Windsurfing Sail and mast design

Discussion in 'Hydrodynamics and Aerodynamics' started by matt_bob, Apr 13, 2015.

1. Joined: Jan 2013
Posts: 56
Likes: 6, Points: 8, Legacy Rep: 22
Location: Lodz, Poland

myszekJunior Member

Have anybody tried to streamline himself? Like speed skiers do

regards

krzys

2. Joined: May 2003
Posts: 1,350
Likes: 142, Points: 63, Legacy Rep: 215
Location: Sydney Australia

CT249Senior Member

Yes, at least one guy tried a streamlined suit in the mid '80s, but without the calf fairings and helmet. He wasn't particularly fast.

How does the design you are analysing compare to the gear used by the fastest windsurfers?

3. Joined: Oct 2005
Posts: 441
Likes: 25, Points: 28, Legacy Rep: 97
Location: France

ErwanSenior Member

Swedish, Russian and more generally Scandinavians, have "Wing-Skates"
Some are just a wing with large thickness, the crew is inside the wing, there is a transparent wind-screen on the leading-edge, just like a motorbike helmet, and just the feet with ice-skate which appears at the bottom are not streamlined.
To sail, you must move your body to windward in order to create righting moment.
Not sure it can help

4. Joined: Feb 2002
Posts: 2,319
Likes: 295, Points: 83, Legacy Rep: 1673
Location: Port Gamble, Washington, USA

tspeerSenior Member

I suggest you work the problem backwards, and develop requirements for each stage of the design. Start with the speed record you want to break. If it's Paul Larsen's outright speed record of 65.45 kt, you probably can't do it with a windsurfer. The last record set with a windsurfer was 49.09 kt, so 50 kt would be a reasonable target for a windsurfer.

The relationship for sailing speed is Vb = Vt*sin(gamma-beta)/sin(beta), where Vb is the boat speed, Vt is the true wiind speed, gamma is the course relative to the true wind (0 = straight upwind), and beta is the apparent wind angle, measured between the air and water velocity vectors. You can choose gamma so that sin(gamma-beta)=1, and achieving the performance comes down to minimizing beta. beta = arctan(D_aero/L_aero)+arctan(D_hydro/L_hydro), so now you have to start making a lift & drag budget for the aerodynamics and hydrodynamics. If you assume a hydrodynamic lift/drag ratio and a target wind speed, then you can determine your target lift/drag ratio for the aerodynamics.

The heeling moment has to equal the righting moment, and if you take the mast step as the moment reference center, the righting moment comes mostly from the sailor's weight and positioning. The amount of lift on the rig will be determined by the location of the center of effort of the sail rig and the allowable heeling moment.

The aero L/D includes all the aerodynamic lift and drag, including the drag of the sailor and rigging, like the wishbone boom. When you divide the lift by the lift/drag ratio, you get the total drag. From the total drag, you can subtract the windage of the sailor, etc., to get the drag budget for the sail itself.

So once you've made some assumptions about the geometry of the sailor and craft, you can start to fill in your lift and drag budgets. Induced drag may be the largest source of aerodynamic drag. The taller the rig, the less induced drag there will be, but the less lift as well, because of the heeling moment constraint. The induced drag will also depend on how the lift is distributed along the span. A more bottom-loaded lift distribution will have a lower center of effort, allowing more lift for a given heeling moment, but also a higher induced drag. But for a given heeling moment, the bottom-loaded lift distribution will allow a longer span, which reduces the induced drag. The tradeoff looks something like this:

Parasite drag of the sail will depend largely on how much sail area there is. You will want to size the sail so it operates near the section's best lift/drag ratio.

By iterating on these assumed values, you can get to where you know the size of the rig and how the lift is distributed along the span. Then you can make some assumptions about the twist and calculate the planform shape from the lift distribution. This lifting line spreadsheet may be useful for doing that. Now you can calculate the lift coefficients for the various spanwise stations.

Once you have the design lift coefficients, the last step is to design the section shapes. You need to come up with something that can be achieved with a membrane sail and accommodate the mast in the luff. The camber will determine the zero lift angle of attack. Since you know the total aerodynamic twist from your assumptions, you can subtract the zero lift angle of attack from the aerodynamic twist to get the physical twist. The physical twist needs to be achievable from the leech tension.

An airfoil design code like XFOIL can be used to design sections to these requirements. You will need to iterate on the design pressure distribution and use the inverse design modes to calculate the section shape.

Once you have the 2D drag polar from the section code and the induced drag from the lifting line, you can put together the aerodynamic characteristics of the rig and see if you are meeting your lift and drag budgets. You'll have to go back and revisit your original assumptions several times until you get things to add up.

5. Joined: Feb 2002
Posts: 2,319
Likes: 295, Points: 83, Legacy Rep: 1673
Location: Port Gamble, Washington, USA

tspeerSenior Member

Yes, you will want to include the mast in your section design. The structural thickness of the mast, and its shape (round or elongated) will be a major constraint on the section shape. Otherwise, you'll find that a very thin section will produce the least amount of drag at your design lift coefficient, but this won't be something that can be built.

Start with the approximate shape that you've shown. You can use it for a starting shape in XFOIL. In XFOIL, the OPER mode will allow you to evaluate the shape. You can use the MDES mode to alter the pressure distribution to get a better shape.

Where there was flow separation in OPER, you need to use MDES to flatten the slope of the pressure distribution. In OPER, the VPLO command will plot the boundary layer shape function, H, and the pressure coefficient, Cf. Separation is indicated by H>~3 and Cf=0. The more adverse the pressure gradient, the larger H will become, so flattening the pressure gradient will lower H. You can push the pressure increase further upstream, because a fresh boundary layer can tolerate a much more rapid increase in pressure than a boundary layer that's already been stressed. (Pretend the boundary layer is a cyclist trying to climb a big hill, and you have to shape the hill so the cyclist can make it to the top.)

6. Joined: Apr 2015
Posts: 9
Likes: 0, Points: 1, Legacy Rep: 10
Location: Pretoria, South Africa

matt_bobJunior Member

Hey.

Apologies for the slow reply, been getting bogged down with projects and stuff and I haven't had time to write a meaningful response.

@Daiquiri
The design goals were left for me to decide as the person undertaking the project. This is what goals I am working with:
- Windsurfer weight = 80kg
- Maximum board speed of 25m/s
- True wind speed of 41 knots [this is so that each part of the sail experinences at least 30+ knots as per the requirement stated earlier calculated as a result of this paper Estimation of windspeed by M. Tahbaz]
- sail size has to be close to 6 sqm
- I am going to develop the forces then see what material properties are required to handle those forces
-professional windsurfer is in operation of the rig

I am skeptical about using that information from M. Tahbaz to calculate the true wind information, but so far it is all I have.
Thanks, I changed the Ncrit value and the re-did the Cl and CD simulation

@Erwan
Thanks, I re-ran the simulation using NCrit of 1
The work paper is was referring to by Larsson et al is this textbook Principles of Yacht Design

We are assuming that a professional is using the rig so we don't really need to consider the surfer, but there is a limit to what the human body is able to handle. But we also need to be aware that the rig is meant to break a record so comfort is not a big priority here as he will only use the rig for a short period of time.

@Doug
Yeah, I appreciate the concern. I think the supervisor likes to windsurf and wants his students to design something that he can ride or make. I am not sure. He doesn't give much feedback. But other students are doing the board and the CFD is so complex that it requires years of experience to do. I did propose some new ideas like a fixed laminated sail similar the type used for Americas cup but he quickly shut that down. He wants us to just design something and compare it to currently available product to see if we have improved the current technology.

@SukiSolo
I didn't consider windsurfer drag. How did I miss that. I am going to assume a shape for the windsurfer that will approximate the irregular shape. Thanks for that.

@tspeer
Wow. Thank you so much. I appreciate that post so much. This is the record that I am attempting to break Antoine Albeau : 52.05 knots in Luderitz, Namibia which means that the board speed of 25m/s that I have chosen is too small, but not by much.
I might be following this type of method anyway, these are the steps that I have completed so far:
1. Developed a initial sail planform, broke the sail into solvable shapes, made an excel form to solve it for 6 sqm. From this got the dimensions of the sail.
2. Calculated the true wind based on a chosen 41 knots at 47 degrees [which more than meets the requirement of 30+ knots]
3. From that calculated the apparent wind velocity and apparent wind angle
4. From that, worked out the Reynolds numbers for the flow over the sail at each batten position as the chord length is known
5. Developed cambers that include the mast in XFLR5 and did Cl and CD plots for those at the reynolds numbers for each chord length and at the various heights.
6. Each height has a different apparent wind angle, so I found that angle on the plot and captured the Cl and Cd values.

That's where I am at the moment. Paring the cambers with the proposed batten positions. I am not sure that I have time to restart the detailed design before hand in. I unfortunately don't really understand how to use Xfoil. But XFLR does also have inverse design for the aerofoils. Not sure what pressure distribution I need to achieve.

From the Principles of Yacht design I am deciding between camber ratios of 1/7, 1/10 and 1/20. Should I redo the camber section of the project using the method u have mentioned here? I can post my results so far, if that will help.

Guys I just want to thank you all for all the help that you are giving. As an aspiring naval architect the information you are sharing with me not only is giving me such a passion to pursue this path even more but also going to help me later in life. Thank you so much.

7. Joined: May 2003
Posts: 1,350
Likes: 142, Points: 63, Legacy Rep: 215
Location: Sydney Australia

CT249Senior Member

Nice post Matt. I'm no NA but I've been lucky enough to talk to many of the world's greatest designers of sailing craft, and the one thing they all seem to share is a thorough knowledge of previous developments in their particular section of their field. Out of interest, does your assignment allow you to use the lore obtained from earlier designs, or does it require you to develop everything from theory?

BTW a camber ratio of 1 in 10 is enormous by windsurfer speed standards.

8. Joined: Oct 2005
Posts: 441
Likes: 25, Points: 28, Legacy Rep: 97
Location: France

ErwanSenior Member

Thank you Matt,

I would have 2 questions if you don't mind:
Do you use any Excel file for your mast+ sail shape before xflr5?
Have you identify on xflr5 the the VPLO command and the bonundary layer shape function H, aboved mentionned by Mr Tom Speer?

Regards

EK

9. Joined: Apr 2015
Posts: 9
Likes: 0, Points: 1, Legacy Rep: 10
Location: Pretoria, South Africa

matt_bobJunior Member

Thanks um, I didnt think it was so enormous when I made it. Here is a picture of the cambers I used for the CFD.

1. I only used excel to calculate the appropriate Reynolds numbers to run the XFLR analysis at. It does this by calculating the chord lengths from dimensions, as well as the apparent wind velocity at the relative height and then calculates the reynolds number from those values.

2. I have not gotten to that point yet as I have kept that cambers that I designed before I read his post. I unfortunately do not have time to re design the cambers at this stage in the project.

Attached Files:

• Camber comparison.PNG
File size:
11.7 KB
Views:
1,037
10. Joined: Oct 2005
Posts: 441
Likes: 25, Points: 28, Legacy Rep: 97
Location: France

ErwanSenior Member

Matt,
If sailing conditions mean some gusts and changeS in true wind directions, the control of the rig by the dude can be included in the design process, when considering wing section Cmo and how it changes with AoA & apparent wind speed.
According to the envelope of true wind & apparent wind speed, and your righting moment I 'll be surprised you'd need more than 1.5% to 3% mean line camber for your wing/sail section.

Nice project, my best wishes for your research

EK

11. Joined: Oct 2007
Posts: 2,936
Likes: 148, Points: 63, Legacy Rep: 1593
Location: Arlington, WA-USA

PetrosSenior Member

I have been lurking this thread for a while, interesting subject.

After watching several Youtube clips of the record setting runs of sailboarders, several things came to mind: the amount of bouncing around of the rig and rider, would not allow for any stead flow conditions over the surface of the sail (it was quite a violent shaking from the surface ripples at those speeds), and the drag of the rider would not be insignificant. go look at the videos and observe the shaking and jiggling of the sail.

so it seems to me if there was a way to smooth the ride (foil perhaps?) and to put the operator in a stream lined suite or "shell", much higher speeds could be reached. consider that ice sailors reach speeds several times higher than any water vessel powered by sails.

12. Joined: Apr 2015
Posts: 9
Likes: 0, Points: 1, Legacy Rep: 10
Location: Pretoria, South Africa

matt_bobJunior Member

Hey. Thanks for that. Yeah, I cant avoid it anymore, I need to redesign these cambers properly. From what I can see, 1% to 10% range is more than adequate but the cambers I have made either give lift coefficients of 0.9 or of near 2. no in between which is making the selection process difficult.

I dont know how to get proper aerofoil dat files for these things.

Well the water is assumed flat, it is a man made channel so chop wont be an issue, it is the flow of the wind that might but will proper rigging of the sail cloth you can get it really close to a fixed wing type sail. which unfortunately, we cant use a fixed sail :/

Not too sure how to account for the drag of the rider. I was going to assume a cylindrical approximation and get the drag of that, see if the mast can handle it.

13. Joined: Oct 2005
Posts: 441
Likes: 25, Points: 28, Legacy Rep: 97
Location: France

ErwanSenior Member

I think there is a 53% relative thickness wing section* (Joukovsky) with a 0.077 drag coefficient, while for cylinder or human body Cd is probably around 0.8 or at least 0.5

You can find something about this wing section in a workpaper: A.O. Smith / High Lift Aerodynamics, or something very similar.

Cheers

EK

14. Joined: Apr 2015
Posts: 9
Likes: 0, Points: 1, Legacy Rep: 10
Location: Pretoria, South Africa

matt_bobJunior Member

Alright, so I wasn't sure about my sail cross sections, and the sail cross section needs the mast to be included. So what I did was go to airtools.com and use the NACA generator with 1 thickness, chose the cambers and draft position. Then copiedthe .dat file and put that into XFLR 5 to edit it and make the mast lump.

Does this sound like a better way to make them? I don't know what I want my pressure distribution to be like so I have done it pure inverse.

The coefficients are still so high, the mas really has an effect on them dramtically. Never expected that.

15. Joined: Oct 2005
Posts: 441
Likes: 25, Points: 28, Legacy Rep: 97
Location: France

ErwanSenior Member

You can have a look at pressure distribution of wing sections like:
E211 E214 E216
or
S3021 S4053 S4061 S4062 S4083 S6061 S 6062 S7012

It can give you some insight for the pressure distribution

Cheers
EK

Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.