COMPARISON OF HYDRODYNAMIC PERFORMANCES OF AN IMOCA 60 WITH STRAIGHT OR L-SHAPED DAGGERBOARD

COMPARISON OF HYDRODYNAMIC PERFORMANCES OF AN IMOCA 60 WITH STRAIGHT OR L-SHAPED DAGGERBOARD L. Mazas and Y. Andrillon, HydrOcean, France, loic.mazas@hydrocean.fr A. Letourneur and P. Kerdraon, VPLP, France, letourneur@vannes.vplp.fr G. Verdier, Verdier Architecture Navale, France, gverdier@guillaumeverdier.com The milestone of the 8 th edition of the Vendée Globe in 2016 was the apparition of foils on IMOCA 60. These foils are L-shaped daggerboards pointing out of the sides of the hull; they were designed so as to create lift on leeward. The combined effect of lifting the hull partially out of water and increasing the righting moment improves the yachts performances. The match between foilers and classical IMOCAs gave advantage to foilers. The evaluation and prediction of sailing yachts hydrodynamic performances in waves is crucial to draw reliable speed polars. IMOCA 60 dynamic behaviour and hydrodynamic forces and moments depend very much on their appendages. The aim of this study was to compare the performances of two types of daggerboard: one classical with a straight shape, one foil with a L-shape, called CLOUT ; they were appended on the same IMOCA hull. Some of the CFD simulations were performed by HydrOcean, an engineering company specialised in marine CFD, for naval architects at VPLP design and Guillaume Verdier Architecture Navale, while another CFD group was operating foil and hull resistance analysis under the direction of Leonard Imas. This specific comparison of CLOUT versus STRAIGHT daggerboard was performed by HydrOcean. Two sailing configurations were evaluated: upwind with 30 incident waves and downwind with 135 incident waves. The simulations showed that CLOUT daggerboard was very disadvantageous upwind, with an increase of 15% in drag and a loss of 7% in righting moment. However, it seemed interesting downwind, with a drag reduction of 9% and a gain in righting moment of 8%. Still, these advantage are to be balanced with the sensitivity of CLOUT s dynamic behaviour to waves: the boat drifts to leeward. That study enabled to point out the advantages and drawbacks of this daggerboard concept and it gave some clues for the improvements that could be brought to foil design. NOMENCLATURE λ Wave length (m) DOF Degree of freedom Fx Drag force (N) Fy Side force (N) Fz Lift force (N) H Wave height (m) Mx Righting moment (N.m) Ry Dynamic sinkage ( ) T Wave period (s) Tz Dynamic sinkage (m) 1 INTRODUCTION The sailing yacht studied is an IMOCA 60 heeled at 25, appended with a keel canted 38 leeward and a daggerboard. Two geometries of daggerboards were evaluated with the same IMOCA 60 hull form: one straight daggerboard; called STRAIGHT, and one L-shaped called CLOUT (for C-shaft, L-tip, outward pointing). Figure 1 presents views of the two daggerboards appended on the same hull. STRAIGHT has a 590 mm chord while CLOUT s is 450 mm. Figure 1 - IMOCA hull appended with a canting keel and with both daggerboards 2 MODEL SETUP 2.1 COORDINATE SYSTEM All presented hydrodynamic forces (Fx, Fy, Fz) and moment (Mx) are averaged values and given in the coordinate system linked to boat s centre of gravity (cf. Figure 2). x-axis is oriented to the bow, y-axis portside and z-axis up. Positive leeway is leeward, positive trim is bow down and positive dynamic sinkage means the boat goes up. It is fixed in orientation and its origin follows the boat s motions. The centre of gravity varies depending on sailing upwind or downwind due to different ballast configurations. 163

2.3 CALM WATER SIMULATION SETUP First, simulations in calm water were performed, for upwind and downwind sailing configurations. The boat was set free in heave and pitch; its speed was imposed: 10 kts upwind and 25 kts downwind. Figure 2 - Coordinate system 2.2 NUMERICAL MODEL SETUP Unsteady simulations were performed at full scale using CD-adapco s STAR-CCM+ RANSE solver, release 8.04.007. Menter s k-ω SST turbulence model was used. The simulation domain was composed of two regions, one large background region, defined with the boundaries of the domain, and one small overlapping region, containing the IMOCA, used for the motion. The overset region was free to move within the background region, thanks to the overset mesh method of the solver. Figure 3 shows the two regions of the simulation domain. 2.4 WAVES SIMULATION SETUP For these simulations, a 5 th order Stokes regular wave model was imposed gradually by restarting the calm water computation at its final time. The waves characteristics are listed in Table 1. Case Heading ( ) λ (m) H (m) T (s) Upwind 30 80.0 1.5 7.2 Downwind 135 25.0 0.75 4.0 Table 1 - Waves characteristics Similarly the boat was set free in heave and pitch for upwind simulations. For downwind simulations, the DOF in sway was also set free. In this case, the combination of the motions during the simulation enabled to have a leeway angle varying over time. An aerodynamic force equals to the boat calm water resistance was imposed, with a different towing point for upwind and downwind configurations due to different sails. Table 2 summarizes the various configurations. Figure 3 - Simulation domain: background and overset regions The boundary conditions imposed on the domain were velocity inlets on all faces (red), except the further aft the boat (orange), which was a pressure outlet. The boundaries of the overset region were of a special type, overset mesh, to handle the overlapping. The mesh contained approximately 0.4 million cells for the background region and 4.1 million for the overset region. High levels of refinement were applied on the hull, the keel and the daggerboard as well as near the free surface location. A refinement box was also set around the overset region, so as to ensure correct cell dimensions between the two regions for the overset algorithm to work. A view of the volume mesh for a simulation in waves is presented on Figure 4 and the boundaries of the overset region are highlighted in red. CALM WATER WAVES Wind Heel ( ) Speed (kts) DOF Up 25 10 Ry, Tz Down 25 25 Ry, Tz Up 25 10 Ry, Tz Down 25 25 Ry, Tz, Ty Table 2 - Sailing configurations on calm water and waves 3 RESULTS IN CALM WATER 3.1 UPWIND AVERAGED VALUES For upwind computations in calm water, a different leeway angle was imposed between STRAIGHT and CLOUT in order to compare both cases with the same total side force (iso-fy). This means both boats have the same aerodynamic forces. Thus, CLOUT was imposed 0 leeway angle while STRAIGHT was imposed -3 (negative leeway is windward). On Figure 5 are plotted the hydrodynamic loads and their relative differences for both configurations at 10 kts, upwind sailing. The relative differences are computed according to the following: diff. (%) = (CLOUT - STRAIGHT)/STRAIGHT Figure 4 - Front view of the volume mesh and the overset region CLOUT is penalized by an increase of 11.7% in drag compared to STRAIGHT. This increase comes from the hull (+19.7%) while the daggerboard in itself has less 164

drag (-20.0%), partly because its chord is smaller (-24%). The keel drag is also increased for CLOUT (+10.7%). The different leeway angles between the two configurations give close values for the total side force Fy (-2.7%). However, noticeable differences are observed on its repartition: CLOUT daggerboard loses 26.1% while the keel brings three times more Fy than for STRAIGHT. Concerning the lift Fz, a small difference of 1.4% is observed; it is because the boats are not exactly at the equilibrium with the displacement imposed (9.2 t). The differences with the displacement imposed is -1.2% for STRAIGHT and 0.2% for CLOUT: the convergence is not fully reached. Again, the repartition between CLOUT and STRAIGHT is very different on the daggerboard (- 76.4%) which is compensated with the keel (+60.1%). Fz on the hull does not vary much (1.4%). As for the righting moment Mx, it is slightly lower for CLOUT (-4.5%). It shows a huge loss on the daggerboard (-148.0 %): Mx becomes positive (1920 N.m) whereas it is negative for STRAIGHT (-3996 N.m). For CLOUT, the daggerboard tends to heel the boat upright whereas for STRAIGHT, it tends heel the boat more. The decrease in total Mx for CLOUT is due to a higher decrease on the keel (+58.1%). Table 3 gives the values of dynamic sinkage and trim. The absolute difference in sinkage shows that CLOUT is lower in altitude than STRAIGHT, which is in agreement with a higher wetted surface, higher drag and more displacement in the hull. Dag. Tz (m) Ry ( ) STRAIGHT 0.000 1.702 CLOUT -0.050 1.482 diff. CLOUT- STRAIGHT -0.050-0.220 Table 3 - Averaged values of trim and sinkage for upwind sailing in calm water Thus, during upwind sailing on calm water, the L-shaped daggerboard CLOUT seems to penalize the boat s hydrodynamic performances. Indeed, for the same side force Fy than STRAIGHT, CLOUT increases drag by around 9% and reduces the righting moment by around 5%, which means that the boat can bear less sail surface. The CLOUT daggerboard certainly could be improved by increasing the lift on both tip and shaft through a change in implantation angles on the boat. This could help drift less, and bring righting moment back. 3.2 DOWNWIND AVERAGED VALUES For downwind computations in calm water, a different leeway angle was imposed for STRAIGHT and CLOUT in order to compare both cases with the same righting moment (iso-mx). This means both boats can bear the same sail surface. Thus, CLOUT was imposed -1.5 leeway angle while STRAIGHT was -6.5. On Figure 6 are plotted the hydrodynamic loads and their relative differences for both configurations at 25 kts, downwind sailing. Figure 5 - Hydrodynamic forces and moments for upwind sailing in calm water 165

First, the total Mx values are quite different (-7.9%), so the comparison is not thoroughly meaningful. However, it still gives interesting trends to analyse. Contrary to upwind, the drag seems to be at the advantage of CLOUT with a decrease of 5.6% compared to STRAIGHT. It is essentially due to the decrease of the hull drag (-12.2% i.e. 1.3 t), since the daggerboard and keel add drag (resp. +10.4% and +4.5%). It is because the CLOUT daggerboard makes the boat more bow up, so the keel and the daggerboard lift more the hull out of the water: that decreases the drag. The difference in wetted surface on hull body between CLOUT and STRAIGHT is -11.2%, which is in agreement with the decrease of the hull drag. This is the major gain. The comparison of side force also gives advantage to CLOUT (+42.0%) with a net increase of the hull contribution (+131.0%) and the keel (+84.1%), but a small loss on the daggerboard (-7.1%). The main interest of the CLOUT daggerboard lies on the vertical lift Fz: it decreases the load on the hull by almost 2 t (-28.3%) because the force on the daggerboard almost doubles (+81.9%, i.e. 0.7 t) and also increases on the keel (+69.0%, i.e. 1.5 t). As observed for upwind sailing, CLOUT shows a loss in Mx, this is due to the following factors. For CLOUT, the keel creates more lift and the hull is lighter, this leads to a loss of righting moment on the hull and the keel. The daggerboard lifts more leeward, and does not compensate this loss in righting moment. Concerning the dynamic behaviour, CLOUT is higher in altitude and more bow up than STRAIGHT, which is in agreement with a smaller wetted surface area, less drag and 2 t less in the hull. Dag. Tz (m) Ry ( ) STRAIGHT 0.006 0.785 CLOUT 0.071 1.243 diff. CLOUT- STRAIGHT 0.065 0.458 Table 4 - Averaged values of trim and sinkage for downwind sailing in calm water Finally, during downwind sailing on calm water, CLOUT daggerboard seems to give advantage of around 6% drag reduction and 2 t less of displacement in the hull. Figure 6 - Hydrodynamic forces for downwind sailing in calm water 4 RESULTS IN WAVES 4.1 UPWIND AVERAGED VALUES It is important to note that this CFD study in waves was performed at constant speed. In real, the driving force should be constant and the speed should vary together with the yaw/pitch/roll. It thus largely affect results but 166

still, the presented ones are interesting for showing trends. For upwind simulations in waves, a different leeway angle was imposed between STRAIGHT and CLOUT in order to compare both cases with the same total side force (iso-fy). Thus, CLOUT was imposed 0 leeway angle while STRAIGHT was imposed at +3.5. On Figure 7 are plotted the hydrodynamic loads and their relative differences both configurations at 10 kts, upwind sailing in waves. As observed for upwind sailing in calm water, CLOUT does not perform well in waves concerning drag. Indeed, the total drag is increased by 15.3% compared to STRAIGHT, mainly because of the hull. The wetted surface increases by 3.7% for CLOUT. Side force Fy is very close for STRAIGHT and CLOUT (0.6%), so the different leeway angles are appropriate. Concerning lift distribution, CLOUT daggerboard does not fulfil its role because it brings negative lift (it tends to make the boat sink). Thus the hull is more heavily loaded (+6.4%) and so is the keel (+47.2%). As for Mx, here again CLOUT performs less than STRAIGHT, with a loss of 7.4% in total Mx, mainly due to the keel (-8.7 t.m). Dag. Tz (m) Ry ( ) STRAIGHT -0.043 1.793 CLOUT -0.058 1.744 diff. CLOUT- STRAIGHT -0.015-0.049 Table 5 - Averaged values of trim and sinkage for upwind sailing in waves Differences on sinkage and trim are very small according to the values in Table 5. Finally, during upwind sailing in waves, CLOUT daggerboard drastically penalizes the boat with an additional 15% drag and -7% of righting moment. Figure 7 - Hydrodynamic forces and moments for upwind sailing in waves 4.2 DOWNWIND AVERAGED VALUES For downwind simulations in waves, the boat was set free in sway, that is, translation along y-axis. Initial leeway angle was -5.1 for STRAIGHT and -3.5 for CLOUT with the aim to obtain the same side force for both configurations. Since the boat was free in sway, leeway angle changed during the simulation; the 167

averaged effective leeway angles were -6.6 for STRAIGHT and -1.4 for CLOUT. On Figure 8 are plotted the hydrodynamic loads and their relative differences for both configurations at 25 kts, downwind sailing. As far as drag is concerned, CLOUT gives a real advantage with a reduction of 9.3%. The hull drag does not vary much between the two daggerboards (-0.9%). It is because the decrease in drag given by the lift force of the daggerboard (generating a decrease of 4.7% in wetted surface) is counterbalanced by an increase in drag due to the difference in leeway angle: CLOUT drifts more (-1.4 ) than STRAIGHT (- 6.6 ), thus the hull form sails more off-axis its lines, generating more drag. The daggerboard drag is reduced by 24.7% for CLOUT. There is also a decrease in drag on the keel (-27.7%) because it generates more lift so its upper part near the hull is more often outside of the water. Concerning the total side force, it is slightly different (+7.4%) because of the different speeds in sway. It is interesting to note that the distribution of the side force is totally different for both configurations. For STRAIGHT, the hull does not brings side force (0.5%): it sails in-axis. For CLOUT however, the hull brings around 19% of the total Fy because it sails off-axis. The daggerboard brings around 55% of total Fy for STRAIGHT whereas it accounts only for 28% for CLOUT. That s why the keel develops more side force for CLOUT (53%) than STRAIGHT (45%). The lift increase due to daggerboard for CLOUT (+79.5%, i.e.0.9 t) enables to lighten the hull (-27.4%, i.e. 1.6 t). The increase in lift brought by the keel (+23.6%, i.e. 0.7 t) can be explained by a higher leeway and trim angles, leading to a better angle of attack of the keel fin. For righting moment, there is a huge difference between the two daggerboards: CLOUT has around 9.4 t.m more righting moment than STRAIGHT (in the global coordinate system) but this increase goes with a loss in Mx of the hull (-5.8 t.m, ). As for the keel, it tends to decrease slightly Mx on CLOUT (-0.07 t.m). All in all, the total righting moment of the boat, is increased by around 3.5 t.m for CLOUT. To conclude concerning averaged values for downwind sailing in waves, CLOUT shows an advantage with a drag reduction of 9% compared to STRAIGHT. For side force, the averaged leeway angle differs greatly between the two configurations. This shows that CLOUT daggerboard could be improved, by an increased shaft area to bring the hull back in its lines. As for the righting moment, there is a real gain of around 3.5 t.m in favour of CLOUT. Figure 8 - Hydrodynamic forces and moments for downwind sailing in waves 168

4.3 DOWNWIND BEHAVIOUR IN WAVES Screenshots of both boats sailing in waves are presented for different times in Figure 9: 9.8 s, 11.2 s, 12.6 s, 14.0 s, 15.4 s and 16.8 s. It should be noticed first that both boats do not encounter the waves at the same times. While STRAIGHT s bow is climbing up a wave on the first two pictures, CLOUT s bow is still going down in a trough. Then, looking at the sequence of images, it is remarkable that CLOUT drifts slowly after each wave crest, whereas STRAIGHT seems to drift to windward. Pressure coefficients shown on Figure 10 help to understand the differences observed in the boat s courses. In the sequence, variations of the pressure coefficient on the keel and daggerboard are hardly visible for STRAIGHT. For CLOUT however, pressure coefficient is varying greatly over time, especially on the shaft of the daggerboard and the keel fin. It gets close to zero at 11.2 s and 12.6 s (second and third images), when the boat is at the top of the wave, the bow going down. Figure 9 - Boat s position in waves at different times The pressure coefficients increase in absolute value (deep blue) when the boat is riding down the wave (last image). 169

Figure 10 - Pressure coefficient at different times 4.4 DOWNWIND TEMPORAL ANALYSIS The aim of this subsection is to understand the substantial drift of CLOUT configuration when sailing downwind in waves. The study of the time series of forces and motions help to analyse the differences in dynamic behaviour of the boats. On the time series of Figure 11, dotted vertical lines represent the times at which the pictures shown previously were taken. Figure 11 - Temporal evolution of side force, trim, leeway and velocity in y-direction for downwind sailing in waves 170

The bottom plot of Figure 11 shows the time evolution of the sway velocity Vy. One can see that CLOUT endures important variations while almost constantly drifting leeward whereas STRAIGHT has a small negative Vy. The trim time series show a 2 s phase shift between STRAIGHT and CLOUT which is due to the difference in leeway leading to different period of encounter An important observation is that the amplitudes of all plotted parameters are higher for CLOUT than for STRAIGHT, which means CLOUT is more sensitive to the waves. For example, there is a difference of almost 1 in pitch behaviour: maximum trim is 4 for CLOUT and 3 for STRAIGHT. As far as yaw is concerned, there is also an important difference with an amplitude of 2.7 for STRAIGHT and 6 for CLOUT. The dynamic behaviour of CLOUT, as shown by the previous computations, presents very disappointing results in terms of seaworthiness. They can be explained as follow. When a wave crest reaches the boat, its bow rises. This increases the angle of attack of the daggerboard tip, thus producing more vertical force Fz and trimming the boat even more. As such, a L-shaped foil has reversed low pressure side between shaft and tip, this phenomenon leads to great loss of the shaft side force Fy. Therefore the leeway increases, which decreases the tip angle of attack and brings the bow down. In this way, the boat dynamic trim behaviour is amplified and its averaged leeway increased dramatically, compared to sailing in calm water. This analysis showed that, even if the daggerboard concept of CLOUT is interesting to decrease drag, it could be improved to give more side force, by specially studying the interaction of the shaft and the tip of the daggerboard as well as the position of the daggerboard on the hull. 5.2 DOWNWIND Figure 12 - Summary of forces (upwind) Downwind, significant differences are observed on the hull. For STRAIGHT, the hull does not provide side force, only lift. As for CLOUT, the hull provides both lift and side force. The keel gives more lift for CLOUT than for STRAIGHT, because of the smaller leeway angle. As there is less side force on CLOUT daggerboard than on STRAIGHT, the side force is more distributed on the keel and the hull. Then, the lift brought by the daggerboard is higher for CLOUT than for STRAIGHT. 5 SUMMARY OF SIDE AND LIFT FORCES 5.1 UPWIND The drawing on Figure 12 summarizes in a simplified manner the side and lift forces acting on each element of the boat for both daggerboards. Plain green arrows are for CLOUT while red dashed arrows are for STRAIGHT. If the component on the hull is very similar in magnitude and direction, there are noticeable differences on the keel and daggerboard. Indeed, CLOUT s keel give more side force and lift than STRAIGHT s. However, the arrows of the daggerboard are very different. It is horizontal (even slightly pointing down) for CLOUT, meaning it does not bring any lift at all, only side force. For STRAIGHT, the daggerboard is inclined and gives the boat some vertical lift as well as side force. Figure 13 - Summary of forces (downwind) 171

6 CONCLUSIONS Although CLOUT daggerboard seems very promising on paper because of its versatility a vertical area, the shaft and a lifting area, the tip, bringing side force upwind and lifting one downwind this study revealed it had some major drawbacks. From the calm water simulations, it seems that CLOUT could be improved by altering the daggerboard angles on the hull as well as the involved surfaces. However, the behaviour on waves totally discredits CLOUT concept of profile camber inversion between shaft and tip. Another daggerboard concept with no inversion was therefore developed with a different role distribution: due to an increased cant angle on hull, the shaft is responsible for the lifting force while the tip creates side force. This is the Dali foil concept that has been used on the 2016 Vendée Globe (cf. Figure 14, yellow). A. Letourneur holds the current position of Sailing Performance Engineer at VPLP design. He is in charge of CFD studies on offshore sailing yacht. His previous experience in yacht design includes aerodynamic and hydrodynamic studies on foiling catamarans, maxitrimarans and IMOCA monohulls. P. Kerdraon will begin a PhD with VPLP design and École Centrale de Nantes in sailing dynamics. G. Verdier holds the current position of Naval Architect at Guillaume Verdier Architecture Navale. He has designed many competitive sailing yachts, among which the first IMOCA with foils of the Vendée Globe 2016. Figure 14 - View of STRAIGHT, CLOUT and DALI daggerboards ACKNOWLEDGEMENTS This CFD study on foils on IMOCA was among the first of a long series. CFD computations were performed by HydrOcean for this specific study. Further work, computations and analysis were performed by Guillaume Verdier together with his collaborators (Leonard Imas, Romain Garo, Robert Kleinschmit, Véronique Soulé, Benjamin Muyl and Romaric Neyhousser) and VPLP design (Daniele Capua, Quentin Lucet, Xavier Guisnel, Simon Watin). Their common work ended up with the present Dali foils, quite different from CLOUT studied here. AUTHORS BIOGRAPHY L. Mazas holds the current position of Hydrodynamic & Project Engineer at HydrOcean. He is in charge of CFD studies on sailing yachts. His previous experience includes hydrodynamic studies on maxi-trimarans, foiling catamarans and IMOCA monohulls. Y. Andrillon holds the current position of Sail Designer at North Sails France (Vannes). He previously worked at HydrOcean as responsible for CFD studies on sailing yachts. His previous experience includes hydro and aerodynamic studies on IMOCA and multihulls 172