Basiliscus

Members
  • Content Count

    802
  • Joined

  • Last visited

  • Days Won

    2

Basiliscus last won the day on October 24

Basiliscus had the most liked content!

Community Reputation

237 F'n Saint

About Basiliscus

  • Rank
    Anarchist

Profile Information

  • Location
    Port Gamble, WA, USA

Recent Profile Visitors

6,510 profile views
  1. Basiliscus

    Boats and foils comparison

    Why do you think both foil wings produce the same lift and the strut produces zero lift?
  2. Basiliscus

    Boats and foils comparison

    I just realized that I was using a different definition for apparent wind angle. I typically use the angle between the velocity vector through the water and the apparent wind as the apparent wind angle. But I'll bet Juan K was using the apparent wind measured from the yacht centerline. The difference is the leeway angle. You'd need to add leeway to Juan K's apparent wind to get the angle I was using. So if you add a couple of degrees of leeway you get 15 deg. 15 deg upwind at 45 deg TWA is a boat speed that is 1.9 times the true wind speed. 15 deg downwind at 140 deg TWA is 3.2 times TWS. I think those numbers are more reasonable 20 deg AWA (including leeway) is 1.2 x TWS upwind and 2.5 x TWS downwind. I'll leave it to you as to which set of numbers are more likely.
  3. Basiliscus

    Boats and foils comparison

    13 deg AWA at 45 deg TWA implies a boat speed that is 2.4 times the true wind speed upwind. Maybe. 13 deg at 140 deg TWA implies a boat speed 3.6 times the true wind speed. Sounds optimistic to me, but not infeasible.
  4. Basiliscus

    Boats and foils comparison

    Yes, the induced drag from the lifting line analysis has to be added to the profile drag from Xfoil. And all the other drag sources. It's really a big bookkeeping exercise. Each source of drag has a different physical cause. With regard to the profile drag, in the past I've added a sheet with section data that cover a range of angles of attack and Reynolds numbers. Then I used the chord and local lift coefficient from the lifting line to look up the profile drag coefficient and add that to the induced drag to get the total drag at each station. Then integrate the total drag over the span to get the final number. Yes, exactly. So why go to the bother of two sails? Landyachts experience 40 kt apparent wind and a lot more. I've personally clocked landyachts doing 5 - 6 times the true wind speed, and the apparent wind upwind is approximately the same as the yacht's speed is downwind. You're probably talking around 20 degrees apparent wind angle for the AC75 and a landyacht apparent wind angle is more like 14 deg.
  5. Basiliscus

    Boats and foils comparison

    In the lifting line analysis, the section shape shows up in two ways. First, as you mention, is the zero lift angle of attack. The more camber the section has, the more negative the zero lift angle will be. A double skinned main is no different from a single skinned mainsail in this regard. Both shapes would create lift at zero angle of attack (if you could keep them from luffing). The second way is the lift curve slope. In inviscid thin airfoil theory, this has a value of 2 pi per radian. In principle, a thick section can have a little higher value, but the boundary layer thickness tends to offset this. For viscous flow, a value of 0.1 per degree is a reasonable guess. You can get both the lift curve slope and zero lift angle from Xfoil. Just tell it to calculate the pressure distribution for zero lift and it'll give you the zero lift angle. Do it again for zero angle of attack and divide that by minus the zero lift angle and you'll have the lift curve slope that is good for the linear lift range. Or you can fit a line to the lift range of interest and use the intercept and slope of that line. The lifting line method is really concerned with the influence of the wake and doesn't care what the actual angle of attack is. You can get the same lift by having a symmetrical section at an angle of attack, or a cambered section at a low angle of attack. The same goes for the chord. You can get the same lift with a long chord at low angle of attack or a short chord at high angle of attack. The associated induced drag will be the same. The difference comes when backing out the planform shape or twist distribution from the spanwise lift distribution. On the whole, I don't really get the motivation for the double skinned mainsail. It can eliminate the separation bubble you'd get on the windward side with a single skinned mainsail and rotating wingmast. But the soft sail rig on the BOR90 trimaran USA17 achieved attached flow on both sides of the mast all the way back to the sail track over 2/3 of the span. It was only in the lower 1/3 of the span, where the angle of attack was quite low because of the jib that there was a windward separation bubble. And that might have been eliminated if the mast could have been twisted Crazy Coyote style. I suppose it helps to fair mechanisms used to control twist, but those are only at the head and foot. But it seems like a lot of weight and complication to achieve that. Landyachts quite successfully use large wingmasts with single skin mainsail at low apparent wind angles and higher apparent wind speeds than the AC75s will see. And the best of the landyacht wingmast/sail combinations are competitive with the best of the rigid wingsail rigs. So you don't need separate surfaces to go fast.
  6. Basiliscus

    Boats and foils comparison

    And here I thought FORTRAN was the "go to" language for data analysis. ;-)
  7. Basiliscus

    Boats and foils comparison

    What you need to do is to trim the sails for best performance for the whole boat. That's not necessarily best L/D for the sail itself. I learned this when I made a simple VPP for a rigid winged landyacht. I arranged the numbers so max L/D was below stall, and I thought I'd get best performance at max L/D. But it turned out I needed to trim for a lift that was higher than that, even though the induced drag was greater. The reason turned out to be the tires. A tire acts a lot like a centerboard or keel. There's a linear range where the side force is proportional to the leeway angle. Yes, a tire has a leeway angle because side force makes the tire flex as it rolls. While the aerodynamic lift is under the control of the pilot, the side force from the tires depends on the aerodynamic load, and the tires can't operate at their best L/D by themselves. For example, consider sailing dead downwind. The side force is zero, so the tire L/D is zero. By sheeting in harder to take the wing past best aerodynamic L/D (including the aerodynamic drag of the body and wheels), the side force loaded up the tires so they got closer to their best L/D. The net effect was an improvement in the yacht's performance, even though the aerodynamic performance was not at its best. With regard to sail profile drag, you may not be able to realize those laminar drag buckets. Laminar flow requires a very smooth surface, and you're not going to see that a sail with its stitching, seams, and waviness from the weave. The atmosphere may also have a lot more ambient turbulence at low altitude. Xfoil has a parameter, Ncrit, that can be adjusted to account for the ambient turbulence on the transition location. You can also artificially set the point where boundary layer transition occurs. You may want to look at the profile drag by tripping the flow to make it essentially fully turbulent. I like to use 2.5% chord for that purpose. In Xfoil, if you use a negative number for the transition location, say -0.025, it locates the transition from the stagnation point instead of the leading edge. That's a good option when analyzing a section at high lift.
  8. Basiliscus

    The new sailing twin skin setup

    It's just the definition of the drag coefficient. You've been using the same relationships all along. You just never thought to look at things in dimensional form instead of the nondimensional coefficients. Compare the dimensional form with the nondimensional form: Dimensional: Di = L^2/(pi*e*b^2*qbar) Nondmensional: CDi = CL^2/(pi*e*AR) Where there is a force term in the dimensional form, it has been nodimensionalized by dividing by (qbar*S). Where there's a length^2 term, it has been nondimensionalized by dividing by the planform area. Aspect ratio is often viewed as being a measure of how skinny the surface is, but here you can see it is really the nondimensional version of span-squared. If you give something a lot of span for its area, of course it has to be skinny. The velocity, density, and area are the biggest factors driving the forces. The nondimensional coefficients are a convenient way of filtering out the biggest factors so you can see the lesser contributions. But it's easy to lose sight of just what the big influences are.
  9. Basiliscus

    The new sailing twin skin setup

    Something to think about. The equation for induced drag can be misinterpreted because of the use of nondimensional lift coefficient and aspect ratio. If you substitute Di=CDi*qbar*S, CL=L/(qbar*S) and AR=b^2/S, what you get for the actual induced drag is Di=L^2/(pi*e*qbar*b^2). Notice aspect ratio completely goes away, and span (b) appears. Area (S) doesn't appear, either, so that may simplify how you handle the jib. It really doesn't matter what the lift coefficient is - what matters is how much lift is being produced. Since qbar=1/2*density*Va^2, the induced drag actually drops with apparent wind speed (squared) when the lift is kept constant. And since the lift is driven by the available heeling moment, it doesn't vary all that much with apparent wind speed.
  10. Basiliscus

    Boats and foils comparison

    For hull hydrodynamic resistance, take a look at Michlet. For hull aero drag, you can get some idea from Hoerner's Fluid Dynamic Drag. On page 3-11, he describes how to estimate the drag of slender shapes by taking the bluff-body drag in the cross-flow plane and the skin friction drag in the direction of flow. 20 degrees apparent wind angle would be a good guess for the hull drag.
  11. Basiliscus

    Boats and foils comparison

    There's a simple way to look at the performance aspects of accelerating to get on foils. Assume that the crew will be trimming out the fast dynamics to keep the boat on her feet, so the slow dynamics are of interest. You can add a fictitious drag component as a D'Alembert force to represent the acceleration of the boat. Now calculate the boat speed just like you did for the static case. You'll get the sail trim and speed for that level of acceleration. If you run the whole polar, you'll get a polar that nests inside of the max speed polar. Repeat for different levels of acceleration. Now you have a set of contours of constant acceleration. If you pick the locus of the maximum points of those contours, that will be the fastest way to accelerate from low speed to high speed. You need to be able to model hull drag at various levels of displacement or heave, of course, in order to add to the foil forces to model takeoff. That's a whole effort in itself. It's possible for the AC75 that there will be two contours that correspond to a given level of low to moderate acceleration. One will be at high speed, when the acceleration drops off as the boat approaches maximum speed. The other will be at low speed hullborne, when the boat doesn't have enough righting moment for the crew to sheet in hard. That's the most interesting aspect of the AC75 takeoff. As the boat accelerates, the foils become more effective, providing more righting moment, allowing the sails to be sheeted in more, and generating more acceleration. You can explore this behavior with a quasi-static VPP using the D'Alembert force approach. You can also assume fictitious thrust to model negative acceleration. The boat can get to points outside the static polar by accelerating to a higher speed and then turning up or down to a new course. To a first approximation, which is probably pretty good for the rates a foiler can turn, assume the transition takes place in zero time with zero loss of boat speed. On a polar diagram, this corresponds to following a ring of constant speed until it intersects the course of interest. This can be used to explore tactical situations. For example, how quickly does the boat slow down when shooting the finish line, so how close to the line does it make sense? How best to get to sailing upwind - is it better to get maximum acceleration on a reach until the boat achieves its upwind target speed and then turns up, or should the boat follow the locus of points that have the best Vmg for each acceleration polar? Is it better to sail a steady course upwind, or to oscillate between footing and pinching?
  12. Basiliscus

    Boats and foils comparison

    Not so. It turns out lines of constant apparent wind speed/boat speed ratio are circles on a polar plot. The lines of constant apparent wind angle are also circles. The two sets of circles are perpendicular to each other, everywhere. Here are the two sets of circles shown on a typical polar plot. The red line is a polar for the BOR90 trimaran. If the boat has a Vmg downwind that is greater than half the true wind speed, the boat speed will be greater than the apparent wind speed.
  13. Basiliscus

    The new sailing twin skin setup

    It's unfortunate that he uses the center of pressure concept. This obsolete approach unnecessarily complicates understanding the notions of stability and trim. The problem comes because an airfoil creates not just the lift and drag forces, but a pure moment as well. The moment does not come from the location of the lift force, which you can see if you consider the angle of attack at which the net lift is zero. There is still a moment generated by the section. This requires the location of the lift to go to negative infinity as the zero lift angle is approached with positive lift and then suddenly switch to being at positive infinity as the angle of attack crosses over into negative lift territory! This mathematically consistent, but isn't helpful with understanding the physics. Instead, consider a moment reference center where the moment is approximately constant as the angle of attack changes. This is known as the aerodynamic center. You can visualize the lift and drag as acting at the aerodynamic center, as well as a pure couple acting on the section. The moment about the aerodynamic center is the same as the moment when the lift is zero. For most sections, the aerodynamic center is not too far from the quarter chord location. The more camber the section has, the greater the moment will be about the aerodynamic center. But, to a first approximation, that moment won't change with angle of attack. In the absence of good drag predictions when you are setting up a VPP, it may be useful to simply assume a lift/drag ratio and a moment coefficient. This will allow you to get on with the force and moment balance so you know approximately what it looks like. Then you can refine the drag prediction once you have a better idea of what the operating point is. You can also vary the lift/drag ratio to see how sensitive the balance is to the drag. If it's not very sensitive, then you may not even need to come up with a good number if you are interested in stability and handling instead of performance.
  14. Basiliscus

    Boats and foils comparison

    When you're trying to calculate the 3D aerodynamics of a wing or sail, if the span is several times the chord, you can approximate the flow by considering two 2D problems. The first is the flow in a plane at right angles to the axis of the surface to get the flow around a cross section. This is what Javafoil and Xfoil do. The second problem is the conditions in the wake as it passes through a plane at right angles to the freestream direction. This is known as the Trefftz plane and is used to calculate the lift-induced drag from the spanwise lift distribution. You need to add the profile drag from Xfoil/Javafoil to the induced drag to get the total drag. With regard to Hoerner's books, he used to publish them himself. When he died, he left a warehouse full of them as his way of supporting his widow. I'd be surprised if she was still alive, so I wouldn't feel too bad about downloading a pdf that someone has scanned and posted online. Rather than trying to calculate the lift from first principles, it may be more productive to estimate it from the righting moment. No matter how hard the wind is blowing, the crew has to balance the boat on the foils. So they will adjust the angle of attack of the sails in order to get the heeling moment that just matches the righting moment available. If you estimate the drag vs lift for a range of lift, you can look up the drag based on the lift derived from righting moment.
  15. Basiliscus

    Boats and foils comparison

    1. Yes. Check out Profili. However, it can only handle single element sections and not multi-element sections like a main + jib. Javafoil is the only free application I know that can handle multi-element sections. Don't forget to add induced drag and allowances for windage and other drag sources. For that, I suggest Hoerner's Fluid Dynamic Drag. You may be able to find it at a local library, or a scanned online copy. 2. No. A closed form solution for CL and CD is not possible for sail-relevant Reynolds numbers. You might be able to curve fit empirical data. If the data were available.