Page 81 and 82:
duct as it approaches the blocker c
Page 83 and 84:
Figure 1. Natural Blockage Thrust R
Page 85 and 86:
Figure 7. Post-Exit rake Total Span
Page 87 and 88:
Figure 11. Comparison of Static Pre
Page 89 and 90:
2.2 Data Aquired The Embedded Laser
Page 91 and 92:
First, plots of the turbulent Reyno
Page 93 and 94:
Particles seeding tube 3 m 1 m mode
Page 95 and 96:
000000000 MEASUREM ENT S MEASUREM E
Page 97 and 98:
Figure 5: Influence of the pseudo t
Page 99 and 100:
Figure 7: Turbulent Reynolds' numbe
Page 101 and 102:
Figure 9: Turbulent Reynolds' numbe
Page 103 and 104:
Figure 11: Vortex Shedding cycle fo
Page 105 and 106:
Figure 13: Comparison with experime
show all

Integrating CFD and Experiment in Aerodynamics - CFD4Aircraft

Create successful ePaper yourself

Delete template?

Save as template?