Trans RINA, Vol 152, Part B1, Intl J Small Craft Tech, 2010 Jan-Jun
adaptive mesh refinement scheme was developed to track the wake according to the maximum axial velocity loss.
The mesh of the hull was realized independently and joined to the one used for the rig, using a domain interface, as proposed by Miyata and Lee [10]. 40,000 hexahedral cells were used to mesh the sub-domain containing the hull. The hull was situated 0.06m below the sails’ foot, as in the experiments. For the simulations without hull, the bottom of the domain (floor) was positioned according to the deck level to ensure a similar gap below the sails.
3.2 (b) Boundary conditions
A no-slip wall boundary condition was assigned to the sails, mast, floor and, whenever applicable, to the hull. The remaining tunnel’s walls (sides and roof) were modelled as free-slip walls to save on computing time. The inlet was assigned a uniform velocity profile with turbulence level similar to the one recorded in the wind- tunnel: turbulence intensity of 0.2% and eddy length scale of
imposed at the outlet. 3.2 (c) Numerical scheme
Each simulation was performed using a second order advection scheme and a convergence
the Shear Stress Transport
turbulence model. 3.3 FORCE CALCULATION
It is common practice when testing sails in wind-tunnels to subtract the contributions of the hull and mast to the measured
force. This is sometimes termed windage
corrections. For consistency, such an approach was also taken with CFD; thus requiring two simulations: with and without sails. The forces acting on the sails were thus obtained as follows:
Lift Jib+Main = Lift Jib+Main+Mast (+Hull) – Lift Mast (+Hull) alone Drag Jib+Main = Drag Jib+Main+Mast (+Hull) – Drag Mast (+Hull) alone
4. RESULTS AND OBSERVATIONS 4.2 AXIAL VELOCITIES AT MID-SPAN 4.1 LIFT AND DRAG DIFFERENCES
The differences of drag and lift between wind-tunnel measurements and
the numerical predictions were
calculated using lift and drag coefficients to allow for the slight fluctuations in air density during the experiments. During each run the air temperature was also was noted. This allowed for the fluctuation of the air density during the experimental period to be dually accounted for.
Diff Drag) = ( Cd
Tunnel Cd
− CdCFD Tunnel
The velocity fields plotted in figure 6(a) and (c) show significant differences between the Large Main and the Small Main. In fact, on the one hand the flow on the latter seems nearly fully attached, but on the other hand, the Large Main exhibits a large separation bubble on its windward side, which is characteristic of a very small angle of attack, as noted by Wilkinson [11]. However, when increasing the angle of attack of the whole rig by 5°, the length of the separation bubble reduces ,as seen in figure 6 (b). This would tend to confirm the hypothesis described in section 4.1.
Figure 5: Differences between experimental results and CFD predictions for the two rigs, with and without modelling the hull
RMS(Residuals)
criteria of (SST)
Diff Lift) = ( Cl
Tunnel Cl
− ClCFD Tunnel
With or without modelling the hull, both drag and lift tended to be underestimated compared to the experimental results, as seen from figure 5. However, it is worth noting that the simulations of the Large Main rig led to differentials 1.7 times higher than for the Small Sail. One hypothesis to explain such a difference resides in a possible small error in the acquisition of the angle of attack from the picture. A new simulation was thus set- up with the angle of attack of the whole rig increased by just 5°. Doing so, the numerical prediction led this time to negative differentials (i.e. over-prediction) of 6.9% for drag and 16.7% for lift. This would thus confirm the hypothesis that a small error of the order of 1-2 degree in the acquisition of the various angles of the sail would be large enough to cause a significant difference between the experiments and the CFD.
Having noted 0.4m. A zero static pressure condition was the influence of the accuracy in the
acquisition of the angles, it is also interesting to highlight the influence of the presence of a hull in the simulation. As seen from figure 5, modelling the hull reduces the difference between the experimental force measurements and the numerical predictions. In fact, the presence of a hull in the simulation tends to increase the drag of the sails by around 6% and increases lift by around 3% for both rigs.
B-4
©2010: The Royal Institution of Naval Architects
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60