Development of a New System for Wind Tunnel and Full Scale Testing

PRESSURE MEASUREMENTS ON YACHT SAILS: DEVELOPMENT OF A NEW SYSTEM FOR WIND TUNNEL AND FULL SCALE TESTING

Fabio Fossati Department of Mechanical Engineering, Politecnico di Milano, Milano, Italy Ilmas Bayati Department of Mechanical Engineering, Politecnico di Milano, Milano, Italy

Sara Muggiasca Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 Department of Mechanical Engineering, Politecnico di Milano, Milano, Italy Ambra Vandone Department of Mechanical Engineering, Politecnico di Milano, Milano, Italy Gabriele Campanardi Department of Aerospace Science and Technology, Politecnico di Milano, Milano, Italy Thomas Burch CSEM, Centre Suisse d’Electronique et de Microtechnique SA, Alpnach Dorf, Switzerland Michele Malandra North Sails Group, Carasco, Italy Manuscript received January 30, 2017; revision received March 8, 2017; accepted April 17, 2017.

Abstract: The paper presents an overview of a joint project including Politecnico di Milano, CSEM and North Sails, aiming at developing a new sail pressure measurement system based on MEMS sensors (an excellent compromise between size, performance, costs and operational conditions) and pressure strips and pads technology. These devices were designed and produced to give differential measurement between the leeward and windward side of the sails. The research has been developed for the final employment on the Lecco Innovation Hub Sailing Yacht Lab, a 10 m length sailing dynamometer which aims at being the reference contemporary full scale measurement device in the sailing yacht engineering research field, to enhance the insight of sail steady and unsteady aerodynamics. The pressure system is described in details, the data acquisition process and system metrological validation are reported; furthermore, some results obtained during a wind tunnel campaign carried out at Politecnico di Milano Wind Tunnel, as a benchmark of the whole measuring system for future full scale application, are reported and discussed in details. Moreover, the system configuration for full scale testing, which is being finalized at the time of this paper, is also described Keywords: Sail aerodynamics, wind tunnel, full scale, pressure measurements, upwind sailing

NOMENCLATURE !"! Apparent wind angle [deg] # Length of the instrumented section [m] $%& Longitudinal position of the center of effort [-] $%ℎ Height of the center of effort [-] $( Pressure coefficient [-] $) Driving force coefficient [-] $+ Heeling force coefficient [-] ,) Driving force [N] ,+ Heeling force [N] ℎ-./0 Mast height 1 Tube length [m] 1"1 Length water line [m] Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 2) Heeling moment [Nm] 23 Yaw moment [Nm] 4 Actual Measured Pressure [Pa] 45 Reference static pressure [Pa] 6 Incoming/apparent wind speed [m/s] 7 Pressure tap position along the instrumented section [m]

8 Thickness of the wind tunnel boundary layer [m] 9 Air density [kg/m^3]

L1 L2 L3 Instrumented sails sections at 25%, 50% and 75% of the sail heights ORC Offshore Racing Congress TOF Time of Flight VPP Velocity Prediction Program

INTRODUCTION The possibility of knowing the effective pressure distribution over the sail plan is of great interest for the aerodynamic and structural design of sails and for the selection and the optimal use of materials and production techniques. Integral measurements alone may not be sufficient for understanding how a sail plan can be optimized for specific purposes, if any information about the complex local fluid-structure interaction are provided. In the last few years there has been a revival of pressure measurements on yacht sails and recently several contributions can be found in literature aiming to assess sail pressure distribution detection (Viola et al. 2012, Deparday et al. 2013, Motta et al. 2015, Pot et al. 2014, Viola et al. 2011a, Lozej et al. 2012, Le Pelley et al. 2012, Motta et al. 2014). The present paper presents an overview of an ongoing joint project developed among Politecnico di Milano, CSEM and North Sails aiming at assessing a new sail pressure measurement system based on MEMS pressure transducers, connected to strips and pads. These devices were designed and produced with the scope to provide differential measurements between the leeward and windward side of the sails. The project has been developed within the Lecco Innovation Hub Sailing Yacht Laboratory project, a 10 m length sailing dynamometer which is hoped will be the modern reference full scale measurement setup in the sailing yacht engineering research field, in order to enhance the insight of sail steady and unsteady aerodynamics.

An overview of the Sailing Yacht Lab project is provided in (Fossati et al. 2015a): a brief summary of the origin and building steps of the vessel’s design are given, along with a description of principal design and performance criteria. Also the project management and commissioning are described, as well as the measurement capabilities and data acquisition procedure. Furthermore, an important feature of this project is the availability of measurement systems for pressure distribution acting on the sails at full scale. In the following, the pressure measurement system is described in detail, as well as the data acquisition process and system metrological validation is provided. The pressure measurement system has also been tested in the wind tunnel using a scale model of a sailing yacht and compared with a different pressure measurement system already available at Politecnico di Milano Wind Tunnel. For wind tunnel tests strips and pads adequate for the model sails were used. Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 Some results obtained during a wind tunnel campaign carried out during an Offshore Racing Congress (ORC) project aimed at revising sail aerodynamic coefficients and the VPP (Velocity Prediction Program) aerodynamic model are reported and discussed in detail. In conclusion, the pressure measurement system designed for full scale testing is described.

PRESSURE MEASUREMENT SYSTEM The pressure distribution on the sails is carried out by means of MEMS sensors (an excellent compromise between size, performance, costs and operational conditions) and dedicated pressure strips and pads which have been designed and produced aiming to provide both differential measurement between the sail leeward and windward side and actual pressure on each side when a reference pressure is available as during wind tunnel tests. The pressure sensors are designed and built to provide the differential pressure measurement with the possibility to select the pressure reference signal. In the pressure measurement system design lightness and flexibility were considered a priority even accepting lower level of accuracy. The idea was also to identify a system usable both for wind tunnel and for full scale tests introducing few adjustments. In the following a detailed description of the pressure scanners will be provided, as well as of the other main components of the system. Pressure strips scanner description The scanner CSEM C16 is a miniaturized electronic pressure scanner in a slim, lightweight and waterproof package (Figure 1). It provides 16 differential pressure sensors and a CAN bus interface for the communication. High attention was given to dimensions and shape of the scanner box. The scanner height, of only 6 mm, has minimal impact on the airflow, which makes it possible to place the scanner directly in a custom-built sleeve close to the actual measurement section on the sails. Each of the 16 sensors has its own reference input which makes the scanner especially suited for measuring the pressure difference between leeward and windward side on dedicated spots on the full scale sails. The commercial MEMS pressure dies, integrated in the scanner, are a new generation of piezo-resistive differential low-pressure dies to reach very low full scale ranges below 1000 Pa. Despite the die size of only 2 mm x 2 mm x 0.5 mm, that is much smaller than traditional low-pressure dies, it provides improved zero-stability, reduced g-sensitivity and reduced sensitivity to humidity. This added stability permits use with added amplification to

achieve accurate performance in ranges much lower than its nominal 1000 Pa rating. The key specification of the scanner is given in Table 1. The MEMS sensors are cost efficiently bonded to a FR4 substrate using innovative die bonding techniques based on elastic adhesives (Figure 4). The sensors are packaged in a sensor array with minimal air cavity to ensure optimal performance in combination with the micro-channels of the pressure strips. A dedicated pressure flange system makes the scanner compatible with either the pressure strips or with standard tubing. Three different pressure adapters have been developed, which can be screwed to the scanner. The first adapter provides 32 tubes (2 per pressure sensor, one facing to the front side and one to the reference side of the sensor). A second adapter combines all reference inputs to a single tube, in order to connect all MEMS sensors to the same reference, cf. Figure 2 (a) and (b). Finally, the third adapter provides direct access to the pressure strips without the need of any tubes. Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Table 1 - Pressure Scanner Specification Parameter C16 Unit FS pressure range ±1000 Pa Number of pressure inputs 16 Number of reference inputs 16 Measurement resolution 0.01 % FS Static accuracy after zeroing 0.25 % FS Total thermal error 0.01 %FS/°C Sample rate 1 - 100 Hz Input voltage 12 V Operation current 60 mA Communication CAN Interface 1 Mbit / s Maximal CAN cable length 40 m Internal flash data memory size 8 Mbit Operating temperature range -10 to 70 °C Size 65x55x6 mm Weight 50 gram Parameter C16 Unit FS pressure range ±1000 Pa Number of pressure inputs 16 Number of reference inputs 16 Measurement resolution 0.01 % FS Static accuracy after zeroing 0.25 % FS Total thermal error 0.01 %FS/°C Sample rate 1 - 100 Hz Input voltage 12 V Operation current 60 mA Communication CAN Interface 1 Mbit / s Maximal CAN cable length 40 m Internal flash data memory size 8 Mbit Operating temperature range -10 to 70 °C Size 65x55x6 mm Weight 50 gram

④ ②

①

③

Figure 1 - Pressure Scanner C16 with auxiliary parts. 1) Scanner, 2) CAN Cable, Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 3) CAN connector, 4) Tube Adapter

(a) Pressure tube adapter 2 x 16 tubes (left) (b) Pressure flange with gasket and threads to and 1 + 16 tubes (right) connect and seal pressure tube- or strip-adapters

Figure 2 - Pressure system details

The scanner C16 supports a standard CAN interface (CAN 2.0A) with a proprietary CAN protocol, allowing remote access to the essential commands required when integrating the unit into an instrumentation system. The serial CAN bus topology ( Figure 3) allows for up to 128 scanners in a single network. In practice, the number of scanners per network should be below 16 (i.e. 256 pressure sensors) in order to reduce the data traffic on the bus and to guarantee synchronized data sampling. The CAN interface has been preferred over a wireless solution due to its robust data transmission capability, the guaranteed data rate of 1 Mbit per second and the possibility to directly supply electrical power to the scanners via the flat CAN cable. Thus, no battery is required in the scanner which reduces both, the dimensions and the overall weight of the scanners. All measurement data and configuration commands are sent over the CAN interface. A correctly received command is always acknowledged by the scanner with the transmission of a response message. Two basic data sampling approaches are supported either autonomous sampling or master sampling. The desired option can be configured and stored in the configuration flash. In auto

sampling mode each scanner in the network generates its own sample timing according to a programmed sample rate and transmits the measurement data of each sample to the CAN bus autonomously. The measurement data can be collected on-line or can be stored in the internal flash of the scanner and downloaded off-line after the measurement session. In master sampling mode the user programmed instrumentation system (SW running on PC or Laptop) acts as sample master and broadcasts each sample start with a SINGLE_SHOT sample command. All scanners in the network receive the sample command at the same time and start the measurement immediately and synchronously. Each scanner writes the measurement data to the CAN bus following a bus collision avoidance protocol. The master collects the response messages of all active scanners in the CAN network, and initiates the next sample according to the desired sample rate. The master sampling mode has the advantage that all scanners connected to the CAN bus are synchronized by the master, even over a long sample period of several hours. The 16 sensors of each scanner are sampled sequentially with an internal scan rate of up to 4 kHz. Hence in master sampling Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 mode all sensors in the CAN network can be sampled nearly synchronous within 4 ms.

Figure 3 – CAN bus topology

Figure 4 - Pressure scanner electronics with MEMS sensors bonded to FR4 substrate Pressure strip technology The pressure strip system is suited for aerodynamics testing on full scale objects in their natural environment but also for models in a wind tunnel (Figure 5). Its main advantage is the light weight and thin, flexible foil appearance which allows non-invasive application to the test surface. The pressure strips are made of thin polymer films and the strip geometry can be customized for nearly seamless fitting to the test object (Figure 6). Tiny micro- channels in the pressure strip propagate the pressure from the tap to the connected pressure scanner. Manufacturing processes have been developed successfully using laser

and micro-milling to produce strips with comparatively deep channels. Laser fabrication has the advantage that it can produce channels in soft materials such as silicone or soft PVC, thus increasing the flexibility of the strip significantly without having to reduce the thickness of the strip. On the other hand, channels can be manufactured approximately 5 times faster using micro-milling (Figure 5). The base material with milled or laser ablated channels is laminated with transparent adhesive tape in order to obtain sealed channels.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 5 - 400 μm wide and 400 μm deep laser fabricated micro-channels in a silicone strip (left) and micro-milled in a polycarbonate strip (right).

Figure 6 – Pressure strip with 40 taps on three sections tailor-made for the jib of a 1:10 scale model yacht

Pad technology Pressure pads are a specific form of small pressure strips and provide a simple solution to attach a pressure tube to a very thin structure like a spinnaker. They are very useful also for full scale tests where, due to the dimension of the sail, it is preferable to place a small pad

and to realize the pneumatic connection to the scanner through small tubes. The pads can be individually placed on the test section and connected to the pressure scanner with small plastic tubing (Figure 7). The pads provide two pressure taps on one end of the pad and a pressure tube adapter with two metal tubes of 1 mm diameter on the other end. Each of the two taps faces to one side of the sail (leeward or windward). A small hole of 0.8 mm in diameter is made in the sail directly beneath the respective pressure tap to realize a pressure passage to the opposite side of the sail. The pad thickness is between 0.5 and 1 mm and therefore introduces only minor interference to the airflow. The length of the pad depends on the size of the test object. For model sails the length is as short as 30 mm while for real sails the length is up to 150 mm to keep the tube adapter a certain distance out of the air flow of the measurement section (Figure 7). The micro channels are cut in the base layer using the same laser ablation and micro milling techniques as for the pressure strips. The base layer is made of a soft material like PVC soft, silicone or acrylic foam tape and usually has a thickness of Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 0.3 - 0.5 mm. The cover layer is a transparent adhesive tape of 0.2 mm. The cover layer overlaps the base layer by a few millimeters which results in a smooth transition between the sail and the pad layers after application on the sail.

Figure 7 - Conceptual drawing of the pressure pad

METROLOGICAL VALIDATION Some preliminary tests were performed to verify the measurement quality of the system in terms of static accuracy of the pressure measurements and dynamic response of pressure strips. Static accuracy The analysis of the static accuracy of the CSEM pressure measurement system was carried out in the 1 m x 1.5 m test section of the close circuit wind tunnel of the Aerodynamics Laboratory of Politecnico di Milano, using a constant section NACA 23015 airfoil model. The model has 0.3 m cord length, aspect ratio 3.1 and it is instrumented with pressure taps along

the mid-span section. A dedicated pressure strip was installed spanning from the 25.5% chord position of the lower surface to the 87.5% chord position of the upper surface of the model (Figure 8). The strip is provided with a double series of pressure taps (Figure 8): at the same chord position of each pressure tap on the strip another hole through the strip was created to be connected to proper tubes, so that comparative measurements could be taken both for the novel pressure system with strip micro-channels (CSEM) and the consolidated high accuracy pressure scanner system (Figure 8). More specifically, the latter relies on an Esterline Pressure Systems DTC ESP miniature pressure scanner with 1 PSI range, controlled by a Chell QUADdaq System.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 8 - NACA23015 airfoil model instrumented with pressure strip for static wind tunnel test Static measurements were gathered for fixed incidence angles, ranging from -2 degrees to 14 degrees, at respectively 2.5 m/s, 5 m/s, 10 m/s, 15 m/s and 20 m/s wind speed. For the sake of clearness, in Figure 9 the results for the wing model at different angles of incidence, are reported. The data plotted in Figures 10 and 11 are reported in terms of pressure coefficient, defined as 4 − 4 $ = 5 1 ( 1 96> 2 ? where 4 is the pressure on pressure tap, 45 the reference pressure, 9 the air density and 6 the wind speed of the incoming flow. It can be noticed that there is good agreement between pressure tap data and the pressure strip data, except for a few points near the airfoil leading edge, where the airfoil has strong curvature. In this region the strip installation, even though it was executed with particular care, presents some tiny surface deformations - visible as small air bubbles around the pressure taps next to the leading edge - which are the main source of the differences. In Figure 10 the pressures measured for an angle of incidence equal to 0° by both systems are reported more in detail (the pressure taps are numbered in clockwise direction): in addition to the mean pressure coefficients an indicator of the 5% deviation from the reference system is reported. It can be noticed that the differences are in general lower than 5%, higher deviations are observed in the points placed where the installation of the strip was not perfectly executed. This deviation level was considered acceptable in order to have a system

as light and flexible as possible. A similar comparison was performed also during the tests on the sail plan: one section of the mainsail was both connected to the new system and to the reference system (see Figure 11 where 7 is the position along the section and # the length of the sail section): also in this case, the agreement between the two systems was considered acceptable.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 9 - Pressure distribution along the airfoil for different angles of incidence at @=20 m/s

Figure 10 - Comparison between the two pressure measurement systems

for 0° of angle of attack: a 5% error band is indicated Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 11 - Comparison between the two systems for level L2 of the mainsail (leeward side) at an apparent wind angle of 20°

Dynamic response Further experimental investigation was also carried out to better understand the dynamic capabilities of the pressure strips. A truck hooter has been utilized as pressure wave generator, driven by a signal generator and an amplifier. The pressure measurements were taken by the means of two CSEM pressure scanners: one with a pressure port connected directly to the pressure wave source by a very short tube (Figure 12), the second with a pressure port connected to the strip channel under test in the same way to be used during the wind tunnel testing, Figure 6.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 Figure 12 - Test setup for dynamic response evaluation A special attention was paid to the pressure connections to the source. Two tubes of the same length were adopted to connect the scanner and the strip channel to be tested, Figure 12. In such a way it can be reasonably assumed that the pressure wave measured near the source has the same amplitude and phase of the pressure wave reaching the pressure tap on the strip. The connection and the sealing of the tube on the strip was done by means of modeling clay. Measurements were carried out on the pressure taps connected with the longest channel of each pressure tap array (bottom, middle, top) both on the mainsail and the jib strip (red circles in Figure 13). The tests were conducted generating single tone sinusoidal pressure waves and sinusoidal sweeps in the frequency range 0 - 3 Hz the expected frequency range for this phenomenon. The pressure data acquisition was started simultaneously on the two scanners. In Figure 14 are reported the results obtained for the pressure tap on the top array of the mainsail both with single tone and sweep excitation. It is possible to note a good agreement between the two sets of data. The obtained transfer function highlights that the pneumatic connection between the measurement point ant the sensor acts like a low pass filter reducing the signal amplitude as a function of the frequency. On the other hand, the linear trend in the phase means a constant shift of the signal in time. The precise definition of the transfer function of each channel is useful in order to correct the time histories during the post-processing procedures. In order to understand how the length of the connection affects the transfer function, in Figure 15 the results obtained for three different channels are reported. In the transfer function definition, geometrical characteristics of the channel section also have an influence; nevertheless, in this case all the channels have the same section dimensions. The experimental transfer functions were then compared with the numerical ones obtained from the analytical approach described in Tijdeman et al. 1965: numerical transfer functions are calculated using a model that takes into account tube dimensions and the presence of connections and that assumes laminar flow in the tubes. The comparison is reported in Figure 16 for a mainsail channel: a good agreement can be observed. The frequency range chosen for the characterization of the pressure system is consistent with the interest of investigating the physics of slow varying aerodynamic phenomena connected to the sailing yacht motion, due to the combined wind and wave loading. The

cutoff frequency for wind tunnel tests in order to reproduce the same reduced frequency of the full-scale phenomenon is approximately 2 Hz (Fossati et al. 2013). Frequencies higher than this range (e.g. turbulence) are not expected to have any relevant influence on the overall dynamics of the boat, in that it represents a mechanical low pass filter. For the full scale system, a similar characterization will be planned considering the tubes and the connections used.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 13 – Positions of the pressure taps tested

Figure 14 - Dynamic response of the pressure tap on the top array of the mainsail

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 15 - Dynamic response for different channels

Figure 16 - Comparison of the dynamic response with numerical transfer function (Mainsail, A = B. DE F)

WIND TUNNEL TESTS The new pressure measurement system was tested during a wind tunnel experimental campaign together with the standard instrumentation used for sail plan characterization. In the following, preliminary results are reported. The main goal of this test session was to check the capability of the system, further tests will be planned for a more systematic study.

Test apparatus, program, and procedure The experimental campaign was performed in the Boundary Layer test section of the Politecnico di Milano wind tunnel. For this section the boundary layer thickness 8 is about 0.2 m, defining 8 as the height where the wind velocity is equal to 99% of the undisturbed flow velocity. The maximum velocity deviation across the section is less than 3%. The turbulence intensity indexes outside the boundary layer are just below 2%. The average along wind integral length scale is evaluated equal to 0.15m. The tested model was a complete 1:10 scale model of a 48’ cruiser-racer, consisting of a yacht hull body (above the waterline) with deck, mast, rigging and sails (see Figure 18). The model was installed on the wind tunnel turntable in order to change !"! (apparent wind angle) during the tests. The large size of the section enables yacht models of quite large size to be used, so that the sails are large enough to be made using normal sail making techniques. Moreover, the Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 model can be rigged using standard model yacht fittings, commercially available, that can be used in order to trim the sails as in real operating condition. The sheet trims are controlled by the sail trimmer who operates from the wind tunnel control room with a 7 multi-turn control knobs that allow winch drum positions to be recorded and re-established if necessary. The sail plan tested during the experimental campaign is reported in Figure 17 while the main geometrical characteristics are summarized in Table 2. In Figure 17 are also indicated the sections instrumented by pressure strips.

Figure 17 - Tested sail plan Table 2 - Main characteristics of the tested sail plan Luff [m] Leech [m] Foot [m] Area [m²] Main 1.94 2.00 0.637 0.810 Jib 2.00 1.89 0.610 0.600

During the tests the model was instrumented in order to measure aerodynamic forces, pressure distributions on the sails, as well as sail shape. Aerodynamic forces where measured by a six components balance placed inside the yacht hull. The balance connects the model to the ground and it is completely covered by the hull has it is possible to see in Figure 18. Pressure distributions were measured by the special system designed for this application and described in the previous chapter: both mainsail and jib were equipped with custom- made pressure strips, each providing three test sections and a total of 40 pressure taps. An example of such pressure strip is given in Figure 6. The wind tunnel pressure system was set up to measure pressure distributions on both sides of the sails in order to deeply investigate the flow field around them. The pressure reference for wind tunnel tests is the mean static pressure of the test section. The pressure maps obtained will be useful also for the analysis of the full scale data where only differential pressures within the two side of the sails will be measured. In full scale differential measures are generally preferred because it Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 is difficult to have a reference pressure signal available in the real operating environment and because the analysis of the differential signals allows evaluation of just the aerodynamic contribution in the pressure field, avoiding pressure variations related to atmospheric pressure.

. Figure 18 – Yacht model in the boundary layer test section of the Politecnico di Milano Wind Tunnel The wind tunnel experimental set-up was completed by a novel sail flying shape detection system, based on Time of Flight technology (TOF). A thorough characterization and validation of this TOF novel technology can be found in Fossati et al. 2015b. Basically, a laser pulse is emitted by the TOF sensor and by measuring the time the pulse takes to hit the target surface and to come back to the receiver, it is possible to estimate the target distance.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 19 – The Flying Shape Detection (TOF) System and its position for wind tunnel tests: one system per sail The device performs measurements in terms of spatial coordinates of thousands of points belonging to the sail surface without impairing its shape since no contact occurs between object and sensor. These data, coming singularly from two different laser scanners, one for main sail and the other for jib (or gennaker), as in Figure 19, are then processed together to reconstruct the 3D sail surfaces (Figure 20), which can be sectioned coincident with the pressure strip heights allowing comparison of the pressure distribution for a given sail shape (Figure 30). The scanning procedure takes only few seconds and gives the steady state configuration of the sail plan. The test procedure was set to characterize the aerodynamic behavior of a sail plan at fixed heel and apparent wind angle. The sail trimmer acts on the control system to obtain the desired sails regulation. In order to properly trim he can use some cameras placed onboard the model to give a view similar to the real life (Figure 21) and the aerodynamic forces signal available in real time. The tests were performed at a velocity of about 6 = 4 H/J, a compromise between higher signals and lower forces acceptable for the sail control system: the wind tunnel speed is mostly limited by the strength of the model mast and rigging and the power of the sheet winches.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 20 - Example of a point cloud acquired for mainsail (blue) and jib (green): 3D reconstruction sail shapes, highlighted sections in correspondence of the pressure strips

Figure 21 - Onboard camera view on jib during testing Upwind sails tests and results The usual way to analyze data of this type of measurements is comparison of non- dimensional coefficients (Fossati et al. 2006, Fossati et al. 2008), allowing to compare the efficiency of sails of different total area at different conditions of dynamic pressure. The first useful parameter to be analyzed is the variation of driving force coefficient $7 versus heeling force coefficient $+. Figure 22 shows $7 vs $K curves for the 6 Apparent Wind Angles (!"!) tested in this campaign. It can be seen that there are some settings at the highest values of heeling force coefficients where the driving force is lower than the maximum value (e.g. below the maximum efficiency curve, isolated points in Figure 22). These non-optimum values were due to an over-sheeting of the sails, such that the mainsail generally had a tight leech and the airflow separated in the head of the sail, (Abbott et al. 1959). After having maximized the driving force, the sails were adjusted to reduce the heeling force while observing the reduction of the driving force. The reduction in heeling force was achieved by initially easing the main sheet, to twist the mainsail and minimize flow separation, then adjusting the traveller to reduce the angle of attack of the wind on the main. Envelope curves

have been drawn through the test points with the greatest driving force at a given heeling force (e.g. depower curves, Figure 22). Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 22 - Driving force coefficient LM versus lateral force coefficient LN Two other interesting parameters used to characterize a sail plan are the positions of the center of effort in terms height ($%ℎ) and in terms of the longitudinal position ($%&). They are calculated as:

2) $%ℎ = 2 ,+ℎ-./0 O $%& = P 3 QRSTS where 2) is the roll moment, ,+ the heeling force, ℎ-./0 the height of the mast, 23 the yaw moment and 1"1 the length of the water line. Both $%ℎ and $%& reported in Figure 23 are normalize over a reference length. As it can be seen, the center of effort height tends to decrease as the heeling force coefficients decrease. This is explained by the way in which the sails are de-powered to reduce $+: increasing the twist reduces the loading in the head of the sails and then depowering the mainsail leaving the same genoa trim, which has a lower center of effort, tends to reduce it. On the other hand, $%& moves forward as $+ reduces: this is again explained by the way the sails are de-powered.

(a) Centre of effort height (b) Centre of effort longitudinal position Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 vs heeling force coefficient vs heeling force coefficient Figure 23 - Position of the center of effort

For each point plotted in Figure 22 and in Figure 23 pressure distributions over the two sails were measured: For a preliminary qualitative analysis, the schematic pressure drawings published in Viola et al. 2011a (Figure 24) can be used as reference: the obtained distributions, especially on the middle sections (see Figure 25) of main and jib, present evident agreement with the expected trends. Comparing Figure 24 and Figure 25 it can be noticed that the main trend is visible, whereas the negative pressure in Figure 25 are bit higher: this can be explained considering the reasonable differences in terms of sail shape and surface roughness. Both sails present a suction region, on the leeward side, where the flow is accelerated and the pressure drops, and a region of positive pressure on the windward side where the flow is slowed down. At the leading edge of both sails, the flow detaches from the surface and reattaches further downstream creating the so-called separation bubble, which is a region of recirculating flow. At the trailing edge, depending on the sail trim and the angle of attack with the incoming air, the flow might separate again leading to turbulent phenomena and vortex shedding. In general, the aerodynamic loads on the headsail are greater and oriented more towards the forward direction compared to the main sail; this characteristic is due to the different shape of the two sails and their interaction with each other. The mainsail positively affects the jib by increasing its effective angle of attack (up-wash effect) and accelerating the flow on the leeward side thanks to the suction region created; to the contrary, the headsail acts oppositely on the main (down-wash effect) which therefore is subjected to lower aerodynamic loads. Some studies expanded this subject, identifying different regions in which attached flow, separation bubbles, reattaching phenomena and detached flow alternate with each other (Wilkinson 1989, Viola et al. 2011a, Viola et al. 2011b, Crompton et al. 2000). It was interesting as in the case of the mainsail, to see that the mast plays an important role in the creation of the separation bubble while for the headsail, even though it does not have any large bluff body in front, the leading edge often presents a separation bubble and the high curvature of the sail might lead to flow separation at the trailing edge.

With reference to Figure 26 (Viola et al. 2011b), the separation bubble on the leeward side occurs when the angle of attack is higher than the ideal one, that is when the sail leading edge is not exactly aligned with the streamlines of the incoming flow. In case of soft sails, a slightly over-tightened trim of this kind is desirable in order to prevent the sails from flapping; therefore, separation bubbles are often present. Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 24 - Expected pressure distributions on the mid sections of the two sails (Viola et al. 2011a)

Figure 25 - Pressure distribution measured at VWV = XE°, level L2, maximum drive force condition

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 26 - Expected L[ distributions along the headsail sections at varying angles of attack (Viola et al. 2011b)

Next, for a more thorough analysis of the local pressure characteristics, the sail sections will be presented by means of two-dimensional scaled graphs. First, the differences of load distributions on sails due to the depowering procedures will be discussed; second, the influence of the variation of wind angle will be evaluated. As depicted in Figure 22, the points laying on the de-powering curve at maximum and minimum apparent wind angle !"! = 35° and !"! = 20°were chosen for a more thorough investigation in terms of pressure distributions and corresponding flying shapes. More specifically, as can be noticed in Figure 13 and in Figure 17, three stations at different height, both on mainsail and jib, port/starboard sides, were chosen for application of the pressure strips on (Figure 13 and Figure 20). L1, L2 and L3 stand for the three levels at 25%, 50% and 75% of the sail heights respectively, so the corresponding pressures at points 1-4 are reported in Figure 27 and in Figure 28

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 27 - Pressure distributions on the three stations along the span of the sail plan: points on the maximum power curve for an apparent wind angle VWV = ^_°

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 28 - Pressure distributions on the three stations along the span of the sail plan: points on the maximum power curve for an apparent wind angle VWV = XE°

From a first look, all the trends obtained present characteristics in accordance to the reference studies present in literature as observed before: it is possible to recognize the different flow features described, such as leading edge separation bubble, curvature-related suction peaks and trailing edge pressures, are all clearly visible both on main and jib experimental distributions. This is particularly true for the middle sections where the flow is not disturbed by tip effects. Starting from the main sections, as expected, the mast has great influence on the pressure distribution and interestingly, the effect varies along the sail height and with trim. For both of the angles considered an increase of the negative pressure coefficients can easily be seen in the leeward side near the leading edge, where the mast is placed: as discussed before this is related to a large separation bubble that extends up to 10% of the chord. The highest negative values in this region were obtained for the powerful configurations while when easing the main the flow is less accelerated and consequently

the negative values became lower. In these cases, the greater sail curvature leads to higher (more negative) pressure gradients that allow reaching similar values at the middle of chord; from that location, the downstream distribution remains almost identical within the same depower curve. Furthermore, moving along the airfoil chord, up to the trailing edge, another greatly negative pressure coefficient region is experienced. For !"! = 20° all the considered points show a second separation region, while for !"! = 35° only the points at lower driving force coefficient $) (i.e. 3-4in Figure 22) have this region. For points with greater efficiency (i.e. 1-2 of Figure 22) this bubble seems not to be occurring: therefore, after the separation due to the mast, a monotonic trend up to the trailing edge is evident, indicating a gradual and efficient reattachment of the flux on the leeward side of the sail. These differences related to the angle of attack will be discussed more in detail in the following. At the trailing edge, the pressure tends to level off at constant values indicating partial flow

detachments that prevent the pressure to fully recover. Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 Regarding the headsail, it is worth recalling that its shape was not changed during the depowering procedure; therefore, the pressure distribution differences between the various trims must be attributed to the changes of the mainsail’s influence only. In particular, a more powerful suction on the main’s leeward side, corresponds to a stronger up-wash effect which, in turn, results in greater induced angles of incidence. With reference to the 20° case (Figure 28), it is evident how the changing angle influences the jib’s leeward pressure distribution, especially at the leading edge. More specifically, when the up-wash effect is very strong, i.e. the angle of incidence very high (trims 1 and 2), the pressure at the leading edge reaches considerably low values and the separation bubble enlarges. As the main is eased, the up-wash induced angle of attack of the jib decreases and the leading edge bubble gets smaller. Also for the 35° case a similar behavior was observed (Figure 31 particularly in the middle section (L2). The different lifting attitude of the jib revealed by the measured pressure distributions can be considered the main source of the driving force reduction along the de-power curve.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 29 - Comparison of the pressure distribution of different apparent wind angles VWV (20°, 25°, 30° and 35°) both considered at the maximum driving force coefficients

In order to consider the !"! effect on the pressure distributions, in Figure 29, a comparative plot of 4 different !"! (20°, 25°, 30° and 35°), all from maximum driving force coefficient configurations (see Figure 22), is reported. Considering again the section L2 both for jib and for main, without loss of generality for the other sectional airfoils, the following explanations can be drawn: the pressure distributions of the jib have basically similar trends and values, indicating there exists an optimum shape that maximizes the aerodynamic efficiency of the sail. Nevertheless, at increasing wind angles, this same regulation results in growing angles of attack that lead to strong leading edge negative peaks. The main differences, especially in terms of values can be noticed on the mainsail’s distributions. The different orientation of the mast, which has an elliptical section, together with the different sail curvatures for the angles considered, generates suction peaks of

different intensity especially at the leading edge. At 30° and 35° the mast remarkably accelerates the flow, resulting in a separation bubble with high mean velocity, as the !"! decreases, the angle of attack of the mast reduces weakening the flow acceleration and, consequently, the intensity of the depression. This difference results in different driving forces and in different center of effort positions. In order to better understand how the pressure distributions are related to overall aerodynamic force, pressure and sails shape are considered at once. Figure 30 shows the pressure distribution for !"! = 20° in the maximum driving force condition corresponding to section L2: the pressure are directly plotted on the measured section with blue arrows corresponding to positive pressure coefficients (windward), whereas red ones to negative (leeward). From this representation, it is clear that the driving force, computed as the integral of the pressure distribution on the corresponding section, is mainly contributed by the jib. Comparing Figure 27 and Figure 28 is also interesting to notice that a less efficient sail plan Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 (e.g. pt.4 of !"! = 35°, Figure 27) is aerodynamically similar to a sail plan set to a smaller effective apparent wind angle (!"! = 20° Figure 28), which is something that is also known and commonly experienced in sailing. As a matter of fact, great similitude can be found in the results reported in Figure 27 and Figure 28, with the extensively investigated aerodynamics of airfoils with external airfoil flaps (Abbott et al. 1959). The same considerations can be drawn analyzing pressure distributions along with the sail shapes detected by TOF system. More specifically, in Figure 31 a comparative visualization of the pressure distributions on the effective flying shape, at the corresponding section (L2), is reported for the apparent wind angle !"! = 35° and for maximum and minimum efficiency (i.e. pt.1 and 4 of Figure 22).

Figure 30 - Visual pressure distribution on the sail plan based on measurements at section L2, VWV = XE° The pressure distributions are the same as reported in Figure 27. It is evident how basically the same jib’s shape of the section, which was not modified along the de-power curve of Figure 22, results in different pressure trends due to the different efficiency of the mainsail, whose suction effect on the jib is greatly lower in a de-powered setting. Also the different overall lateral forces ,+depicted in Figure 30, computed from integrating the pressures over the L2 section, confirm the integral measurements reported in Figure 22. Furthermore, in Figure 32, the comparison of different !"! results (20° and 35°, also reported in Figure 28), both in a maximum power configuration, for L2 section, are shown in terms of pressure/shape visualizations. It is easy to notice that the fairly similar pressure distribution on the jib due to the maximum power based trimming, is associated with different shape, leading to different driving forces on the jib and comparable lateral forces on the mainsail

(i.e. Figure 22, for approximately the same $+, !"! = 20° shows a lower driving force coefficient $) with respect to !"! = 35°). Also Figure 33, in which apparent wind directions are highlighted, supports the consideration reported above: more specifically, the less efficient sail plan at !"! = 35° shows similarities with lower !"! configurations in terms of mainsail pressure distribution. In fact, a reattachment of the flow right behind the mast combined with the co-alignment of the flow with the luff of the mainsail is quite evident. It is worth mentioning that these tests represent, according to author’s knowledge, one of the first wind tunnel set of measurements of the pressure distributions on flexible sails, instead of rigid, which is a testing condition closer to the real navigation (i.e. full scale), whose measurements (forces, pressures and flying shapes) on Sailing Yacht Laboratory, will take advantage of the present research. Furthermore, the rationale behind the way the sails were trimmed, corresponds to a procedure which is consistent to the real sailing navigation. However, a systematic investigation on the aerodynamics of sail plans, relying on this experimental setup, could be Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 based on different depowering approaches, aiming at assessing the effect of single trim parameters (involving the jib also) on the local and integral efficiency of the sails.

Figure 31 - Pressure distributions for L2 section for VWV = ^_°, maximum and minimum power configurations

Figure 32 - Pressure distributions for L2 section, maximum power configuration, VWV = XE° and VWV = ^_°

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 33 - Pressure distributions for various VWV and power configurations

FULL SCALE PRESSURE SYSTEM LAYOUT On the Sailing Yacht Lab, the pressure distribution on the sails is carried out by means of specifically designed pressure pads which have been designed and produced aiming to provide the differential measurement between the sail leeward and windward side. The above reported metrological characterization of the strips and the wind tunnel tests were mainly to assess the capabilities of the measuring system, in advance with respect to the final full scale implementation with pads. However, during the wind tunnel tests some tests on pads themselves were conducted just to verify they functioned as expected, Figure 34. With reference to the Sailing Yacht Laboratory sail inventory, Figure 35 and Figure 36 show the proposed sail pressure measurement system, the sections considered and a relevant number of pressure taps for the complete sail plan. The full scale pressure system is being finalized at the time of writing this paper.

Figure 34 - Example of the full scale pads on gennaker sail for wind tunnel model

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 35 - Full scale mainsail and jib pressure taps layout

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021

Figure 36 - Full scale gennaker pressure taps layout

CONCLUSIONS A joint project developed among Politecnico di Milano, CSEM and North Sails, aiming at developing a new sail pressure measurement system is presented in the paper. The system was designed for the specific application of full scale measurements on Politecnico di Milano Sailing Yacht Laboratory. The capabilities of this system were evaluated through a metrological validation of the system alone (static and dynamic tests) and through wind tunnel tests in the upwind configuration. The pressure system was integrated in the existing set up, in particular together with flying shape detection based on time of flight technology. Wind tunnel tests allowed checking the reliability of the new system and investigating thoroughly upwind soft sail aerodynamics, with the possibility of carrying out sail trim as in real navigation. During the wind tunnel test session, each side of the sail pressure distribution was measured with reference to a common static pressure, by means of pressure strips, whereas for full scale implementation, pressure pads for differential measurements were developed. Wind tunnel results give a promising preview of the potential of the system described in explaining the dependency of sail plan aerodynamics on sail trim, relying on the combined measurements of forces, pressures and flying shapes. Furthermore, these measurements will represent a great reference database for validation of CFD codes and can be used to complete the interpretation of full scale results. Further developments in the visualization techniques (particle image velocimetry) are expected to be combined with this methodology in the near future.

ACKNOWLEDGEMENTS The authors want to thank the master student Riccardo Romagnoni for his help in performing the calibration tests.

REFERENCES Abbott H., Doenhoff E., Theory of Wing Sections: Including a Summary of Airfoil Data, Dover Publications, Inc. New York, 1959. Crompton, M. J., Barrett, R. V, “Investigation of the separation bubble formed behind the sharp leading edge of a flat plate at incidence”, Proc. Inst. Mech. Eng. Part G J. Aerosp. Eng., vol. 214, pp. 157–176, 2000. Deparday, J., Bot, P., Hauville, F., Motta, D., Le Pelley, D.J., Flay, R.G.J., Dynamic measurements of pressures, sail shape and forces on a full-scale spinnaker, in 23rd HISWA Symposium on Yacht Design and Yacht Construction, Amsterdam, 2014 Fossati, F., Muggiasca, S., Maria, I., Zasso, A. Wind tunnel techniques for investigation and optimization of sailing yachts aerodynamics (2006) 2nd High Performance Yacht Design Conference 2006, pp. 105-113.

Fossati, F., Muggiasca, S., Martina, F. Experimental database of sails performance and Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021 flying shapes in upwind conditions, in INNOVSAIL International Conference on Innovation in High Performance Sailing Yachts, Lorient, RINA, pp. 99-114, 2008. Fossati, F., Muggiasca, S. Experimental investigation of sail aerodynamic behavior in dynamic conditions (2013) Transactions - Society of Naval Architects and Marine Engineers, 120, pp. 327-367. Fossati F., Bayati I., Orlandini F., Muggiasca S, Vandone A., Mainetti G, Sala R., Bertorello, Begovic E: “A Novel Full Scale Laboratory For Yacht Engineering Research”, Ocean Engineering 104 (2015) 219-237. Fossati F., Mainetti G., Sala R., Schito P., Vandone A., “Offwind Sail Flying Shapes Detection, 5th, High performance Yacht Design Conference, Auckland, 2015. Le Pelley, D., Morris, D., Richards, P., "Aerodynamic force deduction on yacht sails using pressure and shape measurements in real time", in 4th High Performance Yacht Design Conference, Auckland, pp. 28–37, 2012. Lozej, M., Golob, D., Bokal, D., "Pressure distribution on sail surfaces in real sailing conditions", in 4th High Performance Yacht Design Conference, Auckland, pp. 242–251, 201, 2012. Motta, D., Flay, R.G.J., Richards, P.J., Le Pelley, D.J., Deparday, J., Bot, P., "Experimental investigation of asymmetric spinnaker aerodynamics using pressure and sail shape measurements", Ocean Engineering, 90, pp. 104-118, 2014. Motta, D., Flay, R.G.J., Richards, P., Le Pelley, D.J., Bot, P., Deparday, J., An investigation of the dynamic behavior of asymmetric spinnakers at full-scale, in 5th High Performance Yacht Design, Auckland, 2015 Pot, P., Viola, I.M., Flay, R.G.J., Brett, J.S., Wind-tunnel pressure measurements on model- scale rigid downwind sails, Ocean Engineering, 90, pp. 84-92, 2014 Tijedman, R., Berg., M. T. "Theoretical and experimental results for the dynamic response of pressure measuring systems", Technical Report NLR-TR F.238, 1965 Viola, I.M., Flay, R.G.J., "Sail pressures from full-scale, wind-tunnel and numerical investigations", Ocean Engineering, 38 (16), pp.1733–1743, 2011a. Viola, I.M., Flay, R.G.J, “Sail aerodynamics: Understanding pressure distributions on upwind sails”, Exp. Therm. Fluid Sci., vol. 35, no. 8, pp. 1497–1504, 2011b.

Viola, I.M., Flay, R.G.J., "Sail aerodynamics: on-water pressure measurements on a downwind sail', Journal of Ship Research, 56(4), pp.197-206, 2012. Wilkinson, S. “Static pressure distributions over 2D mast/sail geometries”, Mar. Technol., vol. 26, no. 4, pp. 333–337, 1989.

Downloaded from http://onepetro.org/JST/article-pdf/2/01/1/2205466/sname-jst-2017-03.pdf by guest on 26 September 2021