Conference PaperPDF Available

DEVELOPMENT OF A DIRECTIONAL LOAD CELL TO MEASURE FLYING SAIL AERODYNAMIC LOAD

Authors:

Abstract and Figures

Full-scale measurement of forces on sails is a difficult but achievable task. This paper describes the development of a system which allows determination of the aerodynamic performance of flying sails directly through measurement of the forces and moments produced. A directional load cell (DLC) has been developed by combining a tension load cell and an Inertial Measurement Unit (IMU), enabling the orthogonal force components at each corner of the sail to be measured. By attaching one of these devices to each corner of a yacht's offwind sail, the overall forces and moments produced by the sail can then be deduced. Tests have been carried out on a Stewart 34 yacht to determine the difference in forces produced by an asymmetric and a symmetric spinnaker. NOMENCLATURE A Sail cloth area (m 2) AWA Apparent Wind Angle (°) AWS Apparent Wind Speed (m/s) CF X drive force coefficient () CF Y side force coefficient () CF Z vertical force coefficient () CM X roll moment coefficient () CM Y pitch moment coefficient () CM Z yaw moment coefficient () DLC Directional Load Cell FEPV Force Evaluation via Pressures and VSPARS FSI Fluid Structure Interaction IMU Inertial Measurement Unit ρ air density of air (kg.m-3)
Content may be subject to copyright.
5th High Performance Yacht Design Conference
Auckland, 10-12 March, 2015
DEVELOPMENT OF A DIRECTIONAL LOAD CELL TO MEASURE FLYING
SAIL AERODYNAMIC LOADS
David J. Le Pelley1, d.lepelley@auckland.ac.nz
Peter J. Richards2, pj.richards@auckland.ac.nz
Arthur Berthier3
Abstract. Full-scale measurement of forces on sails is a difficult but achievable task. This paper describes the development of a
system which allows determination of the aerodynamic performance of flying sails directly through measurement of the forces and
moments produced. A directional load cell (DLC) has been developed by combining a tension load cell and an Inertial Measurement
Unit (IMU), enabling the orthogonal force components at each corner of the sail to be measured. By attaching one of these devices to
each corner of a yacht’s offwind sail, the overall forces and moments produced by the sail can then be deduced. Tests have been
carried out on a Stewart 34 yacht to determine the difference in forces produced by an asymmetric and a symmetric spinnaker.
1 Yacht Research Manager, University of Auckland
2 Associate Professor, Department of Mechanical Engineering, University of Auckland
3 MEngSt Graduate, Yacht Research Unit, University of Auckland
NOMENCLATURE
A Sail cloth area (m2)
AWA Apparent Wind Angle (°)
AWS Apparent Wind Speed (m/s)
CFX drive force coefficient ()
CFY side force coefficient ()
CFZ vertical force coefficient ()
CMX roll moment coefficient ()
CMY pitch moment coefficient ()
CMZ yaw moment coefficient ()
DLC Directional Load Cell
FEPV Force Evaluation via Pressures and
VSPARS
FSI Fluid Structure Interaction
IMU Inertial Measurement Unit
ρair density of air (kg.m-3)
R 
  positionvector
TWS
  TrueWindSpeedm/s
VA apparent wind speed (m/s)
VS boat speed (m/s)
VSPARS Visual Sail Position And Rig Shape
1. INTRODUCTION
For a sailing yacht, the sole source of propulsion is the
sails, with everything else providing a retarding force. To
optimise performance it is therefore essential to optimise
the driving force provided by the sails, both upwind and
downwind. Closing the loop by measuring the
performance of sails in service is hard to achieve but can
provide valuable data for validation of simulation tools
as well as a direct means of sail optimisation.
The most common method of deducing the performance
of sails is by careful measurement of the boat speed
compared to a previous baseline [1] or to another
identical yacht sailing in similar conditions. This
procedure has been commonly used by teams developing
designs for the America’s Cup [2]. In the 32nd
America’s Cup in 2007 teams were permitted to build
two boats, and thus 2-boat testing took place formally in-
house. For the 34th America’s Cup the rule restricted the
amount of time that yachts could sail, making 2-boat
testing much harder. Teams would attempt to line up
informally against competitors to gain relative
performance information without knowing the
configuration of the other team.
One of the main problems with trying to determine sail
performance from two-boat testing is that, in spite of best
efforts, the boats don’t sail in identical environmental
conditions. There are often small hydrodynamic
differences between boats even if the hulls are identical.
Due to time pressure it is not unusual for sails to be
tested on the same day as other elements, such as
different foils or crew combinations. Differences such as
these mean that very long tests, significant data analysis
and noise filtering are required to make any sense of the
results.
Instead of deducing the performance from analysis of
boat speed changes, researchers have attempted to
measure the aerodynamic loads directly. Sailing
dynamometers have been developed [3-6] where all of
the rigging elements are connected to an internal frame,
enabling the overall aerodynamic forces and moments to
be measured directly. Using this method it is not possible
to break down the individual components of headsail,
foresail and windage contributions.
Other studies [7] have used strain-gauged rigging
elements, such as shrouds and genoa sheets, to determine
the tension forces of individual components. These have
been used for tuning and validation of FSI models.
However, without knowledge of the direction along
which the force acts, it is impossible to resolve this into
orthogonal force components. Whilst shroud angles may
66
be measured directly, sheet angles which change with
sail trim must be constantly measured.
Several attempts have been made to measure the pressure
distribution on the surface of sails [8-11]. This provides
more detail than an overall aerodynamic force but is time
consuming to interpret and requires knowledge of the sail
shape to be able to convert the pressures into a force
vector.
Two methods to measure sail forces have been developed
by the Yacht Research Unit at the University of
Auckland. The first, known as Force Evaluation via
Pressures and VSPARS (FEPV) [12, 13] uses a
combination of the shape of the sail measured by the
VSPARS camera system [14] and differential pressure
data measured on the sail surface using an array of
pressure transducers. The pressures are interpolated and
extrapolated over the sail surface, and the shape is used
to find the surface normal direction so that the orthogonal
force components at each discretised cell on the sail
surface can be calculated. Summing all the components
allows the global forces and moments to be calculated
about the datum location. This technique can be used on
both upwind and downwind sails. However, it is time
consuming to set up, relatively expensive and care must
be taken with the pressure transducers.
The second method developed by the Yacht Research
Unit is described in this paper. It provides a method of
directly measuring the forces produced by flying sails.
These sails are used offwind and are those which are
attached to the yacht at only 3 places (head, clew, tack)
rather than in a luff groove or on a tight forestay such as
a mainsail or jib. By placing a load measuring device in
line with each corner of the sail the individual tension
forces can be measured. An Inertial Measurement Unit
(IMU) then provides the direction of the load through
yaw, pitch and roll angles. The orthogonal force
components from each corner can be calculated and then
summed giving the aerodynamic force on the sail. The
individual forces can also be combined with the distance
of the corner’s anchor point from the yacht’s datum to
deduce the overall aerodynamic moments.
2. DIRECTIONAL LOAD CELL SYSTEM
2.1 Overview of the DLC System
Once a yacht sails deeper than approximately 50° to the
apparent wind direction, a flying sail is typically set
which is only attached to the yacht’s structure at three
points – the head, the clew and the tack. This makes a
direct force measurement possible. Whilst it is relatively
easy to measure the magnitude of the force applied to
each corner, in order to determine the resultant force on
the sail it is also necessary to know the direction of each
force. Further, if the moment is also required then one
position on the line of action of the force must also be
defined.
Offwind sails can be divided into two types: (1)
symmetric, which have the two vertical luffs of equal
length, and (2) asymmetric, where the luff is longer than
the leech. A symmetric spinnaker is flown from a
spinnaker pole as shown in Figure 1. When the yacht
gybes, the symmetric spinnaker stays in place and the old
trailing edge of the sail becomes the new leading edge.
Likewise, the tack (lower corner of the leading edge) and
clew (lower corner of the trailing edge) remain
physically in place but swap roles and the pole is moved
to the new windward side. In contrast, the tack of an
asymmetric spinnaker is attached to the bowsprit and
remains in place during a gybe, whilst the sail is rotated
around and the clew swaps sides. Many offwind sails are
now set on a removable roller-furler unit which allows
larger sails to be handled with ease.
Figure 1. (a) The tack of the spinnaker attached to the
end of the spinnaker pole and (b) a close-up of the
directional load cell (DLC), which is the small box
with red electrical tape around it.
67
Both types of sail are attached to the top of the mast,
where the halyard used to hoist the sail runs over a
sheave, and have the clew attached to the trim sheet
which runs to a turning block on the deck. The position
of the top of the mast and the block on the deck can be
specified by direct measurement. For symmetric
spinnakers the end of the pole will move as the sail is
trimmed and hence measurement of the direction of the
pole is also needed together with its length and height up
the mast. With asymmetric spinnakers the tack is
connected to the bow of the boat or a fixed bowsprit and
hence the location of this attachment point is more
readily defined. The location of the clew will move
during trimming but the point where the sheet transmits
the force to the deck will be fixed. When a gybe is
performed with either type of spinnaker, changes need to
be made in the acquisition software to reflect the new
DLC roles and fixing locations of the tack and clew
accordingly.
The main components of the DLC system, as illustrated
in Figure 2, are:
• Three directional load cells which measure the force
applied to each corner of the sail and the direction of
that force.
• A base station Inertial Measurement Unit (IMU)
attached to the deck of the yacht which transmits the
orientation of the hull.
• A spinnaker pole IMU which provides the orientation
of the pole relative to the hull.
• Data logging computer.
Figure 2. Schematic of the DLC system.
One of the main advantages of the DLC and FEPV
systems over a method which deduces aerodynamic
performance from boat speed changes is that both of
these systems can capture the high frequency response of
the sail. It is possible to look at the dynamic response of
sails to gusts, trim changes or motion over waves. It is
also possible to deduce the average forces and moments
over a much shorter period than by looking at the boat
speed changes. Rather than sailing in one direction at a
particular trim for several minutes, it is possible to
perform a sweep, pausing at direction intervals for a few
seconds, to enable the whole sail performance envelope
to be captured for a certain wind speed.
2.2 DLC components
Each DLC manufactured by the Yacht Research Unit
comprises four main components:
1. A stainless steel link placed between the corners of
the spinnaker and the halyard/sheet, which is
expected to support a maximum static load of
10kN,
2. Strain gauges measuring the loads, with a
measurement range up to 6kN.
3. An Inertial Measurement Unit (IMU) measuring the
yaw, pitch and roll of the DLC.
4. Electronic hardware acquiring the data and
transmitting it back to the central computer.
One DLC is placed on each corner of the spinnaker, as
shown in Figure 1(a), and is joined to the ring in the sail
and to the sheet by shackles. The overall length of the
DLC is approximately 100mm and hence has little effect
on the location of the sail. For the main link an austenitic
stainless steel is used, both for corrosion resistance and
to avoid magnetic interference of the magnetometers in
the IMU. A full-Wheatstone bridge arrangement of strain
gauges is used with two gauges along the axis of the link
and two across it. This arrangement provides temperature
compensation of the strain measurement. Once
constructed, the load cells were calibrated on an Instrom
tensile testing machine and were found to be linear over
the expected operating range up to 6kN. One unit was
tested to failure which occurred in the connecting shackle
at a force greater than 10kN.
2.3 Electronics within the DLC
Since the DLCs are used in a manner where running
leads to them would be difficult, if not impossible, it was
decided to use wireless technology. Hence each DLC
needed to contain its own power supply and electronics.
These were mounted onto two support plates which were
fixed to just one end of the structural link, the upper end
in Figure 3. The complete unit was then enclosed in a 3D
printed plastic casing. The major components of the
electronic system are:
• A lithium battery which allows the DLC to run
autonomously for up to 5 hours. This can be recharged
through an external socket.
• A Wheatstone bridge chipset which processes the
strain gauge measurements and amplifies the signal
using an operational amplifier circuit. The range of the
output is 0-1500mV.
• An Inertial Measurement Unit (IMU) which is used to
collect the angular position of the DLC in terms of
heading, pitch and heel angles.
• An Arduino Pro Micro microprocessor. This central
chipset processes all the signals and sends the data to
the wireless transmitter.
• An XBee 2mW radio transmitter, which transmits the
data to the data acquisition computer in standard
68
NMEA-type sentences. The range of transmission is
greater than 50m on the water.
Figure 3. The DLC unit showing the stainless steel link,
some of the electronics and half of the casing.
2.4 DLC data processing
The angular information measured by each IMU is first
pre-processed by the DLC’s microcontroller by running
it through a Direction Cosine Matrix (DCM) algorithm
[15]. The accelerometers and magnetometers both
provide absolute references of direction, but are subject
to significant noise in the form of erroneous inertial and
magnetic forces. The DCM algorithm attempts to
eliminate this noise by essentially predicting the new
orientation at any timestep using the more accurate
gyroscope readings. If the orientation of the device at any
time t is given as DCM(t) then the predicted position
DCM (t+dt) can be found by using the current rate of
turn provided by the rate gyros applied over a time dt.
The DCM effectively uses the low frequency
components of the accelerometers and magnetometers
and the high frequency component of the gyros. The
corrected angular positions, along with the load cell data,
are then sent wirelessly from the DLC to the base
receiver.
For each different DLC the angular information is then
processed by determining the difference between that
given by the particular unit and that from the base station
attached to the yacht’s deck. This allows the force
measured by the strain gauges to be resolved into
components along the yacht-fixed axes. Once these
components are known they can be summed to give the
resultant force on the yacht. If required, the heel and
pitch angles determined by the base station can be used
to resolve the forces into components parallel and
perpendicular to the sea surface.
Once the force components are known in relation to the
yacht’s fixed axes the total moment can be determined
from:
x
(1)
where k is the index for each load cell, R is the position
vector of the known points on the line of action of each
force and F is the force. For the tack of a symmetric
spinnaker it is assumed that the spinnaker pole is parallel
to the deck and therefore the height is obtained by
measurement during testing, however the x and y co-
ordinates need to be calculated by taking the difference
in angle between the IMU attached to the spinnaker pole
and that attached to the deck. Then, by using the known
length of the pole, the position of the end of the pole is
easily obtained.
3. FULL SCALE TESTING
3.1 Test description
Tests were performed using a Stewart 34 yacht [16]
equipped with a mainsail and two offwind sails. One was
a symmetric spinnaker with an area of 113.6m2 built to
the class rules and flown from a spinnaker pole. The
other was an asymmetric spinnaker with an area of
88.6m2, designed for general purpose running in the style
of the TP52 class downwind sails. This was flown from
the end of a short sprit which extended 400mm forward
of the forestay at deck height. The purpose of the tests
was to determine the cross-over angle where one sail out-
performs the other for a particular wind speed. The
testing periods were deliberately kept short to examine
the potential of the system to determine sail performance
without lengthy amounts of testing time.
Simultaneous testing was carried out using the FEPV
system on the asymmetric spinnaker to allow comparison
between methods.
The yacht was equipped with:
- 3-axis sonic anemometer 1.2m above masthead
- Base IMU measuring roll, pitch and yaw
- VSPARS camera system
- Pressure sensors on asymmetric spinnaker
- 3 x DLC units on sail corners
- Spinnaker pole angle sensor
- GPS
- Masthead cup anemometer and direction vane
- Internal fluxgate compass
- Paddlewheel log
All of the data apart from the VSPARS images were
obtained using a dedicated data acquisition system at a
maximum frequency of 10Hz. Some of the
69
instrumentation, for example the paddlewheel log, had a
lower data output rate so the data was linearly
interpolated during post-processing so that a complete
10Hz record existed. The VSPARS images were
captured using GoPro cameras and time-synchronised to
the main data log.
The testing was carried out in the Hauraki Gulf, New
Zealand, on 22nd April 2014. Measurements were made
for the asymmetric spinnaker on starboard tack, followed
by a gybe and measurements on port tack over a total
period of around 20 minutes. Then the yacht was
motored back upwind and the symmetric spinnaker tested
on starboard tack only in the same geographical area.
Tests on port tack for the symmetric spinnaker were not
completed due to time constraints. The true wind speed
(TWS) was quite variable, ranging from 12 knots to 18
knots from the southwest.
For each condition, a very slow sweep was made
between tight and deep apparent wind angles whilst the
sails were trimmed continuously. A sweep starting from
the tightest angle at which the boat could sail to deep
downwind typically took around 10 minutes. The
asymmetric could be flown as tight as 70° apparent wind
angle (AWA) whilst the symmetric could only get to
around 90° AWA in the tested conditions. This was
repeated twice for each sail / tack combination.
Data viewing software, which synchronised the data time
history with the videos available, was then used to
average suitable sections of the sweep over short time
periods, around 10s, such that data associated with
discrete apparent wind angles could be examined.
3.2 Results of DLC tests
The force-area and moment-volume coefficients have
been calculated to allow the true cross-over conditions
between the smaller asymmetric and larger symmetric
spinnaker to be determined. This removes the effect of
wind speed variations whilst retaining the principle that a
bigger sail will produce a greater force. The drive force
area and roll moment volume are defined as:
2
.12
X
X
air A
F
CF A V
(2)
2
.12
X
X
air A
M
CM V V
(3)
where VA is the apparent wind speed in m/s and ρair is the
density of air = 1.2 kg/m3. The other force-area and
moment-volume coefficients are calculated in a similar
manner.
All of the results presented here are given in the yacht’s
body-fixed axis system which conforms to a standard
right-hand axis convention. For comparison purposes,
some of the forces and moments have been negated
which is clearly marked in each case.
Ideally, the difference in performance between the two
spinnakers should be apparent from examination of the
boat speed plotted against AWA. Figure 4 shows this
relationship.
Figure 4. Boat speed plotted against AWA.
However, this does not take account of fluctuating wind
speeds during the tests. By dividing the measured boat
speed VS by the calculated apparent wind speed a non-
dimensionalised plot of VS/VA is shown in Figure 5.
Figure 5. Boat speed divided by apparent wind speed
plotted against AWA.
Now the performance of the asymmetric spinnaker at
tighter angles is slightly more obvious. However, it is
unclear where the crossover lies and there is significant
scatter in the results. This is partly due to the fact that the
boatspeed is not directly related to the apparent wind
speed, so this type of analysis is only an approximation.
A similar scatter occurs if the true wind speed is used
instead of the apparent wind speed.
The force and moment results obtained using the DLCs
are shown in Figures 6-11.
70
Figure 6. Drive force-area CFx.A plotted against AWA.
Solid lines show approximate performance envelopes.
Figure 7. Side force-area CFy.A plotted against AWA (port
tack values are negated)
Figure 8. Vertical force-area CFz.A plotted against AWA
Figure 9. Roll moment-volume CMx.V plotted against
AWA (port tack values are negated).
Figure 10. Pitch moment-volume CMy.V plotted against
AWA .
Figure 11. Yaw moment-volume CMz.V plotted against
AWA (port tack values are negated).
Figure 6 shows that the symmetric spinnaker produces
more driving force at deeper angles whilst the
asymmetric spinnaker produces more driving force at
tighter angles. The solid and dashed lines represent the
approximate envelope of the performance, where points
below the line indicate sub-optimal trim. Whilst the port
and starboard tack asymmetric data agree reasonably
well for tight angles, there is a significant difference at
deeper angles. There appear to be almost two distinct
performance envelopes for the asymmetric on port tack,
the lower agreeing well with starboard tack and the upper
having significantly more driving force. However, this is
a very small dataset as all the data for each sail/tack
combination was recorded from only two AWA sweeps.
Further examination also suggests some further
dependence on wind speed not accounted for by non-
dimensionalising by the dynamic pressure. The points
denoted by two larger markers in Figure 6 were obtained
in quite different wind speeds. The one at the tighter
angle was obtained in approximately 12 knots TWS
whilst the one at the deeper angle was measured during a
period of around 18 knots TWS, with the yacht still
accelerating. Whilst there is some scatter, the cross-over
angle between the sails based on driving force alone is
around 110°, where the asymmetric produces more
driving force at tighter angles and the symmetric
produces more at deeper angles.
Figure 7 shows the comparison of the side force. Here
the correlation is much better between tacks for the
71
asymmetric sail. The magnitude of side force created by
the asymmetric and symmetric sails is quite similar.
Figure 8 shows the vertical force. As expected, the
symmetric sail produces significantly more upwards
force than the asymmetric, due to a larger girth at the
higher sections which results in a larger projected area in
the plane parallel to the deck.
The roll moment is shown in Figure 9. There is little
difference between the two sails types at tight angles
where the heel moment is of more importance to
performance. This would suggest that, in this case, the
performance difference between the sails is well
represented by the difference in driving force. Each sail
was tested to the tightest angle possible before broaching
occurred. The symmetric spinnaker could only be held
until 90° AWA whereas the asymmetric could be held up
to 70° AWA. However, there was approximately a 2 knot
higher average TWS for the tightest angle symmetric
spinnaker tests than the equivalent asymmetric spinnaker
tests, which would account for some of this difference.
The pitch moment shown in Figure 10 and the yaw
moment shown in Figure 11 are both very similar
between sails. Both sails exhibit a reversal in direction of
the yaw moment at angles deeper than around 100°
AWA. At tight angles the resultant force line of the drive
and side forces passes very close to the mast location so
the values are small, and then as yacht sails deeper the
sheet is eased and the sail is rotated to windward, moving
the centre of effort to windward and creating a significant
“bow to leeward” yaw moment. This is usually
counteracted by the positive yaw moment of the mainsail
(not measured) to give neutral helm. The magnitude of
this negative yaw moment at deep angles might be
expected to be larger for the symmetric sail as the
spinnaker pole moves the centre of effort further off
centreline than for the asymmetric, but this is not evident
in the data.
3.3 Comparison with FEPV
The overall spinnaker loads were also calculated using
the FEPV system for the asymmetric spinnaker on port
tack. The FEPV system measures pressure on the sail
surface which is then integrated over the measured shape
to deduce the force distribution over the surface. This
was part of a combined study to measure the dynamic
response of the sails to full scale sailing conditions [17].
A comparison of the time history of the forces measured
using both systems is shown in Figure 12.
Figure 12. Comparison of forces measured by FEPV and
DLC systems for asymmetric spinnaker on port tack as a
function of time.
It can be seen that the trends in each force component
agree very well between the two systems. However, there
is a significant offset in the magnitude of the forces. The
dynamic responses seem to be well captured by both
systems. The output of the DLC system might be
expected to appear slightly more damped than the FEPV
system, as the stretch of the sail fabric might be expected
to smooth out loading peaks which would be accurately
recorded by the pressures. However this is not noticeable
in the results.
Figures 13-15 show the comparison of DLC and FEPV
derived force-area coefficients for a selection of the same
averaging periods as presented in Section 3.2. Figures
14-17 show the equivalent plots for the moment-
volumes. The measurement points have been joined by
lines to make the trends obvious, but the measurement
points are not necessarily linked and are taken from
different times along the sail sweeps.
Figure 13. Comparison of CFx.A between FEPV and DLC
systems for asymmetric spinnaker on port tack.
72
Figure 14. Comparison of CFy.A between FEPV and DLC
systems for asymmetric spinnaker on port tack.
Figure 15. Comparison of CFz.A between FEPV and DLC
systems for asymmetric spinnaker on port tack.
Figure 16. Comparison of CMx.V between FEPV and DLC
systems for asymmetric spinnaker on port tack.
Figure 17. Comparison of CMy.V between FEPV and DLC
systems for asymmetric spinnaker on port tack.
Figure 18. Comparison of CMz.V between FEPV and DLC
systems for asymmetric spinnaker on port tack.
It is evident that the trends in the data are extremely well
reproduced yet there is a significant offset between the
two methods. Analysis of dynamic results [17] has
shown that the trends in dynamic behaviour are picked
up well between the two systems, with the force peaks
being of similar relative intensity and duration. Whilst
the relative differences between the FEPV and the DLC
loads vary for each force and moment coefficient, the
resultant force measured by the DLCs is significantly
lower than that measured by FEPV. The reason for this
difference is presently unclear and hopefully will be
resolved by further testing.
3.4 Comparison with wind tunnel tests
A model of the Stewart 34 was constructed at a scale of
1:6.96 and tested in the University of Auckland’s
Twisted Flow Wind Tunnel [18]. Exact scale models of
the mainsail, asymmetric and symmetric spinnakers were
flown in the same configuration as at full scale. The
velocity and twist profiles were calculated and
reproduced for typical downwind sailing conditions. The
model was tested from 50° to 180° AWA in steps of 10°
and a range of heel angles from 0° to 15° in 5° steps.
Different levels of depowering were also examined.
Whilst the main results of this study [19] are outside the
scope of this paper, it is interesting to look at the
equivalent drive force coefficient recorded for the yacht
as a whole, including the contributions of the mainsail
and windage elements. This is shown in Figure 19 for the
condition of 5° heel angle.
Figure 19. Wind tunnel results of drive force coefficient for
asymmetric and symmetric spinnakers.
73
Whilst it is not possible to make a quantitative
comparison, because of the presence of the mainsail, a
number of trends can be seen to agree with the full scale
DLC results from Figure 6. The crossover between sails
occurs at a similar AWA angle of around 100°. The force
produced by the symmetric spinnaker reduces rapidly at
around 140° whereas the asymmetric exhibits a sharper
drop at around 130°. In general the trends are well
replicated by the wind tunnel results.
4. CONCLUSIONS
A system capable of directly measuring the aerodynamic
loads of a yacht’s flying sails using directional load cells
has been presented. The system has been compared in
full scale testing to results from the FEPV method which
deduces the loads indirectly by integration of surface
pressures over a known shape. The DLC results have
also been compared with wind tunnel results of the same
condition. In both cases the trends are well captured.
There is some discrepancy in the magnitude of the forces
and moments between the FEPC and DLC methods.
It should be borne in mind that the results presented here
are based on a very small dataset. All of the data has
been gathered within 2 hours of testing on the water.
Both systems have shown the ability to the capture
dynamic response of the sails to wind loading. The DLC
system offers a relatively simple method of assessing the
performance of a sail directly without knowledge of
other aspects of the yacht.
Whilst the key variable for sail performance studied here
has been the drive force coefficient, all of the force and
moment coefficients are measured using this process.
These results could be entered directly into a 6 degree of
freedom VPP for full performance analysis.
Acknowledgements
The help of staff and students from the University of
Auckland’s Yacht Research Unit is gratefully
acknowledged for assistance with the full scale testing.
References
1. Bergsma, F., Motta, D., Le Pelley, D.J., Richards,
P.J., and Flay, R.G.J. (2000), "Investigation of
shroud tension on sailing yacht aerodynamics using
full-scale real-time pressure and sail shape
measurements." 22nd International HISWA
Symposium on Yacht Design and Yacht
Construction, HISWA, Netherlands.
2. DeBord, F. Jr, Reichel, J., Rosen, B., and Fassardi,
C. (2002), "Design optimization for the International
America’s Cup Class." In Transactions of the 2002
SNAME Annual Meeting, SNAME, USA.
3. Herman, J.S., (1988). “A Sail Force Dynamometer:
Design, Implementation and Data Handling”,
Massachusetts Institute of Technology, Cambridge,
USA.
4. Brandt, H., and Hochkirch, K. (1999). "Fullscale
Hydrodynamic Force Measurements on the Berlin
Sail-Force-Dynamometer." The 14th Chesapeake
Sailing Yacht Symposium, SNAME, Annapolis,
USA.
5. Milgram, J. H. (1993) “The MIT Sailing
Dynamometer”, Technical report, Massachusetts
Institute of Technology, USA.
6. Masuyama Y, Fukasawa T (1997) “Full-scale
measurement of sail force and the validation of
numerical calculation method”, 13th Chesapeake
Sailing Yacht Symposium, SNAME, Annapolis,
USA.
7. Augier, B., Bot, P., Hauville, F., Durand, M., (2012),
“Experimental validation of unsteady models for
fluid structure interaction: Application to yacht sails
and rigs”. Journal of Wind Engineering and
Industrial Aerodynamics 101(0): 53-66.
8. Viola, I.M., Flay, R.G.J., (2010). “Full-scale
pressure measurements on a Sparkman and Stephens
24-foot sailing yacht”. Journal of Wind Engineering
and Industrial Aerodynamics 98, no.12, 800-807.
9. Lozej, M., Golob, D., Vrtic, B., Bokal, D., (2012),
“Pressure Distribution on Sail Surfaces In Real
Sailing Conditions”. 4th High Performance Yacht
Design Conference, Auckland, New Zealand.
10. Puddu. P., Erriu, N., Nurzia, F., Pistidda, A., &
Mura, A. (2006), “Full-scale investigation of one-
design class catamaran sails”, The 2nd High
Performance Yacht Design Conference, Auckland,
NZ.
11. Flay, R.G.J., Millar, S., (2006), “Experimental
Considerations Concerning Pressure Measurements
on Sails: Wind Tunnel and Full-Scale”, 2nd High
Performance Yacht Design Conference, Auckland,
New Zealand.
12. Le Pelley, D. J., Morris, D., and Richards, P.J.,
(2012), "Aerodynamic force deduction on yacht sails
using pressure and shape measurement in real time.",
4th High Performance Yacht Design Conference,
Auckland, New Zealand.
13. Motta, D., Flay, R.G.J., Richards, P.J., Le Pelley,
D.J., Deparday, J., Bot, P., (2014), “Experimental
Investigation of Asymmetric Spinnaker
Aerodynamics Using Pressure and Sail Shape
Measurements”, Ocean Engineering, 90, 104-118.
14. Le Pelley, D.J., Modral, O., (2008), “VSPARS: A
combined sail and rig shape recognition system
74
using imaging techniques”, 3rd High performance
Yacht Design Conference, Auckland, New Zealand.
15. Premerlani, W., Bizard, P., (2009) “Direction Cosine
Matrix IMU: Theory”, http://gentlenav.googlecode
.com/files/DCMDraft2.pdf
16. Stewart 34 website: http://stewart34.co.nz
17. Motta, D., Flay, R.G.J., Richards, P.J., Le Pelley,
D.J., Bot, P., Deparday, J., (2015), “An investigation
of the dynamic behaviour of asymmetric spinnakers
at full-scale”, 5th High performance Yacht Design
Conference, Auckland, New Zealand.
18. Le Pelley, D., Richards, P.J., (2011), “Effective
Wind Tunnel Tesing of Yacht Sails Using a Real-
Time Velocity Prediction Program”, 20th
Cheasapeake Sailing Yacht Symposium, SNAME,
Annapolis, USA.
19. Boscarino, M. (2014), “Development of an
aerodynamic model for a Stewart 34 Class yacht for
VPP applications”, M.Eng.St. project report,
University of Auckland, New Zealand.
75
... However, the physics of downwind sails is by far more complex than the upwind ones, not only because of the highly turbulent flow, but also because these kinds of sails have an inherent unsteadiness, even in quite stable conditions [7]. This happens mainly because downwind sails are made of very light and flexible cloth, and are attached to the yacht's structure at three points, namely, the head, the clew, and the tack [20]. The shape of a downwind sail is formed by self-generated aerodynamic forces that are strongly affected by the sail shape itself [21]. ...
... However, knowing the real flying shape could also allow for optimizing the mechanical characteristics or the manufacturing process of the sails, in order to improve their stability, to reduce their weight, and to optimize their mechanical behavior [18,22,26,27]. Many papers have focused on evaluating the real shape of sails only to predict sailing performances [20,22,25,28]. However, to the best of our knowledge, no relevant papers have tried to measure the flying shape and the loads of a sail, taking into account both the point of view of sailors, whose primary interest is the sail performance, and of the sailmakers, who are also responsible for suitably dimensioning and manufacturing the sail. ...
... With reference to the loads calculated numerically with the proposed methodology, good levels of agreement were found with other case studies in the literature. In particular, it can be observed that the ratios of the loads on the sail corners and the trends of the forces, when AWA varies, are very similar to those found in other studies [2,20,34]. For example, the force ratios F clew /F head and F tack /F head calculated at AWA = 90 • using the new procedure are about 48% and 81%, respectively; the same ratios have been experimentally measured by Deparday et al. [2] for a similar case study and are equal to about 51% and 85%, respectively. ...
Article
Full-text available
The process of designing a sail can be a challenging task because of the difficulties in predicting the real aerodynamic performance. This is especially true in the case of downwind sails, where the evaluation of the real shapes and aerodynamic forces can be very complex because of turbulent and detached flows and the high-deformable behavior of structures. Of course, numerical methods are very useful and reliable tools to investigate sail performances, and their use, also as a result of the exponential growth of computational resources at a very low cost, is spreading more and more, even in not highly competitive fields. This paper presents a new methodology to support sail designers in evaluating and optimizing downwind sail performance and manufacturing. A new weakly coupled fluid–structure interaction (FSI) procedure has been developed to study downwind sails. The proposed method is parametric and automated and allows for investigating multiple kinds of sails under different sailing conditions. The study of a gennaker of a small sailing yacht is presented as a case study. Based on the numerical results obtained, an analytical formulation for calculating the sail corner loads has been also proposed. The novel proposed methodology could represent a promising approach to allow for the widespread and effective use of numerical methods in the design and manufacturing of yacht sails.
... Puisque la voile n'est tenue qu'à ses 3 points, connaître la direction et la norme des efforts à chacun de ces points permet de calculer la force aérodynamique totale créée par la voile. Ces DLCs ont été développées au Yacht Research Unit de l'Université d'Auckland [Le Pelley et al., 2015]. ...
Thesis
Full-text available
A full-scale experimental study on an instrumented sailing yacht is conducted to better assess the aero-elastic behaviour of the sails and rigging in downwind navigations. The downwind sail shape is a non-developable surface with high curvature leading to massive flow separation. In addition, spinnakers are thin and flexible sails leading to a strongly coupled Fluid-Structure Interaction. Due to the non-respect of some rules of similitude, the unsteady behaviour of downwind sails cannot be easily investigated with wind tunnel tests that would need comparison with full-scale experiments. Moreover unsteady numerical simulations modelling the aero-elastic behaviour of the sails and rigging require validations. An inboard instrumentation system has been developed on a 8 meter J/80 sailboat to simultaneously and dynamically measure the flying shape of the spinnaker, the aerodynamic loads transmitted to the rigging, the pressure distribution on the sail as well as the boat and wind data. The shape of the spinnaker while sailing is acquired by a photogrammetric system developed during this PhD. The accuracy of this new system, better than 1.5%, is used to measure the global shape and the main dynamic deformations, such as the flapping of the luff. The aerodynamic load produced by the spinnaker is assessed by the measurements of the load magnitudes and directions on the three corners of the sail (head, tack and clew), and also by the pressure distribution on the spinnaker. The global behaviour of the spinnaker is analysed according to the apparent wind angle. A new representation using Bézier triangular surfaces defines the spinnaker 3D shape. A few control points enable to represent the sail and can easily characterise the type of sail. A typical unsteady behaviour of the spinnaker is also analysed. Letting the luff of the sail flap is known by sailors as the optimal trim but has never been scientifically studied before. It is found that it is a complex three dimensional fluid-structure interaction problem where a high suction near the leading edge occurs, producing a temporary increase of the force coefficient that would not be possible otherwise.
Article
Ground contact and force detection sensors are of critical technical importance in the realization of articulated robots capable of navigating unknown or challenging terrains. Current force sensors for robotic ground contact have limitations including being too heavy, vulnerability to inertial noise, and lacking robustness and redundancy against failure. In this work, we present the design, laboratory testing, and flight testing of a novel force sensor which addresses the aforementioned difficulties. The force sensor presented here is based on a custom elastomer dome with an engineered air cavity adhered to a barometric pressure sensor. Uniquely, we may tailor the sensitivity and range of the sensor by engineering the elastomer shape as well as the air cavity within it. Further, the large deformations incurred by the sensor are exploited to design a sensor enclosure which transfers large loads to structural members rather than onto the force sensor itself. An experimentally validated finite element model is developed to numerically predict the response of the sensor under an applied load. The model is used to perform trade studies on the geometry and material properties of the elastomer dome with a focus on developing design rules for this new type of sensor. With these simple design rules and finite element model, we design, manufacture, and test these new sensors as individuals and within arrays specifically for use on a rotorcraft robotic landing gear.
Article
Full-text available
In the present study we investigate the flapping instability of a light, soft, highly cambered membrane subject to wind loading. An original in-situ experiment is developed where time-resolved pressures and forces are measured on a full-scale yacht downwind sail called a spinnaker. Particular features of this aero-elastic problem are that the membrane is weakly constrained-held only at three corners-, has a free leading edge, and has no proper shape in the absence of wind loading. In usual operating conditions, the soft structure is subject to a flapping instability giving rise to almost periodic folding and unfolding of the fore part of the sail, associated to strong variations of pressures and forces. This dynamic behavior is analyzed in detail and the space-time evolution of pressures on the membrane is linked to the flapping phenomenon. A peak in forces is observed when the membrane recovers its full shape. Thanks to the Bi-Orthogonal Decomposition (BOD) applied to the pressure fields, the dynamic behavior is reasonably well represented by the two first modes where mode 1 mostly carries the global aerodynamic force behavior and mode 2 mostly represents the effects of the membrane flapping. A physical mechanism of the flapping process is proposed based on the discussion of aerodynamic pressures and strains in the membrane.
Article
Full-text available
While sailing offwind, the trimmer typically adjusts the downwind sail “on the verge of luff- ing”, occasionally letting the luff of the sail flap. Due to the unsteadiness of the spinnaker itself, main- taining the luff on the verge of luffing requires continual adjustments. The propulsive force generated by the offwind sail depends on this trimming and is highly fluctuating. During a flapping sequence, the aerodynamic load can fluctuate by 50% of the average load. On a J/80 class yacht, we simultane- ously measured time-resolved pressures on the spinnaker, aerodynamic loads, boat data and wind data. Significant spatio-temporal patterns were detected in the pressure distribution. In this paper we present averages and main fluctuations of pressure distributions and of load coefficients for dif- ferent apparent wind angles as well as a refined analysis of pressure fluctuations, using the Proper Orthogonal Decomposition (POD) method. POD shows that pressure fluctuations due to luffing of the spinnaker can be well represented by only one proper mode related to a unique spatial pressure pattern and a dynamic behavior evolving with the Apparent Wind Angles. The time evolution of this proper mode is highly correlated with load fluctuations. Moreover, POD can be employed to filter the measured pressures more efficiently than basic filters. The reconstruction using the first few modes makes it possible to restrict the flapping analysis to the most energetic part of the signal and remove insignificant variations and noises. This might be helpful for comparison with other measurements and numerical simulations.
Conference Paper
Full-text available
While sailing offwind, the trimmer typically adjusts the downwind sail "on the verge of luffing", letting occasionally the luff of the sail flapping. Due to the unsteadiness of the spinnaker itself, maintaining the luff on the verge of luffing needs continual adjustments. The propulsive force generated by the offwind sail depends on this trimming and is highly fluctuating. During a flapping sequence, the aerodynamic load can fluctuate by 50% of the average load. On a J/80 class yacht, we simultaneously measured time-resolved pressures on the spinnaker, aerodynamic loads, boat and wind data. Significant spatio-temporal patterns are detected in the pressure distribution. In this paper we present averages and main fluctuations of pressure distributions and of load coefficients for different apparent wind angles as well as a refined analysis of pressure fluctuations, using the Proper Orthogonal Decomposition (POD) method. POD shows that pressure fluctuations due to luffing of the spinnaker can be well represented by only one proper mode related to a unique spatial pressure pattern and a dynamic behavior evolving with the Apparent Wind Angles. The time evolution of this proper mode is highly correlated with load fluctuations. Moreover, POD can be employed to filter the measured pressures more efficiently than basic filters. The reconstruction using the first few modes allows to restrict to the most energetic part of the signal and remove insignificant variations and noises. This might be helpful for comparison with other measurements and numerical simulations.
Conference Paper
Full-text available
Since the introduction of the International America's Cup Class for the 1992 America's Cup, more than sixty-five yachts have been designed to the class rule, and most of these have raced for the opportunity to either challenge or defend the America's Cup. The relatively large differences in performance exhibited by these boats illustrate the complexity of the design optimization problem. This paper defines the design problem from a practical Naval Architecture standpoint, reviews engineering methods currently employed to evaluate design trade-offs, and provides insight into how the various analytical and experimental tools available to the designer are used. Constraints of the class rule are analyzed in terms of general limitations and specified quantitative relationships between the major design parameters. Methods used to select major design parameters, including the designer's experience and use of velocity prediction programs are discussed, and limitations of these techniques are identified. Experimental and analytic techniques typically used to provide engineering data for the designer, and to generate input for performance modeling, are reviewed. The experimental techniques considered include tank tests, wind tunnel tests and full-scale testing. Analytic techniques include potential and viscous flow methods for evaluating hydrodynamic drag and lift, wave making drag, and seakeeping performance. The current state-of-the-art for each of these techniques is discussed, their applicability and limitations are reviewed, and specific examples of their use are provided. Finally, recommendations are formulated on how these various design tools and techniques should be used together in an integrated design program to optimize performance.
Conference Paper
Full-text available
This paper presents new results obtained from analysing on-the-water pressure, shape, force, speed and direction data that were obtained from experiments in the Hauraki Gulf in Auckland, New Zealand in April 2014. A fully instrumented Stewart 34 Class yacht sailing downwind was used for the tests. Details of the analysis of the results from simultaneous time-resolved measurements of pressure, sail shape and loads are presented. The dynamic behaviour of the fluid-structure system made up of a light sail cloth and highly curved flow is investigated. Aerodynamic forces on the asymmetric spinnaker are determined from the combination of point pressure measurements across the sail with simultaneous shape measurements. Simultaneous time histories show a strong correlation between the variations of pressure distributions, flapping sail shape and the forces at the corners. Periodic curling and filling of the spinnaker luff influences suctions, in particular at the leading edge, and forces, which can dynamically change on the order of 40- 50%. The results are similar to, and extend, those that were presented by the authors at the 2013 Innov’sail Conference in Lorient, France. It is expected that the results from this work will give reliable benchmark data which may be used to validate unsteady fluidstructure interaction numerical simulations of downwind sails.
Article
Full-text available
Sail forces were measured in a full-scale sailing boat with the use of a sail force dynamometer. This apparatus consisted of an aluminum frame fixed to the hull by way of several load cells. The sailing boat was modified so that the dynamometer frame could be installed inside the hull. The mast, stays, winches, and other sailing rig were fixed on the frame so as to transmit all the forces acting on sail to the frame. By transforming the measured forces, the lift force, drag force, thrust, side force, or the center of effort of the sail force could be obtained. The sailing conditions of the boat, such as the boat speed, heel angle, wind speed, wind angle, and so on, were also measured.Sail shapes of the boat in the up-wind condition were also measured with the use of CCD cameras installed in the boat. The sail shape images taken by the cameras were transformed to bit-map files, and then processed by an SSA-2 D, a sail shape analyzing software. With the use of this software, sail shape parameters were obtained. The relationship between the measured sail forces and the sail shape parameters is discussed in this paper. Moreover, the measured sail shapes were used as the input data for the numerical calculations.Numerical calculations were performed to estimate the sail forces of the boat. In the calculations, two sails, a mainsail and a jib, were modeled in the form of a vortex lattice. The vortex lattice method was adopted as the numerical calculation method. Step by step calculations were conducted up to attaining the steady state of the sail in steady wind. Calculated sail forces were compared with the measured forces, and the validity of the numerical method was studied.
Article
Full-text available
This is the first of a pair of papers on the theory and implementation of a direction-cosine-matrix (DCM) based inertial measurement unit for application in model planes and helicopters. Actually, at this point, it is still a draft, there is still a lot more work to be done. Several reviewers, especially Louis LeGrand and UFO-man, have made good suggestions on additions and revisions that we should make and prepared some figures that we have not included yet. We will eventually incorporate their suggestions, but it may take a long time to get there. In the meantime, we think there is an audience who can benefit from what we have so far. The motivation for DCM was to take the next step in stabilization and control functions from an inherently stable aircraft with elevator and rudder control, to an aerobatic aircraft with ailerons and elevator. One of the authors (Premerlani) built a two axes board several years ago, and developed rudimentary firmware to provide stabilization and return-to-launch (RTL) functions for a Gentle Lady sailplane. The firmware worked well enough, and the author came to rely on the RTL feature, but it never seemed to work as well as the author would like. In particular, satisfactory solutions to the following two issues were never found: • Mixing. It was recognized that in a banked turn, there were two problems arising from the bank angle. First, the yaw rotation of the aircraft around the turn generated a nuisance signal in the pitch gyro, because of the banking. Second, in order to make a level turn, the elevator needed some "up" deflection. The amount of deflection depends on the bank angle, which could not be directly measured. Both issues were opposite sides of the same coin.
Conference Paper
Sail forces were measured in a full-scale sailing boat with the use of a sail force dynamometer. This apparatus consisted of an aluminum frame fixed to the hull by way of several load cells. The sailing boat was modified so that the dynamometer frame could be installed inside the hull. The mast, stays, winches, and other sailing rig were fixed on the frame so as to transmit all the forces acting on sail to the frame. By transforming the measured forces, the lift force, drag force, thrust, side force, or the center of effort of the sail force could be obtained. The sailing conditions of the boat, such as the boat speed, heel angle, wind speed, wind angle, and so on, were also measured. Sail shapes of the boat in the up-wind condition were also measured with the use of CCD cameras installed in the boat. The sail shape images taken by the cameras were transformed to bit-map files, and then processed by an SSA-2D, a sail shape analyzing software. With the use of this software, sail shape parameters were obtained. The relationship between the measured sail forces and the sail shape parameters is discussed in this paper. Moreover, the measured sail shapes were used as the input data for the numerical calculations. Numerical calculations were performed to estimate the sail forces of the boat. In the calculations, two sails, a mainsail and a jib , were modeled in the form of a vortex lattice. The vortex lattice method was adopted as the numerical calculation method. Step by step calculations were conducted up to attaining the steady state of the sail in steady wind. Calculated sail forces were compared with the measured forces, and the validity of the numerical method was studied.
Conference Paper
At the Institute of Naval Architecture, Marine and Ocean Engineering (ISM), Technical University of Berlin, a research project was initiated investigating prediction methods for hydrodynamic forces acting on sailing yachts. For a 33ft. sailing yacht model tests with various keel designs and RANSE calculations have been carried out. In order to verify both, a full scale sail-force­dynamometer was built which enables to record the resulting hydrodynamic forces in a seaway without scaling parameters. In addition, separate six and five component balances have been designed for measuring the forces acting on keel and rudder for the model tests as well as on the full scale boat. The design and construction of the full scale measuring device is presented. Along with some sample measurements a general calibration procedure for inclined multicomponent balances is proposed. A short review of the model tests and RANSE-calculations complement the presentation.
Article
This work is concerned with an experimental campaign carried out on a Tornado class catamaran. Full scale pressure measurements were taken on the boat mainsail both on the pressure and on the suction sides. Pressure taps were applied on the main sail accurately trying not to modify the flow field nor the flexibility of the sails. The distribution of the pressure taps is such that an accurate map of the mean pressure distribution on the sail surfaces has been reconstructed. Every pressure tap was coupled with a pressure transducer and the pressure values were acquired through an electronic scanning system, so that they can be considered to be simultaneous. Static pressure coefficient distribution over the sail span will be presented, giving an insight into the flow structure over the sails and precious information for validation of CFD codes.
Article
This paper presents a review of some of the lessons learnt by the YRU in taking pressure measurements on sails. It has been found that it is technically quite difficult to take multiple pressure measurements on a single surface sail. This is because the presence of the sensors and tubing can influence the flow field and hence the surface pressures. In addition, it is quite difficult to have multiple sensors, as the tubes and cables need to be taped to the sail, and they can be damaged during the sailing operation. It is concluded that whereas an electronically scanned pressure system can be used to obtain pressures on model sails in the wind tunnel, it is fairly impractical to use such systems in full scale, and it is recommended that wireless systems be used.
Article
An innovative method combining simultaneous on-water pressure and sail shape measurements for determining aerodynamic forces produced by sails is described and used on Stewart 34 and J80 Class yachts flying asymmetric spinnakers. Data were recorded in light and medium winds in order to check the reliability, accuracy and repeatability of the system. Results showed similar trends to the published literature. The accuracy of the system was investigated by wind tunnel tests, with determination of the entire sail shape from the stripe images recorded by the camera-based (VSPARS) system, and was found to be relatively good. Generally the pressure distributions show a leading edge suction peak, occurring at 5 to 10% of the chord length, followed by a pressure recovery and then a suction increase due to the sail curvature, with finally a reduction in suction near the trailing edge. The drive force coefficient measured on the Stewart 34 is lower than for the J80 because of a non-optimal sail shape due to light winds. On a reaching course, the standard deviation of the pressure signals was largest near the luff, reducing in the stream-wise direction, while it was high on the entire sail section when sailing on a running course.