Tacking Simulation of Sailing Yachts with New Model of Aerodynamic

Force Variation During Tacking Maneuver

Yutaka Masuyama (M), Department of Mechanical Engineering, Kanazawa Institute of Technology, Japan
Toichi Fukasawa (M), Department of Marine System Engineering, Osaka Prefecture University, Japan
A mathematical model for the tacking maneuver of a sailing yacht is presented as an extension of research by the
same authors. The authors have proposed the equations of motion for the tacking maneuver expressed in the
horizontal body axis system. The calculation method was applied to a 34-foot sailing cruiser and the simulated result
showed good agreement with the measured data from full-scale tests; however, the modeling of aerodynamic force
variation during tacking was insufficient due to lack of information about the sail forces. In this report, the authors
performed full-scale measurement of sail forces during tacking maneuvers using a sail dynamometer boat Fujin. The
Fujin is a 34-foot sailing cruiser which has a measurement system to obtain simultaneously sail forces, sail shapes,
and boat attitude. Based on the results of full-scale measurements, a new model of aerodynamic force variation for
the tacking maneuver was proposed. The equations of motion were also simplified to more easily perform the
numerical simulation. Using this calculation method, the tacking simulations were performed and compared with the
measured data from three full-scale boats. The simulated results showed good agreement with the measured data.
This simulation method provides an effective means for assessment of tacking performance of general sailing yachts.

KEYWORDS: aerodynamics, hydrodynamics, maneuvering, motions, performance assessment, performance prediction, sails.
NOMENCLATURE
B breadth at design waterline
D design draft (including fin keel)
Fn Froude number
GM metacentric height of boat
Ixx, yy, zz moments of inertia of boat about xb-, yb- and zb-axis in general body axis system
Jxx, yy, zz added moments of inertia about xb-, yb- and zb-axis in general body axis system
K, N moments about x- and z-axis in horizontal body axis system
L length on design waterline
lR distance between quarter-chord point of rudder and C.G. of boat
m mass of boat
mx, y, z added masses of boat along xb-, yb- and zb-axis in general body axis system
SA sail area (actual total sail area)
U, V velocity components of boat along x- and y-axis in horizontal body axis system
UA apparent wind speed (AWS)
VB boat velocity
xCG x-coordinate of C.G. of boat
xGCE x-coordinate of geometric centre of effort of sail
X, Y force components along x- and y-axis in horizontal body axis system
zGCE z-coordinate of geometric centre of effort of sail
αR effective attack angle of rudder
β leeway angle
γA apparent wind angle (AWA)
γR decreasing ratio of inflow angle for rudder
δ rudder angle
ρ density of water
ρa density of air
φ heel angle or roll angle
ψ heading angle

Coordinate system
G – x, y, z coordinate of horizontal body axis system

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 G – xb, yb, zb coordinate of general body axis system
od – xd, yd, zd coordinate of sail dynamometer axis system

INTRODUCTION
Tacking of a sailing yacht is a quick maneuvering motion accompanied by large rolling angle changes in a short period of time. To
analyze this type of large amplitude motion, a mathematical model for the simulation was presented by the same authors (Masuyama et.
al. 1993; 1995). In these previous reports the authors employed equations of motion expressed by the horizontal body axis system
introduced by Hamamoto et. al. (1988; 1992). In this coordinate system, the maneuvering motion of the boat and aero/hydro-dynamic
forces acting on it can be expressed easily. Both added mass and added moment of inertia, which are referenced to the body axes fixed
on the boat, can be obtained using the coordinate transformation. The calculation method was applied to a 34-foot sailing cruiser and a
100-foot Japanese traditional tall ship Naniwa-maru (Masuyama et. al. 2003). The simulated results showed good agreement with the
measured data obtained during full-scale tests. Keuning et al. (2005) extended our model to be able to apply the calculation method to a
large variety of yachts using their extensive database of Delft Systematic Yacht Hull Series (DSYHS). In this previous research,
however, the modelling of aerodynamic force variation during tacking was insufficient due to a lack of information about the sail
forces.

To clarify the sail force variation during tacking, the authors (2008b) previously measured the forces using a sail dynamometer boat
Fujin and proposed a new model of aerodynamic force variation. In this paper, which is an extension of the previous report, the
equations of motion are simplified to more easily execute the tacking simulation, and the numerical simulations are performed and
compared with the measured data for three full-scale boats, Fujin, Fair V, and Sea Dragon.

DIRECT MEASUREMENT OF SAIL FORCE USING FUJIN

Sail Dynamometer Boat Fujin

The sail performance was measured using the sail dynamometer boat Fujin. The Fujin was originally built for conducting tests on sails
by the authors for the Japanese America’s Cup entry in 1994. The design of the Fujin is based on the YR-10.3m class, which is an
International Measurement System (IMS) ocean racer designed by Yamaha Motor Co. Ltd. In this boat, load cells and CCD cameras
were installed to simultaneously measure the sail forces and shapes. At the same time, the sailing conditions of the boat (e.g., boat
speed, heel angle, wind speed, and wind angle) are measured. Figure 1 shows the sail plan and general arrangement of the
dynamometer frame of the Fujin as well as the coordinate system of the sail dynamometer. It should be noted that the origin of the sail
dynamometer system is not on the Center of Gravity (C.G.) of the boat but located at the aft face of the mast (xd-direction) and the
height of the deck level (zd-direction). The measurement system installed in the Fujin and the relationships between sail shapes and
performance were reported by Masuyama and Fukasawa (1997a; 1997b). Recently, the measured sail shapes were reanalyzed and
tabulated for use as input data for the numerical calculation. This sail shape database, in association with the results of latest RANS-
based CFD, was also reported by Masuyama et al. (2007; 2009).

Steady Sail Performance for Upwind Condition

The aerodynamic coefficients and the coordinates of the center of effort of the sails for the close-hauled condition are defined as:

 X Sd  YSd
X Sd  , YSd  ,
1 1
 aU A 2 S A  aU A 2 S A
2 2 (1)
N K
xCE  Sd , zCE   Sd ,
YSd YSd

where XSd and YSd are the force components along the xd and yd axes of the sail dynamometer system respectively, and KSd and NSd are
the moments around the xd and zd axes. xCE and zCE are the xd and zd coordinates of the center of effort of the sails (CE). It should be
noted that the xd and yd axes are fixed on the boat and therefore inclined with pitch and heel angles. This means the measured side force
YSd is not in the horizontal plane but is normal to the mast.

Figure 2 shows the measured aerodynamic coefficients and the coordinates of CE as a function of apparent wind angle (γA) for the
mainsail and 130% jib configuration shown in Figure 1. The aerodynamic forces acting on the mast and rigging are included in the
measured XSd and YSd forces. The solid symbols indicate the results of starboard tack and the open symbols indicate the port tack. In
Figure 2(a) the side force coefficient Y'Sd is expressed as positive for both port and starboard tacks. The measurements vary widely,
especially in the Y'Sd data, because they are based on measurements taken with the sails trimmed in different ways.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 Figure 1: Sail plan and arrangement of sail dynamometer frame of Fujin.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 Figure 2: Sail performance variation for mainsail and 130% jib configuration as a function of apparent wind angle (γA).

The apparent wind speed (UA) and apparent wind angle (γA) were measured by an anemometer attached to the bow unit. The bow unit,
which was attached to the bow pulpit, includes a post that can rotate freely to maintain its vertical attitude when the boat heels in order
to measure the wind data in the horizontal plane. Since the measured side force YSd acts normal to the mast, the values of YSd do not
vary with heel angle like the horizontal plane component, but are affected by decreases in both the effective attack angle and the
dynamic pressure on the sails as shown in Appendix 2. This decreasing ratio is approximated by a function of cos φ. However, the heel
angles during measurement were less than 20 degrees, and therefore the difference with the upright condition was less than 6%. For this
reason, the measured data in Figure 2 are indicated without heel angle correction. Under this assumption, the basic sail force curves
used in the tacking simulation in the upright condition, X'S0 and Y'S0, are shown as solid and dotted curves, which refer to the present
results and as well as other wind tunnel test results (Masuyama and Tatano 1982).

Figure 2(b) shows the variation of the CE coordinates of the sails. The xd- and zd-coordinates of the geometric center of effort (xGCE and
zGCE) are 0.63 m aft and 4.80 m above the origin of sail dynamometer system, which are indicated by dashed lines in the figure. The
geometric center of effort is calculated as the center of actual sail area of the mainsail and jib (mast area is not included). The measured
coordinates of xCE are near xGCE and move forward with increasing Apparent Wind Angle (AWA) toward the origin. However, the
coordinate of xCE coincides with the xGCE for the closest angle to the wind. In the experiments with the Fujin using other sail
configurations, such as the mainsail with a 75% jib and mainsail alone, the coordinates of xCE also coincided with the xGCE of each sail
plan in the upwind condition. Therefore, for the tacking simulation, the coordinate of xGCE is used as the point of application of YSd
force for various sail configurations.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
There is a wide scatter in the experimental values for zCE. This is thought to be because the measured heel moment contains a large
component from the mass of the dynamometer frame and rigging (659kg). This moment should be subtracted from the measurement,
taking into account the measured heel angle. If there is a slight error in the position of the center of gravity of the dynamometer frame,
or in the measured heel angle, the error in the calculated moment will be large. Some bias in the values of zCE between port and
starboard tacks might be caused by the slight discrepancy in alignment between the xd axis of the dynamometer frame and the center
line of the hull. However, although the measured data are scattered, the coordinates of zCE coincides with zGCE for the tacking
simulation. Using these results, the heel moment KSd and yaw moment NSd for the tacking simulation are obtained by multiplying the
YSd force by the coordinates of zGGCE and xGGCE, respectively, where the superscript of G means the coordinate from the center of gravity
of the boat.

Sail Force Variation During Tacking

Figure 3 shows two examples of the measured data in the time domain for X'Sd, Y'Sd, K'Sd and N'Sd during tacking for 20 seconds,
beginning five seconds before the initiation of tacking and ending 15 seconds after. Figure 3(a) shows tacking from starboard to port
tack, and Figure 3(b) shows tacking from port to starboard tack. The scatter in the data at the crossing points of the curves is caused by
the crew action on the dynamometer frame in releasing and trimming the jib sheet. In the measured data, the inertia forces and moments
due to the mass of the dynamometer frame are included. These effects appear clearly at the beginning and end of the tacking maneuver
but are not as significant during the middle stage. Hence the measured data are modified only by subtracting the forces and moments
due to the gravitational force acting on the dynamometer frame based on the measured heel angle at every moment. The evaluation of
the effect of the inertia forces and moments on the measured data is discussed in Appendix 1.

Figure 4 shows the variation of sail force coefficients during tacking as a function of the heading angle of the boat ψ, where ψ = 0º
means heading in the true wind direction. During tacking, the jib sheet was released just before the jib was backwinded on the new tack
in order to minimize luffing of the jib and loss of wind power. The curves show the results of 10 tacking cases from starboard to port
tack. It should be noted again that forces and moments are shown using the sail dynamometer coordinate system. The variations start
from the close-hauled condition of starboard tack until the boat is on port tack (i.e., from ψ = -45º to 45º). The corresponding AWA,
from γA= 30º to -30º, are also indicated in the second abscissa in the figure. Figure 4(a) shows the variation of X'Sd. When the boat heads
directly into the wind, X'Sd becomes about -0.1, (i.e., drag force coefficient). Figures 4(b) to 4(d) show the forces and moments become
zero not at ψ=0º, but around ψ=10º, which indicates a delay in the variation of forces and moments compared to the change of heading
angle. This could be caused by the sail filling with wind due to the yawing motion from the former tack to ψ =10º on the new tack when
the jib sheet was released.

Figure 3: Examples of measured sail force coefficients in the time domain during tacking.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
Figure 4: Variation of sail force coefficients during tacking as a function of heading angle of boat (tacking from starboard to port tack).

Figure 5 shows the same variation for the case of port tack to starboard tack. In this case, Y'Sd, K'Sd and N'Sd become zero at around ψ= -
10º, and the variation of forces and moments are almost symmetrical to Figure 4. Therefore, the bias in the zero crossing point of the
forces and moments in the tacking maneuver is symmetrical.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
Figure 5: Variation of sail force coefficients during tacking as a function of heading angle of boat (tacking from port to starboard tack).

Model of Sail Force Variation for Tacking Simulation

The equations of motion for tacking simulation are expressed in the horizontal body axis system, which is discussed in the next section.
In this axis system, the sail forces used in the calculation are expressed in horizontal components, and the moments are expressed
around the C.G. of the boat, rather than the origin of the sail dynamometer system. These components are defined as XS, YS, KS and NS
as:

1
X S  X 'S 0 cos 2    aU A 2 S A
2
1
YS  Y 'S 0 cos 2    aU A S A
2

2
 zG  1 3 (2)
K S   Y 'S 0  GCE  cos    aU A S A 2
2
 S  2
 A
  xG   zGCE
G
 
 sin   cos 2   1  aU A 2 S A 2
3
N S  Y 'S 0  GCE   X 'S 0 
  S   S   2
 A  A 

The derivation of these formulas, including effect of heel angle, φ, are described in Appendix 2.

The basic sail performance curves of X's0 and Y's0 in Figure 2 show the steady state values and do not express the dynamic variation due
to tacking. Therefore, the model of sail force variation for the tacking simulation is defined as bold lines in Figure 6, referring to the
measured data in Figures 4 and 5. Figure 6(a) shows the case of tacking from starboard to port tack. The abscissa indicates apparent
wind angle (γA). In the model, the basic sail performance curves of X's0 and Y's0 are divided into three stages. Stage A is the range of γA
that is greater than 20º. In this region, the coefficients vary with γA according to the basic curves. Stage B is the range of γA= 20º to -10º.
In this region, the coefficients are assumed to vary linearly along the lines determined from the results of Figures 4(a) and 4(b). Stage

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
C is the range of γA= -10º to -30º. In this region, the basic pattern of the coefficients is expressed as basic performance curves.
However, it may take several seconds to recover to the basic curves due to the delay in trimming the sails for the new tack. Therefore,
the coefficients are assumed to increase from the lowest values to the basic curve values with time. The recovery time was chosen from
5 to 10 seconds by taking the simulated heel angle corresponding to the measured heel angle. Figure 6(b) shows the case of tacking
from port to starboard tack, where the variation pattern proceeds in the opposite direction.

In the numerical calculation, the rolling effects on the sail were considered as a variation of apparent wind angle, γA, and speed, UA, due
to the motion at the center of effort of the sail. The sail forces and moments expressed in equation (2) are used for the equations of
motion in the following section.

X' S0 ,Y' S0
2.0
Starboard tack Port tack
A B 1.5 C Y' S0
Increasing with
elapsed time
1.0
X' S0
X' S0
0.5

0.0
50 40 30 20 10 0 -10 -20 -30 -40 -50
-0.5
γA [deg]

-1.0
Y' S0
-1.5
ψ=-45 ° ψ=45 °
-2.0
6a: Tacking from starboard to port tack

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 X' S0 ,Y' S0
2.0
Starboard tack Port tack
C 1.5 B A Y' S0

1.0
X' S0
X' S0
0.5

0.0
50 40 30 20 10 0 -10 -20 -30 -40 -50
Increasing with -0.5
γA [deg]
elapsed time

-1.0
Y' S0
-1.5
ψ=-45 ° ψ=45 °
-2.0
6b: Tacking from port to starboard tack
Figure 6: Model of sail force variation for tacking simulation.

EQUATIONS OF MOTION FOR TACKING SIMULATION

To express the large amplitude motion such as a tacking maneuver of a sailing yacht, we used motion equations expressed in the
horizontal body axis system introduced by Hamamoto et al. (1988; 1992). It should be noted that this coordinate system is different
than the system for the sail dynamometer. The origin of the coordinate system is on the C.G. of the boat as shown in Figure 7. The x-
axis lies along the centerline of the boat on the still-water plane and is positive forward. The y-axis is positive to starboard in the still
water plane, and the z-axis is positive downwards. In this coordinate system, the maneuvering motion of the boat and aero/hydro-
dynamic forces acting on it can be expressed in the horizontal plane even though the boat heels. Both added mass and added moment of
inertia, which are referenced to the body axes fixed to the boat, can be obtained by a coordinate transformation.

The authors previously presented the equations of motion expressed in four simultaneous differential equations but excluded pitching
and heaving motions (Masuyama et al. 1993; 1995). The distance between center of gravity and center of added mass of the boat, xG,
was taken into account. For a sailing yacht with a fin keel, the value of xG is small because the fin keel has a large mass and lateral
projected area, which affect both the centers of gravity and added mass. Therefore, to create an easier simulation procedure, it is
assumed that the center of added mass of the hull coincides with the center of gravity of the boat (i.e., xG =0). In this case, the forces
and moments expressed in the general body axis system are formulated from the Euler-Lamb equations of motion (Lamb, 1932). These
equations are transformed into those expressed in the horizontal body axis system, assuming the pitch angle, θ, is equal to zero, as
shown in the Appendix 3. The equations of motion expressed in the horizontal body axis system for the motions of surge, sway, roll
and yaw are derived as follows. The left sides are formulas (A-23) in Appendix 3 and the right sides are fluid dynamic forces acting on
the hull and sail with reference to the horizontal body axes.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 Figure 7: Definition of coordinate system and forces and moments, (+)-ve is indicated direction.

surge:
(m  mx ) U  (m  m y cos 2   mz sin 2  ) V (3)
 X 0  X H  X V V  X R  X S
sway:
(m  m y cos 2   m z sin 2  ) V  (m  m x ) U  2(m z  m y ) sin  cos   V (4)
 YH  Y   Y  YR  YS
roll:
( I xx  J xx )  ( I yy  J yy )  ( I zz  J zz )sin  cos   2 (5)
 K H  K    K R  K S  mgGM sin 
yaw:
 (I yy 
 J yy ) sin 2   ( I zz  J zz ) cos 2    2( I yy  J yy )  ( I zz  J zz )sin  cos    (6)
 N H  N  N R  N S

These equations are the same as equations (1) through (4) in reference of Masuyama et al. (1995), except the term xG was eliminated. In
equation (3), X0 is the hull resistance in the upright condition, which is calculated from model tests or the DSYHS database (Keuning
and Sonnenberg 1999). The previous traditional expressions (Masuyama et al., 1995) are adopted to describe the hydrodynamic forces
acting on the hull. The steady forces on the canoe body and fin keel are described using hydrodynamic derivatives as follows:

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 1
 V 2  X 
X H  ( X VV   2  X VVVV
 V 4 ) (  VB2 L D)
2
1
YH  (YV V   Y   YV V  2  YVV  V 2  YVVV
 V 3 ) (  VB2 L D)
2
(7)
1
K H  ( KV V   K   KV V  2  KVV   V 2  KVVV
 V 3 ) (  VB2 L D 2 )
2
1
N H  ( NV V   N    NV V  2  NVV   V 2  NVVV
 V 3 ) (  VB2 L2 D)
2
where, V' is defined as:

VB sin 
V     sin 
VB

Equations (7), in which higher order terms are eliminated, are a simplification of equations (5) in reference of Masuyama et al. (1995).

The hydrodynamic forces and moments on the rudder are expressed as:

1
X R  C X sin  R sin  (  VB2 L D)
2
1
YR  CY sin  R cos  cos  (  VB2 L D)
2 (8)
1
K R  C K sin  R cos  (  VB2 L D 2 )
2
1
N R  C N sin  R cos  cos  (  VB2 L2 D)
2
where CXδ to CNδ are coefficients determined by rudder angle tests. The effective attack angle of the rudder, αR, is given by

 l R 
 R     R    tan 1   (9)
 U 
where γR is the decreasing ratio of inflow angle, which is caused mainly by the downwash from the fin keel. The third term indicates
the inflow angle due to the turning motion of the boat, where the lR is the horizontal distance between the quarter-chord of rudder and
the C.G. of the boat. For tacking maneuvers, the turning radius is relatively small and hence the inflow angle at the rudder position
becomes greater than 30 degrees. This means that the rudder is outside of the downwash of the fin keel and that a decrease in inflow
angle might not occur. Therefore γR is not multiplied here by the third term of equation (9).

HYDRODYNAMIC FORCES ACTING ON HULL

Principal Dimensions of Boats for Tacking Simulation
Tacking simulation was performed for three boats, Fujin, Fair V and Sea Dragon. The Fair V is a 10.4 m LOA sailing cruiser designed
by Masuyama. The principal dimensions of the Fair V and a comparison of its tacking performance between measured and simulated
tests have been previously described (Masuyama et al. 1993; 1995). However, the sail plan and shape of the fin keel were modified
after the previous experiments. Therefore, the measurement of tacking maneuvring was repeated by the authors. The Sea Dragon is a
6.4 m LOA small sailing cruiser built in Japan. Using this boat, the authors also conducted many rolling and tacking tests (Masuyama
et al. 2008a). Figure 8 shows the sail plan and principal dimensions of Fair V and Sea Dragon.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 8a: Fair V 8b: Sea Dragon
Figure 8: Sail plans of Fair V and Sea Dragon.

Hydrodynamic Derivatives and Coefficients

The hydrodynamic derivatives of the hull and rudder were obtained from model tests. In order to apply the derivatives for Equations (7)
and (8), oblique towing tests and rudder angle tests were performed. The analysis procedure from the measured data to the
hydrodynamic derivatives is described in a previous paper (Masuyama et al. 1995). The results of the rudder angle tests and oblique
towing tests for the Fujin are shown in Figures 9 and 10, respectively. Because of the restrictions of the experimental facility, heel
moment, K, was not measured. Hence, the K moment around the LWL level was evaluated by multiplying the Y force by 43% of the fin
keel draft according to Keuning et al. (2003). Table 1 shows the hydrodynamic derivatives of the hull and coefficients of the rudder that
were obtained. The calculated results from equations (7) and (8) using these hydrodynamic derivatives and coefficients are indicated as
curves. It should be noted that the N moment is expressed around midships. The decreasing ratio of inflow angle, γR, was also measured
by rudder angle tests under oblique towing conditions for the Fujin, modified Fair V, and Sea Dragon. The results showed agreement
with the data of the original Fair V (Masuyama et al. 1995).

XR' YR'
0.01 0.05

δ
-40 0 40
[d e g ]
δ
-40 40
[d e g ]

-0.02 -0.05

NR'
0.04

○ β= 0°
△ β= 5°
δ
-40 0 40 □ β= 10 °
[d e g ]
◎ β= 15 °

-0.04

Figure 9: Variation of hydrodynamic coefficients of rudder with rudder angle δ for Fujin.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 XH' YH'
0.04 0.1

β β
-20 0 20 -20 20
[d e g ] [d e g ]

-0.04 -0.1

NH'
0.04

○ φ= 0°
△ φ= -10 °
β
-20 0 20 □ φ= -20 °
[d e g ]
◎ φ= -30 °

-0.04

Figure 10: Variation of hydrodynamic coefficients of hull with leeway angle β for Fujin (without rudder forces).

Hydrodynamic derivatives of the hull due to yawing motion, such as X V , Y , and N , are calculated using the following equations
from Masuyama et al. (1995):

1
X V  my Cm  1 L2 D (10)
2

where Cm is assumed to be 0.3 by Hasegawa (1980).

  2d m  1 1
Y    VB L2 D  2a0 h  h0  VB AF c (11)
4 L 2 2
2d  2d 1 1
N  m  m  0.54  VB L3 D  2a0 h  h VB AF c
L  L

2
 (12)
2 2
where,

h  h0  hwb  2
In equations (11) and (12), the first term is the contribution from the canoe body, which is expressed by an empirical formula (Inoue et
al. 1981). The second term is the contribution from the fin keel, which is calculated by lifting surface theory (Etkin 1972). Here, AF,
a0 , and c are the lateral projected area, lift curve slope, and mean chord length of the fin keel, respectively. The terms h, h0, and hwb
are the distances from the C.G. of the boat to the leading edge of the fin keel, the axis of rotation for vanishing lift force, and the
aerodynamic center of the fin keel, expressed by the percentage of the mean chord length, c . The contribution from the rudder is
already considered in the third term of equation (9).

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
The derivative due to rolling, Y , was calculated by considering the change in attack angle of both the fin keel and the rudder caused
by the rolling angular velocity 
 as:

1 D
Y   VB a0  c z dz (13)
2 0

where c is chord length of the fin keel or rudder at z coordinate. The contribution of the canoe body was neglected.

The damping coefficient for rolling , K  , was obtained by a rolling test of the full-scale boat. A logarithmic decrement, σ, and
coefficient, α0, are defined as:

  t   2
  ln   , 0  (14)
 t  T 2 T

where φ is roll angle and T is roll period.

From the rolling test of Fair V and Sea Dragon with mainsail, the value of α0 was evaluated as 0.42 and 0.53, respectively (Masuyama
et al. 1995 and 2008a). The damping effect of the sails was very large, so these tests were also performed under running conditions to
clarify the effect of velocity without sails. In this case the value of α increased linearly with the boat velocity. From these results, the
value of α for these boats is formulated using Froude number as:

   0  0.4Fn (15)

Using the coefficient α, the damping coefficient for rolling, K  , is given by

K  2 I xx  J xx  (16)

where the value of Ixx +Jxx are also obtained from the rolling test using following relation:

2
T 
I xx  J xx  mg  GM   (17)
 2 
The hydrodynamic derivatives and coefficients of these boats are shown in Table 1.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 Table 1: Principal dimensions, hydrodynamic derivatives, added masses and added moments of inertia of
Fujin, Fair V and Sea Dragon.

Boat Name Fujin Fair V Sea Dragon

Class YR 10.3 KIT 34 ―
LOA [m] 10.35 10.39 6.45
LWL (L) [m] 8.80 8.55 5.30
BMAX [m] 3.37 3.04 2.48
BWL (B) [m] 2.64 2.42 1.87
Draft (Canoe body) [m] 0.44 0.41 0.28
Draft (Fin keel) (D) [m] 2.02 1.96 1.30
Displacement* [kg] 4410 3780 1370
GM [m] 1.45 1.31 0.93
xCG (from Midship) [m] －0.60 －0.10 0.0
lR (from C.G.) [m] 4.40 4.22 3.06
2
Sail area(Mainsail) [m ] 33.2 30.2 12.6
2
Sail area(Jib) [m ] 26.1 29.5 11.6
G
x GCE (from C.G.) [m] 0.67 0.87 0.39
G
z GCE (from C.G.) [m] －6.13 －6.20 －4.15
mx [kg] 160 140 50
My (Hull, Sail) [kg] 2130, 280 2410, 280 740, 120
mz [kg] 12000 10400 3600
2
Ixx [kg･m ] 17700 12500 1400
2
Iyy [kg･m ] 33100 22600 2600
2
Izz [kg･m ] 17200 11300 1300
2
Jxx (Hull, Sail) [kg･m ] 7200, 8100 8600, 8100 900, 1200
2
Jyy [kg･m ] 42400 34600 4300
2
Jzz [kg･m ] 6700 7600 1000
1 1 1
X’VV 3.38×10－ 3.37×10－ 2.66×10－
3 4 3
X’φφ 1.40×10－ －9.83×10－ －7.16×10－
X’VVVV －1.84 －1.88 －1.21
2 2 2
X’Vψ． －1.91×10－ －2.35×10－ －2.82×10－
1 1 1
Y’V －5.35×10－ －5.98×10－ －5.85×10－
3 3 3
Y’φ －5.89×10－ 7.39×10－ －8.94×10－
－1 －1
Y’Vφφ 7.37×10 6.04×10 1.25
1 1 1
Y’VVφ －5.53×10－ －3.94×10－ －3.94×10－
Y’VVV 3.07 3.20 1.95
1 1 1
Y’φ． 2.19×10－ 2.42×10－ 2.77×10－
3 2 2
Y’ψ． －4.01×10－ 1.53×10－ －1.88×10－
－1 －1 1
K’V 2.80×10 2.99×10 2.82×10－
3 3 3
K’φ 3.36×10－ －3.60×10－ 4.41×10－
－1 －1 －1
K’Vφφ －4.07×10 －3.03×10 －5.86×10
1 2 1
K’VVφ 2.24×10－ 9.45×10－ 1.39×10－
K’VVV －1.38 －1.30 －1.04
1 1 1
K’φ． －3.53×10－ －3.57×10－ －3.69×10－
2 2 2
N’V －3.23×10－ －2.73×10－ －4.82×10－
2 2 2
N’φ －1.52×10－ －1.49×10－ －1.45×10－
－4 －3 －2
N’Vφφ 2.71×10 8.53×10 4.13×10
2 2 2
N’VVφ －9.06×10－ －6.14×10－ －8.97×10－
－2 －2 －2
N’VVV －2.98×10 1.87×10 －9.02×10
3 3 3
N’ψ． －5.89×10－ －5.21×10－ －6.43×10－
2 2 2
CXδ －3.79×10－ －6.21×10－ －4.83×10－
1 1 1
CYδ －1.80×10－ －1.78×10－ －1.52×10－
－2 －2 －2
CKδ 9.76×10 9.32×10 9.32×10
2 2 1
CNδ 9.74×10－ 9.10×10－ 1.01×10－
* Including crew weight

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
Added Masses and Added Moment of Inertias
Added mass of the hull along x-axis, mx, was assumed to be the same as the value of a spheroid (Newman 1977). Since the frequency of
swaying and yawing motion is low, my and Jzz are calculated with the double model expressed by the Lewis form coefficient, C1 and C3,
as:


my    D 2 C y x  dx (18)
2 L


J zz    x 2 D 2 C y x  dx (19)
2 L

where,

2
(1  C1 ) 2  3C 3
C y x  
(1  C1  C 3 ) 2

Values of mz and Jyy are also calculated as:


mz    B 2 C z x  dx (20)
8 L


J yy    x 2 B 2 C z x  dx (21)
8 L

where,

2
(1  C1 ) 2  3C3
C z x  
(1  C1  C3 ) 2

The value of Jxx, including mainsail, was obtained from the rolling test with mainsail using Equation (17). The added masses of the fin
keel, rudder and sail along the y-axis were calculated assuming they were ellipsoid planes with the same lateral areas. The added
moment of inertias of the fin keel and rudder around the z-axis were evaluated as:

J zz k ,r   my k ,r   x2C .G.k ,r  (22)

where xC .G.k ,r  means the distance from the C.G. of the boat to the center of lateral area of the fin keel or rudder. Added masses
and moments of inertias of the fin keel and rudder were included in the values of my and Jzz of the hull. The values of these added
masses and moments are also shown in Table 1 with inertia forces and moments of these boats.

COMPARISON BETWEEN MEASURED AND SIMULATED RESULTS

The Runge-Kutta method was employed to calculate the equations of motion. The rolling and yawing motions were calculated around
the C.G. of the boat. Input data for the simulation consisted of the true wind velocity and the measured time history of the rudder angle
during the tacking maneuver at increments of 0.1 seconds.

Results of Fujin
Figure 11 shows the comparison between measured and simulated results of Fujin. Figure 11(1) shows tacking from starboard to port
tack, and 11(2) shows tacking from port to starboard tack. The sail force variations in Figures 3(a) and 3(b) correspond to these cases,
respectively. The indicated results were recorded for 35 seconds, beginning 5 seconds before the start of tacking. Figure 11(1)(a) shows
the boat trajectories. Solid circles indicate the positions of measured C.G. of the boat at each second, while open circles indicate the

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
simulated positions. The illustrations of the small boat symbol indicate the heading angle ψ every three seconds. The wind blows from
the right side of the figure and the grid spacing is taken as 15 meters. Figure 11(1)(b) shows the time histories of rudder angle δ,
heading angle ψ, heel angle φ, and boat velocity VB. The solid lines are measured data and the dotted lines are simulated data.

In Figures 11(1)(b) and 11(2)(b), the patterns of rudder angle variation can be considered as standard for tacking maneuvers. As shown,
tacking with a yawing motion of 90 degrees is completed in 7 to 8 seconds. The boat velocity decreases about 30%, and the boat takes
about 15 seconds to recover to the previous velocity after the yawing motion is completed. The measured time histories of ψ and φ
indicate the delay of the zero crossing point of φ compared with ψ. This might be caused by the sail filling with wind due to the yawing
motion until around ψ= 10º on the opposite tack as shown in Figures 4 and 5. The simulated time histories show a slight delay when
compared to the measured data. In particular, the delay of the simulated heel angle is relatively large. This might be caused by the over-
K
estimation of the damping coefficient for rolling,  . For this point further investigation might be necessary. However, the simulated
results of velocity decrement show agreement with the measured results. This suggests that the model of sail force variation proposed
in this report is adequate for the tacking simulation. In Figures 11(1)(a) and 11(2)(a), although the simulated trajectories show slightly
larger turning radiuses than the measured trajectories, the simulated results show agreement with the measured values overall.

Results of Fair V
Figure 12 shows the comparison between measured and simulated results of Fair V. The contents of these figures are identical to
Figure 11. In these cases, the rudder angle variations in the first stage are relatively small. These cause a delay in the yawing motion of
the boat. Hence it takes more than 10 seconds to complete the tacking maneuver. On the other hand, the simulated results show a
prompt response to the rudder angle variation. Therefore the simulated time histories vary slightly earlier compared with the measured
histories. By the same reasoning, the simulated trajectories in Figures 12(1)(a) and 12(2)(a) show smaller turning radiuses than the
measured trajectories.

Results of Sea Dragon

Figure 13 shows the comparison between measured and simulated results of Sea Dragon. This boat has a relatively small fin keel and
old, inefficient sails. The closest measured angle to the wind is therefore large and the tacking angle becomes about 100 degrees. The
simulated trajectories show agreement with the measured trajectories, but there is a slightly smaller minimum angle to the wind than
the measured values due to the better sail coefficients used in the calculation. Unfortunately, the measured and simulated time histories
of the heel angle do not agree well because this boat is severely affected by crew movements on the deck due to the light displacement.
The patterns of rudder angle variation are almost identical to Figures 11(1)(b) and 11(2)(b). However, the simulated time histories of ψ
vary relatively earlier than the measured histories. This also might be caused by the light displacement of the boat.

Overall, although the timing of boat motion indicated in the simulated time histories shows a slight discrepancy, the tendency and
amount of variation of the boat motion indicate good agreement with the measured data, including the decrement of boat velocity.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 measured measured
simulated simulated

U T =5.7m/s s ta rt o f U T =5.4m/s
tac k in g

W IN D W IN D
sta rt o f W IN D W IN D
tac k in g

15 m 15 m

(a ) T ra je c to ry o f b o at (a ) T ra je c to ry o f b o at
sta rt o f ta ck in g s ta rt o f ta ck in g
[d e g ] [d e g ]
70 70
ψ: H ea d in g A n g le
ψ ψ φ: H ee l A n g le
δ: R u d d e r A n g le
35 35
φ φ V B : B o a t V e lo city
δ
δ δ
[m /s] [m /s]
φ 0 5 φ 0 5
δ
ψ 4 ψ 4
VB
3 VB 3
-35 VB -35 VB
2 2
1 1
-70 0 -70 0
-5 0 5 10 15 20 25 30 -5 0 5 10 15 20 25 30
ela p s e d tim e [s e c ] ela p s e d tim e [s e c ]
(b ) B o a t a ttitu de p a ram e te rs (b ) B o a t a ttitu de p a ram e te rs

11a: From starboard to port tack 11b: From port to starboard tack.
Figure 11: Measured and simulated results of tacking maneuver of Fujin.

measured measured
simulated simulated

U T =4.8m/s sta rt o f U T =4.9m/s

tac k in g
W IN D W IN D
sta rt o f W IN D W IN D
tac k in g

15 m 15 m

(a ) T ra je c to ry o f b o at (a ) T ra je c to ry o f b o at
sta rt o f ta ck in g sta rt o f ta ck in g
[d e g ] [d e g ]
70 70
ψ: H ea d in g A n g le
ψ φ: H ee l A n g le
ψ
δ: R u d d e r A n g le
35 35
φ V B : B o a t V e lo city
δ φ
δ δ
[m /s] [m /s]
φ 0 5 φ 0 5
4 δ 4
ψ ψ
VB 3 3
-35 VB -35 VB
VB 2
2
1 1
-70 0 -70 0
-5 0 5 10 15 20 25 30 -5 0 5 10 15 20 25 30
ela p se d tim e [se c ] ela p s e d tim e [s e c ]
(b ) B o a t a ttitu de p a ram e te rs (b ) B o a t a ttitu de p a ram e te rs

12a: From starboard to port tack. 12b: From port to starboard tack.
Figure 12: Measured and simulated results of tacking maneuver of Fair V.
U T =5 .4 m /s

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 measured measured
simulated simulated

U T =5.6m/s sta rt o f U T =5.5m/s

tac k in g

W IN D W IN D
sta rt o f
tac k in g W IN D W IN D

15 m 15 m

(a ) T ra je c to ry o f b o at (a ) T ra je c to ry o f b o at
sta rt o f ta ck in g sta rt o f ta ck in g
[d e g ] [d e g ]
70 70
ψ: H ea d in g A n g le
ψ ψ
φ: H ee l A n g le
δ: R u d d e r A n g le
35 δ 35 φ
φ V B : B o a t V e lo city
δ δ δ
[m /s] [m /s]
φ 0 5 φ 0 5

ψ 4 ψ φ 4
VB 3 VB 3
-35 VB -35 VB
2 2
ψ 1
ψ 1
-70 0 -70 0
-5 0 5 10 15 20 25 30 -5 0 5 10 15 20 25 30
ela p se d tim e [se c ] ela p s e d tim e [s e c ]
(b ) B o a t a ttitu de p a ram e te rs (b ) B o a t a ttitu de p a ram e te rs

13a: From starboard to port tack. 13b: From port to starboard tack.
Figure 13: Measured and simulated results of tacking maneuver of Sea Dragon.

CONCLUSIONS
In this report, the sail force variations during tacking were measured using a sail dynamometer boat Fujin. Based on the results a new
model of aerodynamic force variation for tacking maneuver was proposed. Then the equations of motion were simplified to more easily
execute the numerical simulation by eliminating the term xG. Using these equations of motion expressed in the horizontal body axis
system, tacking simulations were performed and compared with the measured data for three full-scale boats Fujin, Fair V, and Sea
Dragon.

Although the timing of boat motion indicated in the simulated time histories shows a slight discrepancy, the tendency and amount of
variation of the boat motion indicates good agreement with the measured data, particularly the simulated decrement of boat velocity.
This shows that the proposed model of sail force variation is adequate for the tacking simulation. This simulation method provides an
effective means for assessment of tacking performance of general sailing yachts.

ACKNOWLEDGEMENTS
The authors wish to thank Yamaha Motor Co. Ltd. for permitting the description of the principal dimensions and specifications of the
Fujin; Mr. David A. Helgerson for his comments on this article; Mr. H. Mitsui, the harbormaster of the Anamizu Bay Seminar House
of Kanazawa Institute of Technology, for his assistance with the sea trials; and the graduate and undergraduate students of the
Kanazawa Institute of Technology who helped with the sea trials, particularly H. Arakawa and T. Onishi, who performed the sea trials
of modified Fair V and Sea Dragon.

This research is partly supported by the Grant-in-Aid for Scientific Research of the Ministry of Education, Science, Sports and Culture
in Japan.

REFERENCES
CAMPBELL, IMC. (1997). “Optimisation of a sailing rig using wind tunnel data.” Proceedings of the 13th Chesapeake Sailing Yacht
Symposium, Annapolis, MD, 49-63.
ETKIN, B. (1972). Dynamics of Atmospheric Flight. John Wiley & Sons, New York.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
HAMAMOTO, M. and Akiyoshi, T. (1988). “Study on ship motions and capsizing in following seas (1st report).” Journal of the
Society of Naval Architects of Japan, 147, 173-180.
HAMAMOTO, M. (1992). “A new coordinate system and the equations describing maneuvering motion of a ship in waves.”
Proceedings of workshop on prediction of ship maneuverability, West-Japan Society of Naval Architects, Fukuoka, Japan, 61-76.
HASEGAWA, K. (1980). “On a performance criterion of autopilot navigation.” Journal of the Kansai Society of Naval Architects,178,
93-103.
INOUE, S., Hirano, M. and Kijima, K. (1981). “Hydrodynamic derivatives on ship maneuvering.” International Shipbuilding Progress,
28 (321), 112-125.
KERWIN, J. E. (1978). “A velocity prediction program for ocean racing yachts revised to February, 1978.” H. Irving Pratt Ocean Race
Handicapping Project, MIT Report No. 78-11.
KEUNING, J. A. and Sonnenberg, U. B. (1999). “Approximation of the calm water resistance on a sailing yacht based on the ‘delft
systematic yacht hull series.’” Proceedings of the 14th Chesapeake Sailing Yacht Symposium, Ananpolis, MD, 181-200.
KEUNING, J. A. and Vermeulen, K. J. (2003). “The yaw balance of sailing yachts upright and heeled.” Proceedings of the 16th
Chesapeake Sailing Yacht Symposium, Annapolis, MD, 1-17.
KEUNING, J. A., Vermeulen, K. J., and de Ridder, E. J. (2005). “A generic mathematical model for the maneuvering and tacking of a
sailing yacht.” Proceedings of the 17th Chesapeake Sailing Yacht Symposium, Annapolis, MD, 143-163.
LAMB, H. (1932). Hydrodynamics. Cambridge University Press, Cambridge, UK.
MASUYAMA, Y., and Tatano, H. (1982). “Hydrodynamic analysis on aailing (part 4: wind tunnel experiments on yacht sails).”
Journal of the Kansai Society of Naval Architects, 185, 107-115.
MASUYAMA, Y., Nakamura, I., Tatano, H., and Takagi, K. (1993). "Dynamic performance of sailing cruiser by full-scale sea tests."
Proceedings of the 11th Chesapeake Sailing Yacht Symposium, Annapolis, MD, 161-179.
MASUYAMA, Y., Fukasawa, T., and Sasagawa, H. (1995). “Tacking simulation of sailing yachts – numerical integration of equations
of motion and application of neural network technique.” Proceedings of the 12th Chesapeake Sailing Yacht Symposium, Annapolis,
MD, 117-131.
MASUYAMA, Y., and Fukasawa, T. (1997a). “Full scale measurement of sail force and the validation of numerical calculation
method.” Proceedings of the 13th Chesapeake Sailing Yacht Symposium, Annapolis, MD, 23-36.
MASUYAMA, Y., Fukasawa, T., and Kitasaki, T. (1997b). “Investigations on sail forces by full scale measurement and numerical
calculation (part 1: steady sailing performance).” Journal of the Society of Naval Architects of Japan, 181, 1-13.
MASUYAMA, Y., Nomoto, K., and Sakurai, A. (2003). “Numerical simulation of maneuvering of ‘Naniwa-maru,’ a full-scale
reconstruction of sailing trader of Japanese heritage.” Proceedings of the 16th Chesapeake Sailing Yacht Symposium, Annapolis,
MD, 174-181.
MASUYAMA, Y., Tahara, Y., Fukasawa, T. and Maeda, N. (2007). “Database of sail shapes vs. sail performance and validation of
numerical calculation for upwind condition.” Proceedings of the 18th Chesapeake Sailing Yacht Symposium, Annapolis, MD, 11-
31.
MASUYAMA, Y., Fukasawa, T., and Onishi, T. (2008a). “Dynamic stability and possibility of capsizing of small light sailing cruiser
due to wind.” Proceedings of the International Conference on Innovation in High Performance Sailing Yachts, Lorent, France, 57-
64.
MASUYAMA, Y., and Fukasawa, T. (2008b). “Tacking simulation of sailing yachts with new model of aerodynamic force variation.”
Proceedings of the Third High Performance Yacht Design Conference, Auckland, NZ, 138-147.
MASUYAMA, Y., Tahara, Y., Fukasawa, T., and Maeda, N. (2009). “Database of sail shapes versus sail performance and validation of
numerical calculation for the upwind condition.” Journal of Marine Science and Technology, 14(2), 137-160.
NEWMAN, J. N. (1977). Marine Hydrodynamics. MIT Press, Cambridge, MA.

APPENDIX 1
Cancellation of Inertia Forces and Moments due to Mass of the Dynamometer Frame

Inertia forces and moments due to mass of the dynamometer frame are cancelled with the following procedure. Using the finite
difference scheme, the angular velocity  k and angular acceleration 
 k are obtained from the time history of roll angle as:

k 1  k 1
k  (A-1)
2h
  2k  k 1
k  k 1 (A-2)
h2

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
where h is time step interval and φk is roll angle at k-th step.

In order to confirm this cancelling procedure, a rolling test of the Fujin was performed. Figure A-1 (a) shows the time histories of the
measured roll angle, and calculated angular velocity and angular acceleration by formulas (A-1) and (A-2). The measured time step
interval is 0.1 second (h=0.1 sec).

Figure A-1 (b) shows the time histories of the YSd forces. The result without cancellations is shown in solid line, the result with
cancellation of the gravity force alone is dotted line, and the result with cancellation of both gravity force and inertia force is the
alternating long and short dash line. Here, the inertia force Yinertia is calculated as:

Yinertia   rdm  z DG M  1310  N  (A-3)

where zDG is the distance between C.G. of the boat and C.G. of the dynamometer frame (1.98 m) and M is mass of the frame (659 kg).
Figure A-1 (c) shows the same time histories of the KSd moments. Since the KSd moment is expressed around the origin of the sail
dynamometer coordinate system, the inertia moment Kinertia is calculated as:

K inertia   r r  z 0 dm   r 2 dm   z 0  rdm  R 2 M  z 0 z DG M

(A-4)
 8740  N  m
where z0 is the distance between C.G. of the boat and the origin of sail dynamometer coordinate system (1.22 m), and R is the radius of
gyration of the dynamometer frame around xd axis (3.96 m). The cancelled results of YSd force and KSd moment show successful
elimination of the contribution of the mass of the dynamometer frame due to both the gravitational force and the inertial force on the
measured data. Therefore, this cancelling procedure is considered to be adequate.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 30
φ [deg]
20 φ' [deg/sec]
φ'' [deg/sec 2 ]

φ, φ', φ''
10

-10

-20

-30
0 10 20 30 40
ela p se d tim e [se c ]
(a) Rolling angle, angular velocity and angular acceleration

30 00
Y S d (without cancellation)
20 00 Y S d (with cancellation of
gravity force)
Y S d (with cancellation of
Y S d Fo rc e [N ]

10 00 garvity force and

inertia force)

-10 00

-20 00

-30 00
0 10 20 30 40
ela p se d tim e [se c ]
(b) Y S d Force

60 00
K S d (without cancellation)
40 00 K S d (with cancellation of
gravity force moment)
K S d M o m en t [N -m ]

K S d (with cancellation of
20 00 garvity force moment
and inertia moment)

-20 00

-40 00

-60 00
0 10 20 30 40
ela p se d tim e [se c ]
(c) K S d Moment

Figure A-1: Time histories of rolling test of Fujin showing cancelling procedure of inertia forces and
moments due to mass of dynamometer frame.

To determine the effect of inertial force on the data of a tacking maneuver, the same procedure was applied to the measured data in
Figure 3. Figure A-2 shows the results of cancellation on the YSd force and KSd moment. The solid lines are the same as those in Figure
3, which show the cancelled effect of gravity, and the dotted lines represent the results of cancelled inertial forces. The cancelled results
appear clearly at the starting and finishing stages of the tacking maneuver, but are not as significant at the middle stage. Therefore the
measured data are shown only subtracting the forces and moments due to the gravity force acting on the dynamometer frame.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 s ta rt o f ta ck in g s ta rt o f ta ck in g
1.5 1.5
Y 'Sd
1.0 1.0
K 'Sd
0.5 0.5

Y 'Sd, K 'Sd
Y 'Sd, K 'sd

0.0 0.0

-0.5 K 'Sd
C an c e le d g ra v ity -0.5 C an c e le d g ra v ity
fo rc e e ffe c t a lo n e fo rc e e ffe c t a lo n e
C an c e le d g ra v ity C an c e le d g ra v ity
-1.0 fo rc e a n d in e rtia -1.0 fo rc e a n d in e rtia
Y 'Sd fo rc e e ffe c t fo rc e e ffe c t
-1.5 -1.5
-5 0 5 10 15 -5 0 5 10 15
ela p s e d tim e [s e c ] ela p s e d tim e [s e c ]
(a ) Ta c k in g fro m s ta rb o ard tac k to p o rt ta ck (b ) T ac k in g fro m p o rt tac k to s ta rb oa rd ta c k

Figure A-2: Results of cancellation of inertia force on YSd and inertia moment on KSd from the measured time
histories during tacking.

APPENDIX 2
Variation of Sail Force Coefficient with Heel Angle

The equations of motion for tacking simulation are expressed in the horizontal body axis system. The sail forces and moments should
be expressed in their horizontal components.

In the upright condition, the aerodynamic coefficients X'S0 and Y'S0 are expressed using lift coefficient L'S0 and drag coefficient D'S0
as:

X S 0  LS 0 sin  A  DS 0 cos  A

(A-5)
YS0   LS 0 cos  A  DS 0 sin  A

where the subscript of 0 means value at the upright condition.

In the heeled condition, the effect of heel on the aerodynamic forces is produced by the reduction of both the apparent wind angle and
apparent wind speed as given by Kerwin (1978) and Campbell (1997). The apparent wind angle in the heeled condition γAφ is
expressed as follows using apparent wind angle γA and apparent wind speed UA:

 U A sin  A  cos 
 A  tan 1    tan 1 tan  A  cos  (A-6)
 U A cos A 
The apparent wind speed in the heeled condition UAφ is also expressed as:

U A  U A cos  A 2  U A sin  A  cos  2  U A 1  sin  A  sin  

2
(A-7)

For the close-hauled condition, the sail may not stall due to the small attack angle. Therefore, the lift force will decrease proportionally
to the reduction of both the apparent wind angle and the dynamic pressure of flow (i.e., the square of the apparent wind speed). Hence
the decreasing ratio of lift force by the heel angle φ can be described as:

 tan 1 tan  A  cos  

2
  A

 U A


    1  sin  A  sin  
2
  (A-8)
 A  U A   A 

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
The vector of lift force inclines with heel angle and rotates in the normal plane to the apparent wind axis. Since the angle between the
apparent wind axis and the boat center line (heeling axis) is γA, the rotating angle of the lift force vector φ’ in the normal plane to the
apparent wind axis is given by:

   sin 1 cos  A  sin   (A-9)

Therefore, the decreasing ratio of horizontal component of the lift force is expressed as:

 tan 1 tan  A  cos  

2
  A

 U A


 cos        
 1  sin  A  sin    cos sin cos  A  sin  
2 1
(A-10)
 A  U A    A 

Expanding equation (A-10) in a power series and assuming that γA is small, results in

2
  A  U A 
   cos     cos 2   sin 2   cos   cos   cos 3 
1 1
  (A-11)
 A  U A   2  2

Equation (A-11) can further be expanded in terms of φ, and results in

2
  A  U A 
   cos    1   2 (A-12)
 A  U A 
Equation (A-12) is incidentally equal to the first two terms of the power series for the cos2φ function.

Hence the curve of cos2φ was compared with the calculated results of equation (A-10) for three γA cases (Figure A-3). The calculated
results show agreement with the curve of cos2φ in spite of the large γA. Therefore, we adopted the formula of cos2φ to express the
decreasing ratio of the horizontal component of the lift force in place of equation (A-10).

1.0
Decreasing ratio of lift force

0.9

0.8

0.7
cos 2 φ
○ γA = 20 °by eq (A-10)
0.6 ■ γA = 40 °by eq (A-10)
△ γA = 60 °by eq (A-10)
0.5
0 10 20 30 40
Heel angle φ [deg]
Figure A-3: Comparison between the curve of cos2φ and the calculated results of equation (A-10).

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
Finally, when the lift coefficient represents the variation of the lift force including the contribution of dynamic pressure of apparent
wind speed, the horizontal component of lift coefficient in the heeled condition L'S is described as:

LS  LS 0 cos 2  (A-13)

The main part of the drag is caused by the induced drag, which is in proportion to the square of the lift force. The reduction of lift force
expressed by equation (A-8) is also approximated by cos φ. The vector of the drag force is in line with the apparent wind axis and does
not incline by the heel angle. Therefore the horizontal component of the drag coefficient D'S is described as:

DS  DS 0 cos 2  (A-14)

From these results, the aerodynamic coefficients in the horizontal components X'S and Y'S are then expressed as follows using the
coefficients at the upright condition L'S0 and D'S0:

X S  LS sin  A  DS cos  A  LS 0 cos 2  sin  A  DS 0 cos 2  cos  A
 X S 0 cos 2 
(A-15)
YS   LS cos  A  DS sin  A   LS 0 cos 2  cos  A  DS 0 cos 2  sin  A
 YS0 cos 2 

The moment KS is generated mainly by the YS force. However, KS is affected by the component normal to the mast. Hence,

 zG   G 
K S   YS  GCE  cos   YS0  zGCE  cos  (A-16)
 S   S 
 A   A 
where zGGCE is z-coordinate of the geometric center of effort of the sail from the C.G. of the boat and negative upwards.

The moment NS is also generated mainly by the YS force. However, it is well known that the NS is also affected by the heel angle φ due
to the application point of the thrust force XS moving outboard to lee side. Therefore N'S can be written, including the effect of X'S0, as:


  xGCE
G   zGCE
G
  2
N S  Y 'S 0    X 'S 0   sin  
 cos  (A-17)
  S   S  
  A   A  

where xGGCE is x-coordinate of the geometric center of effort of the sail from the C.G. of the boat.

APPENDIX 3
Transformation of Forces and Moments from General Body Axis System to Horizontal Body Axis System

Let the velocities and angular velocities expressed in the general body axis system (G-xb, yb, zb), which is fixed on the boat axes, be u,
v, w and p, q, r. The forces and moments expressed in this axis system are formulated from the Euler-Lamb’s equations of motion
(Lamb 1932) as:

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 X b  m  mx  u  m  m y  rv  m  mz  qw
Yb  m  m y  v  m  mz  pw  m  mx  ru
Z b  m  mz  w  m  mx  qu  m  m y  pv
K b  I xx  J xx  p  m y  mz  wv  I yy  J yy   I zz  J zz  qr
(A-18)

M b  I yy  J yy  q  mz  mx  uw  I zz  J zz   I xx  J xx  rp
N b  I zz  J zz  r  mx  m y  uv  I xx  J xx   I yy  J yy  pq

where Xb to Nb on left side are forces and moments expressed in the general body axis system. For the sake of simplicity, it is assumed
that the center of added mass of the hull coincides with the C.G. of the boat, and the principal axes of inertia of the added mass also
coincide with those of the boat.

Then we willl consider the horizontal body axis system (G-x, y, z), which originates on the C.G. of the boat. The x-axis lies along the
center line of the boat on the still-water plane and is positive forward. The y-axis is positive to starboard in the still-water plane. The z-
axis is positive down as shown in Figure A-4.

The Euler angles, ψ, θ, φ, are used for the translation from the earth fixed axis system into the general body axis system. To express the
tacking motion of the boat in calm water, we assume the pitch angle, θ, is equal to zero. In this case, the angles around x-, y-, z-axis of
the horizontal body axis system, Ψ, Θ, Φ, are described as:

  ,     0,  
The xb-axis also coincides with the x-axis. Therefore, the components of the vector expressed in the horizontal body axis system are
transformed into those expressed in the general body axis system as:

 xb  1 0 0  x
 y   0 cos  sin    y  (A-19)
 b   
 zb  0  sin  cos    z 

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 Figure A-4: Relation between general body axis system and horizontal body axis system.

The velocities and angular velocities expressed in the general body axis system in equation (A-18) are formulated by ones expressed in
the horizontal body axis system such as U, V, W, and  ,  ,  , using equation (A-19) as:

u U
v  V cos   W sin   V cos   W  0
w  V sin   W cos   V sin   W  0
u  U
v  V cos   V sin 
w  V sin   V cos 
(A-20)
p  
q   cos    sin    sin     0
r   sin    cos    cos     0 
p  
q   sin    cos 
r   cos    sin 

Substituting equations (A-20) into (A-18), we find the following relationships:

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 
X b  m  mx  U  m  m y cos 2   mz sin 2  V 
Y  m  m   V cos   V sin  
b y

 m  mz  V sin   m  mx  U cos 

Z b  m  mz   V sin   V cos  
 m  mx  U sin   m  m y  V cos 
K b  I xx  J xx    m y  mz   V sin   V cos  (A-21)

 I yy  J yy   I zz  J zz   sin   cos 

M b  I yy  J yy   sin    cos    mz  mx  U  V sin  
 I zz  J zz   I xx  J xx  cos 
N b  I zz  J zz   cos    sin    mx  m y  UV cos 
 I xx  J xx   I yy  J yy  sin 

The forces and moments expressed in equation (A-21) are transformed to the horizontal body axis system using the inverted matrix of
equation (A-19) as:

X  Xb
Y  Yb cos   Z b sin 
Z  Yb sin   Z b cos 
(A-22)
K  Kb
M  M b cos   N b sin 
N  M b sin   N b cos 
Substituting equations (A-21) into (A-22), we find the forces and moments expressed in the horizontal body axis system using the
velocities and angular velocities expressed in the same coordinate system. To calculate the motion of the tacking maneuver in calm
water, we adopt four equations about surge, sway, roll and yaw as follows:

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 X  ( m  mx ) U  (m  m y cos 2   mz sin 2  ) V

Y  ( m  m y cos 2   mz sin 2  ) V  (m  mx ) U

 2(mz  m y ) sin  cos   V

K  ( I xx  J xx )  ( I yy  J yy )  ( I zz  J zz )sin  cos   2 (A-23)

 (m y  mz ) sin  cos   V 2


N  ( I yy  J yy ) sin 2   ( I zz  J zz ) cos 2   
 2( I yy  J yy )  ( I zz  J zz )sin  cos   
 ( mx  m y cos 2   mz sin 2  ) UV

The third terms in both the K and N moment equations indicate the Munk Moments. When the hydrodynamic derivatives of the hull are
obtained from the oblique towing test using scaled model, these Munk Moments are included in the measured data. For example, when
the φ is assumed to be small and U is constant, the Munk moment in the K moment equation is expressed as:

m y  mz   V 2
and in N moment equation is expressed as:

 mx  m y U  V  mzU   2 V

These terms are functions of φV2, V and φ2V. As noted mentioned the magnitudes of these components are included in the measured
hydrodynamic derivatives of K'VVφ, N'V and N'Vφφ, respectively. Therefore, for the equations of motion, these third terms are eliminated
from the left side of equations (5) and (6).

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
 Discussion
Eugene Miller, Member
I would like to congratulate the authors on developing a successful simulation of a complex dynamic maneuver (tacking a sailboat) and
describing their approach and results in a clear and concise paper. The collection and presentation of data from real tacking maneuvers
is to be commended. I have the following questions / clarifications for the authors.

1. Were comparisons developed for steady state sailing prior to the start of the tacking maneuver? This would include
comparisons of parameters such as speed, heel angle, rudder angle, leeway etc. vs. apparent wind speed and direction. Such a
comparison would validate the steady state part of the modeling / simulation including the sail forces (Figure 2) and the steady
state hydrodynamic forces. The steady state rudder angle would be of particular interest since it would help to validate the
modeling of the overall moments acting from the sails and hull. Getting the steady state condition correct is a prerequisite to
successful simulation of the unsteady tacking maneuver.

2. The approach to the modeling of the sail force variation during the tacking maneuver as presented in Figure 6 seems to be
logical and reasonably represents the X forces shown in the measured data presented in Figures 4 and 5. However, the
measured Y force data in Figures 4 and 5 seem to be more linear with respect to the apparent wind angle than the Y force
model presented in Figure 6. Do the authors think that this would have a significant effect on the results of the simulated
maneuver?

3. Equation 9 presents the effective angle of attack of the rudder which is the key parameter in determining the rudder forces.
This equation introduces γR for the effect of the downwash from the hull and keel on the inflow angle to the rudder. This is
real and important effect. The data presented in Figure 9 on rudder force and moment vs rudder angle for various drift angles
seems to imply that γR may be a function of drift angle. Do the authors have a feel for how significant the exact value of γR is
in the results of the simulation? If it is important, then a refined model for γ R could be developed. In the extreme case, some
sort of vortex tracking model could be developed to track the convection of the vorticity down stream from the keel to the
rudder. Typically, it would take 1 to 2 seconds (about 10% of the time for the maneuver) for changed vorticity (due to change
in drift angle) generated by the keel to reach the rudder.

4. Equations 11 and 12 present the relationships used to determine the contribution of yaw rate to the Y and N forces. Typically,
in a dynamic maneuver, the yaw moment damping (N vs yaw rate) is very important to the results of a maneuver. Given the
importance of this parameter did the authors make any independent estimates of the value? It might also be expected that at
the high turning (yaw) rates that occur during the tacking maneuver that the yaw damping would be a non-linear function of
the yaw rate.

5. If the authors get the opportunity, it would be interesting to conduct some tests to verify the maneuvering part of the
simulation, independent of the sail forces. Conceptually, this would involve tests where the boat (sails down) is towed up to
speed, released and then various maneuvers conducted (turns, Zig-Zag) with the rudder as the boat slows. The results could be
used to validate the maneuvering part of the simulation. If the results do not agree, then some sort of parameter identification
could be used to generate better estimates of the hydrodynamic forces (coefficients) used in the simulation.

6. An obvious application of the simulation of tacking maneuvers is to explore various alternatives (use of larger or smaller
rudder angles, etc.) to minimize the speed loss to windward during the tacking maneuver. Do the authors think that the
tacking maneuver simulation is sufficiently developed or could be developed to be a tool of the optimization of tacking
maneuvers?

Authors’ Closure
We appreciate all comments and questions by the discusser. Our replies to the questions are as follows:

Question 1
At first, we measured the steady sailing condition of each boat by the sea tests and compared them with the results of our velocity
prediction program (VPP). The VPP provided the sailing state parameters such as speed, heel angle, rudder angle, leeway angle etc.
using hydrodynamic and aerodynamic coefficients of the boat, which are the same values as for the calculation of the tacking
simulation. The results of VPP agreed well with the measured data. The calculation of the tacking simulation started from the
equilibrium state evaluated by the VPP.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
Question 2
The authors appreciate the discusser’s point. When we made the model of the sail force variation during the tacking maneuver, we were
strongly affected by the variation curve shown in Figure 2, although these curves indicate the steady state condition, not the dynamic
condition. The measured variation of Y’s in dynamic condition in Figures 4 and 5 is almost linear with respect to the apparent wind
angle, however, the curve of Y’s0 in Figure 2 cannot be linear. This is because we assumed the variation of Y’s0 as shown in Figure 6.
The linear approximation of Y’s0 at stage C as the extension of stage B may provide more concise modeling and easy treatment of the
calculation.

As the difference between the original curve of Y’s0 and the linear approximation at stage C in Figure 6 is small, the results of the
simulated maneuver are not affected so much by the linear approximation, because the hydrodynamic derivatives of the hull and rudder
played more important roles than that of the sail at this stage.

Question 3
The authors also appreciate the discusser’s point. The value of the decreasing ratio of inflow angle γR is very important for the
simulation. We performed free-rotating rudder tests at various leeway angles by oblique towing test using each model. The value of γ R
was obtained as the ratio of the rudder angle indicated under free-rotating condition, δ0, to the leeway angle β. The results are shown in
Figure D-1 for the three models. Although the data scattered due to friction of rudder shaft, the strong dependency on the leeway angle
can be seen. Therefore, we deduced a polynomial expression from the results and used it for the simulation.

In this content, Hoerner (1975) showed a formulation of the down wash angle ε caused by an elliptic loading wing as follows;

CL
ε = 1.6
π A
(D-1)

where CL is lift coefficient and A is aspect ratio of main wing, respectively. However, this formulation can be applied for a narrow
region of a small attack angle.

After that, Keuning et al. (2007) presented an empirical formulation for the case of fin keel of a sailing yacht, as follows:

C LK
ε = a0 (D-2)
AeK
where a0 = 0.136 for heel angle 0 degree

a0 = 0.137 for heel angle 15 degrees

CLK = lift coefficient of fin keel

AeK = effective aspect ratio of fin keel (twice of geometric aspect ratio).

The decreasing ratio of inflow angle at the rudder is expressed as follows;

0  -
R   (D-3)
 
Keuning et al. applied this formulation for several fin keels with different aspect ratios, of which results show good agreement with the
experiments. In figure (D-1), the calculated results using equation (D-2) are also shown as the curves for three models. It can be seen
that the calculated results represent good agreement with the experiments. Therefore, the formulation of equation (D-2) presented by
Keuning et al. can be used effectively for the evaluation of the decreasing ratio of inflow angle at the rudder. However, this formulation
was not available when the authors conducted the simulations.

As for the convection of the vorticity, some sort of vortex tracking model could be considered at the first stage of tacking maneuver.
However, when the boat attains the constant turning motion, the inflow angle at the rudder position becomes very large because the
turning radius is relatively small. This means that the rudder is outside of the downwash of the fin keel and that a decrease in inflow

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
angle might not occur as mentioned in the paragraph after equation (9) in the text. Therefore, it can be considered that the convection of
the vorticity may not affect very much the results of the simulation.

Decreasing ratio of inflow angle γR

0.8

0.6

0.4

0.2

0
0 5 10 15 20
Leeway angle β[deg]

Fujin Fujin (Cal.)

Fair V Fair V (Cal.)
Sea Dragon Sea Dragon (Cal.)

Figure D-1: Variation of decreasing ratio of inflow angle vs. leeway angle for three models.

REFERENCES
HOERNER, S. and Borst, H. (1975). Fluid Dynamic Lift. Hoerner. 11-16.
KEUNING, J. A., Katgert, M. and Vermeulen, K. J. (2007), “Further Analysis of the Forces on Keel and Rudder of a Sailing Yacht,”
18th Chesapeake Sailing Yacht Symposium, SNAME, 49-61.

Question 4
Unfortunately, we could not perform dynamic model tests such as PMM test due to the restriction of our experimental facilities.
Therefore the non-linear terms of the yaw damping due to yaw rate cannot be evaluated. However, as is mentioned below in the answer
for question (5), we performed many maneuvering tests (turns, Zig-Zag) without sails, and the results were compared with the
simulation. The simulated results agreed well with the measured ones. Hence, it can be considered that the contribution of non-linear
terms may not be so significant.

Question 5
For the case of Fair V, we performed many maneuvering tests (turns, Zig-Zag) without sails, and the results were compared with the
simulation. The boat was thrust by a propeller which was set 2.5 meters before the rudder. Therefore the numerical simulation was
conducted considering that the constant thrust force vector acted at the propeller position. Since the distance between rudder and
propeller was sufficiently large, the effect on the rudder by the propeller was not taken into account. The comparison between
measurements and the results of the simulation was shown in references (Masuyama et al. 1993 and 1995). Although the equations of
motion in the references were expressed with the terms of distance between center of gravity and center of added mass of the boat, xG,
the simulated results agreed well with the measurements. Therefore, we concluded that the method used for the derivation of
derivatives and coefficients of the boat was adequate.

Question 6
For the case of Fair V, we showed the results of a tacking maneuver under three different steering angle procedures such as (1) first
small then large, (2) first large then small, and (3) continuously small angle in reference (Masuyama et al. 1995). In our sea tests, the
steering angle procedure of (1) yielded the best tacking maneuver with small reduction and quick recovery in boat speed and the best
VMG. The tacking simulation was performed using the measured time histories of rudder angle. The simulated results indicated good
agreement with the measurements in both trajectories and velocity variations during tacking maneuvers for the three cases. The

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver
predominance of the steering angle procedure of (1) was also demonstrated by the simulation. For the other two boat cases, the
comparison between measured and simulated results under various steering procedures were also performed, which show good
agreement.

From these results, the authors think that the equations of motion and the mathematical models for sailing yacht maneuvers presented
here could be used for any sailing yacht. Therefore, if detailed data of a target boat are provided, the optimization of tacking maneuvers
could be possible.

Again, we would like to express our sincere appreciation to the discusser for his valuable opinions and careful observation of our
results.

Tacking Simulation Of Sailing Yachts With New Model Of Aerodynamic Force Variation During Tacking Maneuver