Skeg box drag

Discussion in 'General Paddling Discussions' started by rider, Oct 29, 2018.

  1. rider

    rider Paddler

    Joined:
    Jul 12, 2005
    Messages:
    1,724
    Location:
    Coquitlam,BC
    Hey everybody. Time for me to pick the brains of those sufficiently nerdy... Context- I recently picked up an Impex Cat5, with the main purpose of recreational racing or efficiently traveling through tedious inlets. Had the boat out a couple times now, it feels reasonably quick, and about as neutral as Switzerland, even in light quartering seas. Which means the skeg won't get much,if any use.
    So the question :
    Does anybody know if the skeg box creates any measurable amount of drag? if there anything to be gained by taping it closed? (before you say "tape it closed any try for yourself", I can,and maybe I will, but I don't trust myself to accurately feel such an increment difference in ever changing conditions)
    Thanks in advance!
     
  2. mick_allen

    mick_allen Paddler & Moderator

    Joined:
    May 15, 2005
    Messages:
    2,906
    Measurable? maybe with great difficulty to get the precision.

    But there must be drag for 2 reasons:
    1) water inflow/outflow from the box will cause there to be some turbulence - which means drag.
    2) water within the box is dragged along with the kayak- which means extra mass, which means extra drag.

    So 1/2 the solution is to tape the box closed. The other half is to tape the box closed so it is watertight and no water gets into the box. But that means that the skeg is no use at all . . .

    so choose the first 1/2, but use a gasket instead of tape and slit it right down the middle - the idea being that turbulence is kept to a minimum, but the skeg could still be usable. [Some sailboats deal with their centreboards in a similar manner - maybe check them out.] And take great care never to damage the fragile gasket . . . .

    idea, anyway.
     
  3. Roy222

    Roy222 Paddler

    Joined:
    Jun 2, 2009
    Messages:
    340
    Check your local Universities to see if any have a drag tank. Drag testing is the only real way to know the answer to your question and is ($$$) probably not going to happen. I suspect a good coat of wax on the hull will have as big drag reduction as tapping off the skeg box hole. Who knows, the tape surface might cause more drag then an open hole.

    You might consider making a formed ( fiber-glass) plug that would snap in and out of the skeg box hole.
    At the very least it would give you a talking point with friends.

    Roy
     
  4. Pascal

    Pascal Paddler

    Joined:
    Nov 12, 2012
    Messages:
    55
    Location:
    Vancouver Island, Canada
    Does anybody know if the skeg box creates any measurable amount of drag?

    "Measurable" ... not sure what you mean... With modern dynamometers, data averaging and careful drag test setup, I would suspect any drag differential due to the skeg box would be very easy to pickup, and it would be below 1%, if I have to guess (unless your hull is super smooth and offers very little resistance to start with)... And the test would cost you several 100s $.

    Significant difference? You could probably detect a drag differential above 5% with a cheap digital hanging scale (mine gives me 20g steps on a 15kg scale, so ~4% measurable variation assuming a ~5kg pull from viscous drag (that's for a hull with barnacles, ie. mine). Load your kayak to simulate your weight + cargo, hold it (off set from your own eddy) with your dyno by the bow in a shallow, steady, smooth 6-8 km/h fast current (you got to find a good spot in a river), record several readings for 1 min, average, repeat with your skeg box tapped. Et voilĂ . 5 min job (once you have found your river spot, which may take a month). But I bet you won't see a difference in that range, and as Roy mentioned, removing the barnacles and a wax job on the hull would probably induce a much more dramatic change.

    "Over an eight-hour day, the paddler with the more efficient boat will arrive at camp 16 minutes before his friend who paddles a 2% less efficient boat." (source)
     
  5. AM

    AM Paddler

    Joined:
    Jan 30, 2006
    Messages:
    718
    Location:
    Vancouver
    What a great answer, Pascal!
     
  6. Pascal

    Pascal Paddler

    Joined:
    Nov 12, 2012
    Messages:
    55
    Location:
    Vancouver Island, Canada
    Correction, that should read "You could probably detect a drag differential above ~0.5% with a cheap digital hanging scale (mine gives me 20g steps on a 15kg scale, so ~0.4% measurable variation assuming a ~5kg pull from viscous drag (that's for a hull with barnacles, ie. mine).

    hmmm... That's not bad if other numbers are realistic. Will test if I can detect my rudder drag sometime.
     
  7. Kayak Jim

    Kayak Jim Paddler

    Joined:
    Mar 5, 2016
    Messages:
    146
    Location:
    Comox Valley BC
    I suspect in a natural stream the reading would bounce around by more than that, but it would be an interesting experiment.
     
  8. Pascal

    Pascal Paddler

    Joined:
    Nov 12, 2012
    Messages:
    55
    Location:
    Vancouver Island, Canada
    Not sure it would bounce around much in a smooth current since the boat+load would have a significant inertia. That's why you'd need to choose a relatively smooth current AND average multiples readings, like 10 at least, over 40 sec. or so. But I have to try to know for sure.

    Also, any systematic bias should be cancelled out since you compute the difference between the two values obtained in the same way. The difference is the only thing you are interested in. An exact measure of the boat drag would be more difficult because that would require calibration with a boat/load system of known viscous drag.
     
  9. Roy222

    Roy222 Paddler

    Joined:
    Jun 2, 2009
    Messages:
    340
    I put my statistical books away a long time ago, so this maybe off center a bit.
    The difference in the 2 averages mean nothing if the variation in data points are
    in the same ball park. Or put another way. If you take 10 readings of the same load and the variation is .5%, then the your experimental difference will need to be a lot greater then .5% .
    Say you take 10 readings with open skeg box and the average is 20, and your data ranges between 18 and 22. Then 10 readings with a taped skeg box and the average is 19, and your data ranges between 17 and 21. A statical study would probably indicate that the experiment showed no differences in drag.

    Roy
     
  10. JohnAbercrombie

    JohnAbercrombie Paddler

    Joined:
    Dec 7, 2011
    Messages:
    1,624
    Location:
    Victoria, BC
    I was told by an experienced paddler:
    "Sea Kayaker found that at three knots a kayak had 10% more drag with the rudder down than when it was out of the water. They didn't believe the result and ran another trial. Same result that time. "
     
  11. Pascal

    Pascal Paddler

    Joined:
    Nov 12, 2012
    Messages:
    55
    Location:
    Vancouver Island, Canada
    It depends on the number of readings AFAIK. In both cases (with and without rudder, or skeg box), measurements will have the same type of distribution around the average (probably close to a Normal distribution - bell shaped) and the more readings you take, the closer your average should get to the "true" value. The uncertainty in the mean is inversely proportional to the square of the number of measurements. Obviously, an average from 3 readings would have a much higher chance of being off than an average derived from say, 100 readings, and therefore difference between averages would then be meaningless. Averaging a high enough number of readings should allow to extract differences that are bellow noise level (the standard deviation).

    The uncertainty associated with differences (or sums) is not the sum of the uncertainty for each measurement (which have the same uncertainties associated to them in our case) but their quadratic mean (square root of the sum of the two uncertainties squared). So if I have an uncertainty of 20g for each measurement, the uncertainty of the mean for N measurements would be 20/Sqrt(N)
    and the uncertainty for the difference of the two means should be Sqrt (2 x (20/Sqrt(N))^2) = 20Sqrt(2)/N^(1/4)

    For 10 measurements, the uncertainty for the difference of the mean is 15 g
    For 50 measurements, it's 7 g, and so on....

    So, yes, 10 reading would be borderline to detect a difference if that difference is close to 20 g. But you just have to make more readings. The method is not flawed in theory.

    Please correct me if I am wrong.
     
    Kayak Jim likes this.
  12. Astoriadave

    Astoriadave Paddler

    Joined:
    May 31, 2005
    Messages:
    5,530
    Location:
    Astoria, Oregon, USA
    I remember this comparison, and three per cent seems about right for the flat plank, nonfoil rudders of the day.

    Early days, for a while SK had free access to a sophisticated (for early 1990s) hydrodynamic testing tank, and conducted tests on a broad range of FG boats, most considered to be very high volume kayaks today, I expect. The data were reported along with reviews of the boats in SK. I believe these hard data were used to test and refine the models eventually used to report expected performance numbers in later reviews. Naturally, data from good tank testing are more valid, but prohibitively expensive.

    Mick Allen may know more on this. I thought the tank was in Vancouver, and access was due to the persuasive skills of John Dowd, the original publisher of SK. John is still kicking, I think.
     
  13. mick_allen

    mick_allen Paddler & Moderator

    Joined:
    May 15, 2005
    Messages:
    2,906
    The tests of 5 single sea kayaks and one [unidentified] whitewater kayak were done at BC Research Ocean Engineering Centre for 2 days in 1986. There was another run for 6 more kayaks that I do not have any info on. I do not see any mention of rudders in that first article.

    It is interesting to read that even with their sophisticated setup and measurement that they could not measure any reportable drag difference between all 5 seakayaks below 3 knots of speed:
    SeaKykr1986Fall-Dawson.jpg - SeaKayaker1986Fall - Dawson

    I also think it's quite funny that even the generic whitewater kayak of the day [probably the typical 10-13' torpedo shape] has drag at those speeds that is not that much greater.
    A much later article refers to the 2 articles:
    SeaKykr1993Winter-Spilman.jpg -SeaKayaker1993Winter - Spilman


    [If anyone has the second part of that sea kayaker article, I wouldn't mind a scanned copy of the pages.]
     
  14. Astoriadave

    Astoriadave Paddler

    Joined:
    May 31, 2005
    Messages:
    5,530
    Location:
    Astoria, Oregon, USA
    Mick, looking at the resistance values at 3 knots, it appears the spread is about 0.3 pounds, maybe 0.4 pounds. At 2 knots, the total spread looks to be 0.2 pounds. Yet their data appear to have no inconsistencies of that magnitude anywhere. So, how could they conclude there were no reportable differences below 3 knots?

    Second issue: the claim has been made here that a difference of 0.5 per cent in drag force should be measurable with a cheap digital scale in field conditions. If the SK article limits represent state of the art in 1986 and 1987, what changes have occurred in the thirty years since then to warrant a limit of only 0.5 per cent?
     
  15. mick_allen

    mick_allen Paddler & Moderator

    Joined:
    May 15, 2005
    Messages:
    2,906
    I can only guess that the graph is a poor representation of the spread of the data, and that they are unsure of stating the definitive accuracy of their results at 3 knots. The statement that the curves are a computer representation of the drag is somewhat confusing to me: ie I would have thought that the curves were a direct representation of what they read.

    I would expect that the instrument readouts were typically extremely accurate at that time [- ie rocket science actually existed then! and they were controlled tests in an isolated tank] - but with all apologies to Pascal [ and I'd sure be interested in what could be carefully done] , I am a little skeptical that most external influences could be eliminated so that valid comparable readings could be taken in a field test with simplistic setups and scales.
     
  16. Astoriadave

    Astoriadave Paddler

    Joined:
    May 31, 2005
    Messages:
    5,530
    Location:
    Astoria, Oregon, USA
    Mick, they show their data points, and the fair curves which fit them. In 1986-1987, there were plenty of Fortran programs around, some in Pascal probably, even ones in BASIC that would run on a PC with only 64K memory, which would develop the curve and give quantitative statistical analysis. As you say, however, there might have been mechanical or electronic flaws in their data gathering, not displayed in the graphs.

    I would definitely give them the benefit of any doubt on their conclusions. My earlier suggestion that the curves seem to show valid separations at and slightly below 3 knots assumes no consistent flaws in their instrumentation. But they did the data gathering, and they may have known something not discussed in the article.
     
  17. mick_allen

    mick_allen Paddler & Moderator

    Joined:
    May 15, 2005
    Messages:
    2,906
    Yes, I agree - so if drag is as they say, likely all semi normal shaped kayaks of similar size and loading to above are quite similar in drag around the 3 knot speed. Kind of interesting.
     
    Astoriadave likes this.
  18. rider

    rider Paddler

    Joined:
    Jul 12, 2005
    Messages:
    1,724
    Location:
    Coquitlam,BC
    Given the year that test was performed, it makes sense that the whitewater kayak was relatively close up to 3 knots because they were also long and pointy.. Great input :) I was sort of hoping that someone brings up a similar test somebody has done and recorded with skeg box as a factor of some sort, but i guess there's a good chance that was never really done.
     
  19. Astoriadave

    Astoriadave Paddler

    Joined:
    May 31, 2005
    Messages:
    5,530
    Location:
    Astoria, Oregon, USA
    Rider, at low hull speeds, given the small differences in drag over the range those studies show, I suspect open or closed, that skeg opening is not going to make much difference, especially if you make flaps so the opening is barely a slit when the skeg is not deployed.
     
  20. Pascal

    Pascal Paddler

    Joined:
    Nov 12, 2012
    Messages:
    55
    Location:
    Vancouver Island, Canada
    Regarding the graph you published, I have no clue ... Are these points single data point or averages? There is no error bar attached to them, so it's perplexing. I don't know how many readings they took. The only way would be to actually have a look at the raw data and the full experimental protocol/set up.

    Regarding the set up I was envisioning to measure rudder drag, it's completely different because you can rapidly alternate between Rudder In and Rudder Out to register differences and log them at very close intervals. The effect on the drag is likely felt within 2 sec. or so max I would think, since it's a simple, small, hyrodynamic form, and any dirty water it generates will not affect drag upstream. The time constant for environmental variations that could affect the pull of the kayak would be longer than this 2 s lag (I think) IF there is no wind AND if the water chosen is relatively laminar with no big eddies coming in and out rapidly. I don't think it needs to be perfect if the kayak is loaded and yaw is somewhat controlled. So you end up with a series of deltas that can be accumulated rapidly, with the accuracy on the average raising as you do so. It would be fun to try anyway. I have an idea how to control yaw, but I still have to find the spot and the time.

    For the skeg box set up, it would be much harder, I agree, as you would not be able to switch rapidly, and environmental variables could easily interfere and ruin the experiment. But I also agree that the drag is probably negligeable compare to the overall drag of the boat.
     
    Last edited: Nov 5, 2018