What is a significance of a wing thickness

Discussion in 'Hydrodynamics and Aerodynamics' started by markmal, Nov 16, 2012.

  1. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    That is what you get for let a general purpose engineer jump into specialty software.

    Since two points define a straight line, I assumed that the software used the line as a boundary surface, and there was no need of a bunch of points on the line.

    I inserted a bunch of points and the results changed drastically.

    Back to the drawing board.
     
  2. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    Ok, the redirection from Mark required taking a few steps back before I could make any forward progress.

    I had to go back to my spreadsheet and re-vamp it to allow all of the manipulation (extract forward top surface, flip it over, add flap upper surface with 20 panels, add flap lower surface with 20 panels). I built in the ability to pick an alpha angle for initial mirroring, define upper flap surface angle and trailing edge (TE) gap.

    The mirroring takes extra work in that the formulas end up with wrong order for XFOIL such that I have to copy, paste values & re-sort.

    The other messy piece is changing the transition point to flat flap surface. This is done with manual moving stuff around and adjusting specific cells.

    The good news is that I got back to better performance for boating than I found in standard foils.

    A big chunk of sailing is operating at less than "fully powered up". This is where you have more Righting Moment (RM) available than required. For these cases, you may be at moderate to low wind and adjusting for max lift (Cl) is likely to be your best bet. I picked a low Reynolds Number (RE) of 100,000 to be representative of the low to moderate wind.

    I tweaked my G679rf to give a good looking low wind operating point at alpha = 10, Cl around 1.58, Cl/Cd of around 40. Compare this to a Cl of around 1.2 for Eppler 748 and USA35a with Cl/Cd values of around 30 and 22 respectively.

    One strange thing about this effort is that XFLR5 showed big low RE gains with a fat TE gap. The big blunt TE hurts high RE Cl/Cd some, but when I got what looked like a decent mix, it was a lot fatter than I expected.

    I picked a RE of 1,000,000 for more top end conditions. I got better on both Cl and CL/CD above alpha = 10. If actual sailing found a big fraction of time with alpha below 10, some adjusting would be worth considering.

    The real advantage for the G679rf would be in the 500,000 to 1,000,000 RE range with an alpha around 12. Cl is up around 2 with Cl/Cd of 70 and 90 respectively.
     

    Attached Files:

  3. tspeer
    Joined: Feb 2002
    Posts: 2,319
    Likes: 294, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    Mark is right - you need to use more points on the flap panels. They only need to be linearly interpolated, however, so it's not difficult to do in a spreadsheet.

    When you read the coordinates into XFOIL, execute the PANE command and save the results as a new coordinate file. PANE will pass a spline curve through the points and interpolate new points, with a higher concentration of points where the curvature is greatest. This will result in an appropriate number of points with a good distribution of the points. Duplicate the coordinates at the flap panel hinges to signal XFOIL that there is a sharp corner at those points.

    The attached files show the result, and the first plot compares these coordinates with the original coordinates. When XFOIL passed its spline curve through the end points of the flap panels, not having any better information about the shape of the flaps themselves, the top and bottom curves intersected with each other. Even if you'd done a PANE at the top level with the original coordinates, this is what XFOIL would have been analyzing. But with the corner points and the intermediate points, the result accurately captures the shape.

    The second plot shows the a pressure distribution around this section. The pressures over the forward portion are a bit lumpy. XFOIL can help you smooth these out. I assume you have a spreadsheet that allows you to vary the flap panel angles. I suggest you generate a symmetrical section that has no flap deflection. Save the coordinates, bring them into XFOIL, and execute PANE as described above. Then go into OPER and generate a pressure distribution at zero angle of attack.

    Next, go back to the main level and into the MDES design mode. Use the SYMM command to constrain the design to be symmetrical. Blow up the plot of the design pressure distribution for one side of the forward half of the section so you can work with it in more detail. The FILT command will smooth the pressure distribution. You can use FILT repeatedly or manually modify the pressures using the MODI command. In practice, it will probably require a combination of both commands to whip the pressure distribution into shape. If you make the design pressure approximate the points of the original pressure distribution, but smoother, the result will be a section that is little changed from your original one, but will be smoother and give you better results. Execute the EXEC command in MDES, return to the top level to execute the PANE command, and then save the revised coordinates.

    You will probably need to bring the revised coordinates into your spreadsheet so you can enforce the shape and change the angle of the flap panels. Use the same process of executing the PANE command when you load the program into XFOIL to adjust the paneling over the newly deflected flap panels.

    You can rarely use just XFOIL to create a section like this, because of the unique geometrical constraints. You need to go back and forth between XFOIL and a spreadsheet to refine the design.
     

    Attached Files:

  4. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    Tom,

    You were apparently working through stuff at the same time as I was. If you look at my 1-28-13 post (post #47) you will see that I now have my spreadsheet set up for more points along with the ability to do all of the manipulations needed to mirror existing foils and then play with flat flap surfaces.

    FYI, even with my "more points from the spreadsheet", XFLR5 is not quite happy. It does some weird stuff when I use certain design manipulation commands unless I use the "refine globally" command to redistribute points (it wants real small panels close to the lower flap hinge point).

    I did play around and confirm that I can just load and run my dat files with no manipulations and the results are real close to what I get when I do the "refine globally" manipulation.

    I think (hope) I am over the hump with basic generating and running polars for this class of foil sections.

    Hopefully, we can transition to discussion of results.

    Partially for the fun of it, I did a couple of runs of a ClarkY mirror as a wingmast with soft sail trimmed flat (actually I was just too lazy to reconfigure to get your old shapes).

    For my for low RE point, the ClarkY wingmast results came out better for CL/CD but with CL max. For high RE, the ClarkY wingmast looks better up to aplha = 8, but much less at high aer alpha and much less CL max.
     

    Attached Files:

  5. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    Ok, sometimes I get a little anal and just have a hard time not going all the way,

    I did get a wingmast to run that is closer to what Tom did long time ago.

    In the polar, the top foil (CY-10.7-0.190) is the new version of the old idea.

    It would not run until I put a little gap between the upper and lower surfaces to the rear.

    It has pretty good overall performance.

    I have just never figured out how to control a soft surface good enough to make it work.

    Looking at it more, I think it would work better if I tweak it to get rid of that point at the nose. However, I need to give up for now.
     

    Attached Files:

  6. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    After post 50, I spent some more time playing around with:

    • Wing mast nose shape
    • flap thickness for flat flaps

    For what I would consider, the post #50 ClarkY stuff nose was too pointy. At high alpha (angle of attack), there was a big pressure spike that messed up performance. Also, the front section was really not thick enough for my planned construction. I tried to blunt it and fatten it up, but it did not work out.

    I went back to what just seems to work best. The GOE 679 mirror has a nose that seems to work smooth at all of the alphas I have tried. It also gets me to a 18% thickness with about the right transition from forward mirror portion to flap.

    I took this and did a wing mast, an three different flap thicknesses. From these four, I think I can draw some conclusions.

    When I look at the polar comparisons of wing mast vs every thing else, I saw good high RE performance at low alphas, but worse every where else. I looked at the boundary layers for the wing mast and it was obvious that the extra curvature was driving flow separation. This can probably be helped with thinner front sections, lower front section fractions, and some "under rotation" of the mast.

    However, the 'flat flap' approach might not be so bad for a wing mast. Rods with control arms in lieu of battens comes to mind.

    On the other hand, the polars show the curvature can get more ultimate lift (max Cl) if lots of drag is not a problem.

    For flap thickness & TE gap, the fat flap helps make the bottom smoother and this seems to help at low REs. At high REs, thin flaps and or medium flaps with smaller TE gaps seems to be better. Having a "fat rear flap" that is really separate upper and lower surface flaps would allow a variable TE gap and a "best of both worlds" approach. I think I have seen this somewhere, but I details escape me.

    Overall from a structural / build considerations aspect, the mid flap is looking pretty good right now (If I stick with a one piece rear flap).
     

    Attached Files:

  7. tspeer
    Joined: Feb 2002
    Posts: 2,319
    Likes: 294, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    A flat flap is a good choice,especially if you are shooting for high lift. FWIW, the flap section on USA 17's wing was flat for 80% of the flap chord (40% of the total chord). The pressure recovery region on the main element was also flat (~40% of the main element chord). Essentially, all of the fabric covered regions were flat.

    A flat profile produces a concave pressure recovery that is not as severe as a Stratford pressure distribution, so it has a gentler stall that begins at the trailing edge and progresses forward. The pressure recovery has to begin a little earlier, so there's less maximum lift potential, but it's a good practical compromise.

    It's not unusual for the forward portion of a flapped section to be set at a negative angle of attack. The upwash still results in the stagnation point being on the windward side of the leading edge. I think you'll find the maximum lift/drag ratio typically occurs when the stagnation point is fairly close to the leading edge. That's because the formation of a leading edge suction peak causes the drag to increase.
     
  8. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    Another day, another set of data points and conclusions (from this rank amateur).

    I see the "reversible camber" wing as most useful in a high aspect ratio application. Tom and others have repeatedly pointed out that minimum overall drag needs full consideration of aspect ratio and shape (planform) along with the foil section.

    A very high aspect ratio wing can also be paired with a foresail. Use of my efforts in this thread for this approach should probably end up similar to Chris White's "mastfoil" that that I first saw in action at http://chriswhitedesigns.smugmug.com/Other/MastFoil/19592678_p56mx6#!i=1534942128&k=spKxqLj

    For a wing, the skin can be designed to handle the bending forces, but to do this right requires materials and methods I do not want to attempt. The other option is an internal main structure (I plan on a rectangular spar). A thick section (I am currently thinking 18% of total chord) helps to minimize the required material (carbon tow or pultruded carbon rods) and to keep the weight down.

    Also note that you really need to think about your performance goals. I expect to operate at near Cl-max until the point where I run out of righting moment. This makes max lift with decent Cl/Cd the focus at some low RE point (I picked 100,000). For upwind with plenty of wind, max Cl/Cd for a given mid range Cl is probably more important (I used RE = 1,000,000 to assess this case).

    With 18% forward sections, it seemed that I could get surprisingly similar performance from any reasonable starting point shape. The trick was just to poke around at percent of chord for the flap, and trailing edge gap to get what looked good on the polars.

    The flap deflection was much less trial and error. The best deflection was usually what it took to be very low curvature at the foil to flap transition on the top side. On some foils, a little convex curvature was OK at the hinge if an overall increase in lift was what I wanted, slightly concave worked a couple of times to keep flow attached longer.

    My current pick is my "Blend1" forward section that started as two foils that XFLR5 interpolated. This foil is more tapered in the first 20% of overall chord but it avoids being too "pointy" at the very leading edge.

    Another idea I tried was simple a NACA 0030 section for the front portion with a flat flap added. This approach did ok, but it requires more alpha than the rest to get the same lift.

    A very notable point is that you seem to be able to choose from pretty wide range of flap percent of cord values without any performance loss.

    For a minimum weight wing, one approach would be to consider maximizing the flap portion since you can build it lighter per square foot. The U35 foil transition to flap is at 44% as compared to 52% for the Blend 1 foil section. Although more flap may be lighter, it may be more of a problem to "actuate" to the desired position.
     

    Attached Files:

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

    Petros Senior Member

    what happens to the pressure distribution if you make the mast only about a third or even a fourth of the total profile? It would have an effective thinner section, and you would have less mast and more fabric.
     
  10. bearfoil
    Joined: Jan 2013
    Posts: 6
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Loch Lomond

    bearfoil Junior Member

    tspeer..howdy. For purposes of discussion, camber is not helpful. What sucks in the beginner is the belief that camber (asymmetric section) produces power by accelerating air over the wing. This is not so, and Bernoulli does not explain this, so a symmetric (planar) exemplar is better, imo.

    Bernoulli support is fostered by the claim that at "0" Angle of Attack, an asymmetric foil produces lift. Again, not so, an honest beginning requires an explanation of "effective" Angle of Attack, which is best understood by a planar exemplar (thin) section.

    In a global sense, when matter is displaced, work can be derived. The best generalized understanding of lift comes from "Flow" (downwash) not velocity. this means Newton, and it is a long passionate discussion.
     
  11. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    A smaller fraction of the chord for the mast is typical for a wing mast. The challenges with a wing mast are that with the smaller faction of chord usually comes less thickness (needed for structural considerations); Also, you have to work out trimming of the sail portion to get the curvature you need.

    I have not seen much posted that gives av average guy much help on how to do a wing mast right. I feel a lot more in control with a fixed flap (even if thin & flimsy) than with cloth. However, some seem much more into the art of sail making and sail trimming.


    I am confident that a really good wing mast with the right shape controls can perform very well.

    If you want to understand what performance to expect, the first step is to use something like a spread sheet that allows you to build a table of points that define the foil shape. The tool needs to be able to do things like change the curvature of the sail portion, the width of the sail portion and the rotation of the wing mast.

    It needs to be broken down with lots of points (say 200). The fist column will be your "x" values that need to be sorted such that the fist point it at the top trailing edge, and the the points are in sequence going from back to front across the top and then front to back across the bottom. These will be saved into a text file with a .dat extension. The foil name should be on line 1 and then just two columns of numbers from line 2 to the end.

    Evaluation is a a snap with the XFLR5 freeware IF it will run the shape you have created. You get a feel for this pretty quick. Steps are:

    • Call up your foil (File, navigate, scroll and select)
    • Go to "Design" and do a Refine Globally as a minimum. I end up with anywhere from 100 to 250 points. If it just hangs up when you run the foil you many want to drop down to less points.
    • You really want to evaluate your foil with a length of 1. If needed the normalize commend under Design will do this for you.
    • If your nose and TE are not centered up and down, you should use the de-rotate command under design.
    • Next go to the Analysis command and select batch
    • I like to use the List method to define RE with just two points as I have previously discussed. One you set this it stays until you change it.
    • Then select Type 1 and hit Analyze down in the lower left to generate the polars for the foil.

    The program will do a lot more. However, when I was doing my evaluations, it was pretty much just call it up, refine global, and analyze. I would then tweak something on the foil in the spreadsheet. My spreadsheets generate a new name based on the changes. This helps keep track and lets you get comparison runs on the polar graphs. I would save the new version as the same dat file, call it up again run it & see how the results change.

    Under the design command the program will actually generate NACA foils inside the program. You may want to do this for comparisons. Regular aviation foil DAT files are also pretty easy to find. Just be careful that the points are in the correct order.
     
  12. Petros
    Joined: Oct 2007
    Posts: 2,936
    Likes: 148, Points: 63, Legacy Rep: 1593
    Location: Arlington, WA-USA

    Petros Senior Member

    I have in mind a means of controlling the sail shape (both camber and twist) of the fabric part of the sail. So I can have the mast much smaller than you show as a percentage of the cord, I need to find a shape for the mast that will work for my experiment.

    Would using the profile you show work for that, or would another shape optimized for the small mast be better?

    BTW, did you see my PM to you?
     
  13. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    I just smile when I see many of these discussions.

    You really do not need much theory at all to do what you really need.

    Air flows over and under the the foil.

    "Laminar" sections are usually low drag, but can be tricky.

    Air flow separating from the top side is usually bad.

    Anyone with much experience trimming sails probably has some feel for how shapes affect air flow and how properly directed airflow generates lift.

    It is sad that people have dream up a bunch of junk to describe how airfoils work.
     
  14. P Flados
    Joined: Oct 2010
    Posts: 604
    Likes: 33, Points: 28, Legacy Rep: 390
    Location: N Carolina

    P Flados Senior Member

    I saw the PM, and did some stuff, but I wanted to focus on finishing up a first pass on trying to understand the basic of what a "reversible camber" needs to look like and what it will do.

    Along the way, I did try playing around with wingmasts some and I tried using a NACA up front for my wing study.

    With a wingmast, you really need to figure total chord, design loading and out how thick you need the front portion (based on build materials and methods), to meet your structural loads. With this, then you can pick some shapes and try them.

    I did put together something of a road map on how to run foils above (the easy part).

    I am better than average with spreadsheets, but setting up a tool to manipulate the applicable parameters and generate the Dat files for a wingmast study is not as easy. Actually taking the txt files I attached to my summary post and renaming them to DAT files could get you started.
     

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

    Petros Senior Member

    I know how to design efficient structures, that is what I have been doing for much of the last 20 years. I just have no motivation to start running foil sections since I would have to learn how to use new software. So I was hoping someone could just give me a good mast section shape I can use for my experiment.

    I am planning on going full cantilever mast, but if it becomes too heavy I may revert to a simple pivoting cable stay, or perhaps even a tripod type strut braced mast.
     
Loading...
Similar Threads
  1. mfly
    Replies:
    2
    Views:
    904
  2. Sailor Al
    Replies:
    183
    Views:
    6,871
  3. Dave_S
    Replies:
    3
    Views:
    941
  4. Inquisitor
    Replies:
    5
    Views:
    1,563
  5. Slingshot
    Replies:
    58
    Views:
    5,327
  6. schakel
    Replies:
    4
    Views:
    2,110
  7. WSW2016
    Replies:
    17
    Views:
    2,886
  8. Surfer Naval Architect
    Replies:
    6
    Views:
    2,177
  9. Wally2
    Replies:
    3
    Views:
    1,420
  10. pmudesign
    Replies:
    14
    Views:
    1,813
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.