Wing sail questions

Many thanks for sharing this link, as it it always good to have in mind the basics when dealing with flows around profiles. I notice that T.Speer also used Ncr = 9 in his theorical approach, having certainly the same concern that I have, regarding the absolut necessity of crossing numerical calculations with experimental results.

(extract from tspeer.com)




When the Günther brothers, in 1925, launched the Baümer Sausewind, Bäumer Sausewind - Wikipedia https://en.wikipedia.org/wiki/B%C3%A4umer_Sausewind, it starts spreading the Prantl's lifting line theory all around, it opened the path to elliptical wings, followed by R.J. Mitchell R. J. Mitchell - Wikipedia https://en.wikipedia.org/wiki/R._J._Mitchell and his aerodynamicist, the canadian Beverly Shenstone, with their famous Supermarine Spitfire, that made his first flew in 1936.
In the meantime, the boundary layer theory, also formulated by Prantl in 1905, had a slower acceptance. The spread of his theory had to wait until 1931, date of the publication by Prantl of his book "Abriss der Stromungslehre". We gain from the further and latest developpements of this theory our actual view on the subject. Not only the induced drag is to be chased, but also the parasite drag,. (https://www.faa.gov/regulations_policies/handbooks_manuals/aviation/phak/media/07_phak_ch5.pdf p.5-5 Lift/Drag ratio). Minimizing the induced drag is only one part of the problem. Again, the structural needs are to be taken into account, for the whole design to be consistent.


To thank you again, please allow me to share, in return, a small insight of this study, made in the scope of the "little America's cup", presentato uno dei profili studiati e le sue prestazioni determinate nella galleria del vento dell'Aermacchi, for the determination of the aeodynamic coefficients of profiles to be used for "aerofoils" (as the wingsails were named, at the time).



I wish my next post will benefit again from your comment. Stay tuned.

Cheers,

 

Hi Dustman, Quite a while since I've started the calculations. Here is a how I study your questions, what results I have, and how, in my opinion, your main question may be answered :

Strategy
Following the path of Bill Steane, with his C-Class Moutain Lion [4], we choose to compare a simple rigid rectangular sail with a classical sail set, composed of a mainsail and a jib, mounted on an hyptothetical catamaran, whose dimensions are those given by Dustman. Bounding our calculations to upwind navigation and to fixed wing mast, to simplify the study, the calculations are made using three combined technics :

• 2D Viscous fluid flow analysis, using XFOIL software [5], to determine the aerodynamic coefficients of the 2D wing profiles. It is widely used, both in naval engineering and aeronautics. The original NACA63012 profile is analysed, as well as a thicker profile, hereby nammed 'NACA63025'. The points defining the NACA63012 profile can be downloaded here [6].

• 2D1/2 lifting line analysis, using an in-house code, developped in the scope of the design of the appendages of maxi-catamarans for the Round the World records, regulary used also for keels, rudders and daggerboards of cruising monohull and multihulls. This lifting line analysis [7] uses the output of XFOIL to integrate the 2D aerodynamic coefficient of the NACA63012 and the NACA63025 into the calculations of the induced drag of the wing planform.

• Velocity performance analysis, using an in-house code, also particulary developped for multihulls. This code differs from a regular VPP in the way that every parameter can be turned into a variable, meaning that you can either calculate the velocity polars of a boat, or optimize any parameter of the boat in order to achieve a specific polar. From the ORC sails models embedded in the code, I determine the stability curve of the catamaran's platform and define a classical sail set for the final comparison, bounding the analysis to the lateral stability, for simplification purpose.

In this study, We first use the Velocity performance analyser to select a regular sails combinaison to be 'our horse to beat, from stability calculations. Then, we analyse the 2D aerodynamic coefficients of the two wing profiles using XFoil, respectively the NACA63012 and the NACA63025. Finally, we compute the aerodynamics coefficient of the two wings, taking into account their planform, and bring the results of the ORC model of the regular sails set, to the comparison.

Sketching the project
The dimensions of the platform of the catamaran given by Dustman are presented here. Please note that the hull beam and the hull height are being arbitrary set, from our personnal experience, and without being required by Dustman. These additionnal dimensions are required for our analysis to be performed, but may not reflect any ideal dimensions for a boat of this size. The same goes with the global hull's shape, that is represented here only for the visual experience.

Boat measurements

• Displacement 455(kg)
• Overall lenght 7.32(m)
• Hull beam 0.580(m)
• Hull height 0.732(m)
• Beam between center of buyoancy 4.88(m)

Above, the Figure 1 summarizes the three configurations tested. The green lines represent the wing sails, whose dimensions are those given by Dustman. The blue lines represent a classical sail set, that is drawn according to the output of the Velocity performance analyser. The planform of the two wing sails coincide in side view. In top view and front view, only the NACA63012 profile is drawn. The NACA63025 profile would be at the same place, appearing thicker in these two views.

The geometrical centers of the each sail and wing sail are represented by a cross. The aerodynamic forces are represented by vectors, whose origin is set at the effective height of their application point.

Stability considerations
The platform of the catamaran has unusual dimensions. Her beam is wider than one would expect, considering her lenght. The stability considerations and the consequent choice of a regular sail set, to be compared with the wing configuration, are affected by this inconsistency. In order to still be able to draw some conclusions, the lesser evil is arbitrary choosen, thus leading to the study of an inconsistent regular sail set. However, as we take caution to also study the catamaran stability, we are able to take into account this inconstancy in our analysis.

The boat displacement is taken as the sum of the boat weight 300(kg) and the crew weight of 150(kg), laterally positionned on the deck line. Again, it's an arbitrary choice, made for maximizing the surface area of the classical sail set that is returned from the Velocity performance analyser. The hull draft is taken equal to 0.125(m) and the height of the center of effort of the hydrodynamic forces, acting on the daggerboard, is taken equal to 0.250(m) below the bottom of the hull.

The maximum righting moment is attained for an heel angle of 3[deg], and reaches the value of 14.6[kN]. Please note that the linear variation of the righting moment curve, below the maximum point, is approximative and is the result of having taken a linear variation of the volume of the hull, with the sinking. The value of the maximum righting moment is un-affected by this choice. On the stability curve, it can also be verified the crew position taken in the calculations, as being laterally positionned on the deck line of the hull, ie at 4.88/2(m) from the centerline.

Now that the stability curve is known, we can take the dimensions of the wing sails, given down below, in accordance with the requirement of Dustman, and define the surface of an equivalent regular sail set, giving the same forces and the same righting moment to the catamaran.

Wing measurements

• Height 4.88(m)
• Cord 1.22(m)

Two approaches are presented downbelow. The first approach uses the Velocity performance analyser, to take into account the sail model given in the ORC VPP [8], resolving in particular the height of the center of pressure, for the heeling moment to be accurate. The second method is a 'other things being equal' approach, used in particular to determine the height of the mast of the regular sail set.


The figure (3) shows, respectively in black, blue and green, the projected surfaces of a minimum regular sail set, allowing the maximum righting moment to be attained, the surface that is choosen for this study and the total surface of the two wing sails, as proposed by Dustman.

This representation confirms that, due to the great width of the catamaran, in comparison with its lenght, the proposed surface under-powers the catamaran. The choosen surface is a compromise between the two, but it should be noted that none of the choosen surface or the proposed surface, are able to make the catamaran attain his righting moment. A direct consequence is that the driving forces of these two sailplane equilibrates the drag of the two hulls.

Second approach. In order to set the height of the mast of the regular sails set, we write the lateral stability equation, in a reference framed fixed to the boat, by requiring the sum of the forces momentums to be null. Keeping the equations as simple as possible, the crew position is assumed to be at the boat axis, as well as the origin of our reference frame. The daggerboard position and its surface stay the same in the regular sail set configurations and in the biplane configuration.

Maths cannot be displayed here, but it can be shown that, from geometric considerations, we see that the height of the center of pressure of the regular sail set is to be at the same height as the height of the center of pressure of the wing sails, for the two sailplanes to deliver the same heeling moment, if their heeling forces are the same. Our next calculations show that the height of the center of pressure on the wing sail is approximatively 2.55(m), above the base. Taken this center to be located at the mid-height of the mast, we establish that a regular sail set equivalent to the proposed wing sails would be composed of a mast, whose lenghts would be 5.10(m), and whose total surface would be 11.9(m²).

As the choosen surface do not respect the 'other things being equal' principle, cares should be taken in the interpretation of the results of this study, especially regarding the concern (G), in relation with the aspect-ratio.



2D Viscous calculations
Viscous flow calculations on the wing profiles NACA63012 and NACA63025 are made with the software XFoil 6.99. The range of incidence is choosen to be in accordance with an upwind navigation, at moderate or high speed. The parameters influencing the boundary layer are set as follows :

Simulation parameters

• viscosity 1.0e-6(m2/s)
• Reynolds number 1.0e6
• Turbulence coefficient 9

The choice of these parameters has been discussed on the forum. In particular, the choice of the turbulence coefficient, which undertakes the sensibility of the boundary layer to disturbances, could have been set to the lower value, to better represent a real flow. A lower value of turbulence coefficient would have for consequence to move forward, on the profile coord, the stall point of the boundary layer, resulting in a lower lift coefficient and a higher drag coefficient. Nevertheless, as it has not being found no reference of measurements made in this conditions, we choose the value of 9, being able to compare our calculations with those that could be find in the litterature [9]. Thus, the transposition in reality of the forces calculated should take into account the optimistic character of the results calculated in this study.

Turbulence coefficient [6]

Situation Ncrit
sailplane 12 to 14
motorglider 11 to 13
clean wind tunnel 10 to 12
average wind tunnel 9
dirty wind tunnel 4 to 8

Also important to notice is the fact that, according to the XFOIL documentation, results above the stall incidence of an arfoil are not considered as being accurate. For our purpose, we have to calculate the aerodynamic coefficients above this limit angle. Although our calculations have all converged, the results may differ from those obtained in reality. Considering the initial approximation made by the choice of the software itself, the results obtained above an Apparent Wind Angle of 12[deg] to 15[deg] should be taken with care. In particular, above this limit, we observe in our results an increase of the lift coefficient. A similar effect has been observed in [10], but nevertheless, additionnal cares are to be taken for results obtained in this range.

During simulations, it has been noticed a great difficulty for the calculations, above the limit angle, to converge. Solutions have been found with panels refinements over the whole profiles (.GDES CADD option) and, locally, in way of the leading edge (PPAR option). In the picture above, we can see a typical output of the software. The boundary layer is represented by the yellow and blue lines, color coding respectively results in relation with the extrado and with the intrado. The pressure coefficient is given in a graphic representations, and the parameters and the result of the calculation are given on the top right.



Under the angle limite of 15(deg), the calculations of the aerodynamic coefficients of the NACA63012, nammely '_12' are in good accordance with wind tunnel results.The limit angle and the value of the lift coefficient are found to be respectively measured at 15(deg) and a coefficient of 1.2, for a Reynolds Number equal to 0.9x10E6. It is thus assumed that the same results for the NACA63025, nammely '_25', could then be comparable.

A remarkable result is the behaviour of the boundary layer on the NACA63025 between 16(deg) to 18(deg) of incidence. The decrease, then the increase in the lift coefficient maybe caused by the presence of a bubble of recirculation in reality, resulting in a bi-stable flow in this range of calculations. The calculations in this domain, for the NACA63025, should be confirmed. Cares should be taken for these calculations.

2D1/2 Lifting Line
The classical lifting line method [11] is used to calculate the induced drag due to the trailing vortices and their downwash on the lift of a finite wing. This theory establishes a relation between the downwash and the aspect-ratio of the planform. Our implementation of the lifting lines couples the viscous calculations of XFOIL with the classical lifting line theory, to produce more accurate results for the drag and lift calculations, taking into account the influence of the boundary layer on the profile. We have successfully used this method to determine the navigation polar of a military drone, equipped with two rigid wings of low aspect-ratio.

The coupling between the 2D viscous calculations and the lifting line is realized by an interpolation of the calculations points, presented above in table [1], in accordance with the local flow angle on the wing. As an example, the figure below shows together a set of calculations points, and its interpolation. The mathematical model used is a cubic spline interpolation.


It could be noted that the interpolation passes by every point with accuracy. Nevertheless, the degree of the polynomials associated with the cubic spline do not represent correctly the linear domain of variation of the lift curves. As a consequences, the results obtained by this method alos present a similar behaviour, in the interval between 0(deg) and 12(deg) of local incidence, that is to be found after the integration of these regressed coefficient along the lifting line. As a result, the calculations of the aerodynamic lift made in this interval suffer from a pessimistic evaluation and from two curvature inflexions in the interval. On the contrary, we should also note that an overshoot of the expected maximum value in the interval 12(deg) to 15(deg). The conclusions should take into account these considerations.

ORC model
A few word about the use of the ORC sails model, as used in the context of this study. Since this model is widely used, and benefit from a continuous development, it constitutes a reference for our study. As described in the part 5.4 Total Aerodynamic Lift and Drag, the induced drag expression is similar to the classic Prantl's equation, with an aspect-ratio that is tuned using a coefficient, representing the contribution of the roach, the relative position of the mainsailhead and the jib, the overlap of the headsail, and the depowering.

As stated in the section 5.1.3 Optimization and De-powering, the whole aerodynamic model is to be adjusted using the FLAT coefficient, which applies directly on the aerodynamic coefficient of the model. A FLAT coefficient tops the lift and drag coefficient to their maximum values, as to reflect the trimming of the sails during navigation. In our calculations, we have de-powered regular sail set, with the following parameters :

ORC model de-powering scheme as used in our method

FLAT = 0.62
cl = CLmax*FLAT
cd = (CDmax+CDi)*FLAT
A notable result is that the lift coefficient, obtained with this FLAT coefficient, does not reach its maximum value of 1.5 in our method. This choice is an attempt to reproduce the loss of lift occuring on a sail plane due to the downwash effect on the lift, that is not directly adressed in the simple ORC sail model, whose values are tuned for an internal usage. This choice should be reviewed in the conclusion of the study.



Results
The calculations that have been presented are agregated in this section. The 2D viscous calculations of the NACA63012 and the NACA63025 are processed by the modified lifting line method, to be compared with the results obtained for the regular sail set, calculated with the ORC model. We first present the aerodynamic coefficient of the sailplane and of the wing, averaged on the height of their respective planform, taking also into account the wind gradient. The results are expressed as functions of the Apparent Wind Angle at an altitude of 10(m), as it is the case for the results that follows.


In the figure (5), we notice that the variations of the aerodynamic coefficient of the ORC model of the regular sail are small, compared with the aerodynamic coefficient obtained with the modified lifting line method. This effect may be due to the combined fact that, in the ORC model, the sails are always tuned as to achieve the highest sailing speed. Also, the wind gradient influence is not integrated along the height of the sailplanes, rather adjusted at the calculated centers of pressure. Which is an evident flaw of the IRC model.

On the contrary, the integration of the wind gradient in the modified lifting line method do produce important variations in the aerodynamic coefficients of the two wings. At low Apparent Wind Angles, the inflexions on the interpolated 2D coefficients occurs on a wider range than in the 2D results, as the local flow incidence on the top of the wing reaches respectively the linear part and the parabolic part of the 2D lift coeffients and drag coefficients. However, this influence is not awaited at medium and high angle of incidence, with regards with the choosen range of calculations, and the large variations of the aerodynamic coefficients are explained by the variation of the 2D aerodynamic coefficients themselves.

As a consequence, it could be noted, on this figure, the following remarks :

• At low angle of incidence, the CL of the wing sail is the highest, together with its CL/CD ratio. Between 20(deg) and 30(deg) of AWA, the NACA 63025 delivers the best CL, together with the best CL/CD ratio. Above 30(deg) of AWA, the regular sail set offers the best CL, together with the best CL/CD ratio.

• The NACA63012 is more efficient at low AWA, when the NACA63025 is more efficient at medium AWA.

• The regular sail set has the wider range of use, if considering the CL/CD ratio, that do not vary much in the medium range and the high range of the AWA.

These remarks applies on the sail and the wings alone. If we now consider the efficiency of the three configurations, with regards to sailing performances of the catamaran, the driving force and the heeling moments should be analysed in details.



About the Figure (4), the following remarks could be made :

• In general, the driving forces decrease with the AWA. The regular sail set has the less increase in the calculated range, and shows the better driving force in the medium and high range. An regular sail set equivalent to the wings would gives the best driving force above 30(deg) of AWA.

• The range of efficiency of the wing sail whose profile is the NACA63025 is narrowed, and its difference with the other configurations, in this range, is small.

• The wing whose profile is the NACA63012 has the greatest driving force at low AWA, below 20(deg).

In addition, these remarks shall be tempered, considering also the comparison between the heeling moments created by these configurations.


In the figure (5) 'MH_' is the Heeling moment, and 'ZCE_' is the height of the ceneter of pressure, above HBI for the wing, and above HBI + BAS for the regular sail set. In general, it should be noted that none of the studied configuration, are able to reach the maximum righting moment of the boat.

In addition, the following remarks could be done :

• The local variations of the lift coefficent and height of pressure, already seen before, define roughly three domains : One domain, from 16(deg) to 19(deg) of AWA, close to the wind, where the NACA63012 profile gives an greater heeling moment thant the one given by the NACA63025. A second domain, from 20(deg) to 29(deg) of AWA, where it is the opposite. And a third domain, above 29(deg) of AWA, where the regular sail set gives the greater heeling moment.

• Since the surface ratio of the regular sail set over the wing sail is 2:1, and with the considerations that has been made regarding the stability of the catamaran, it can be deduced that an regular sail set, strictly equivalent to the wing profile, would stand between the wings with the two NACAs, in the first and second domain defined below, and would give the highest heeling moment in the third domain, above 29(deg).

Answers to the original questions
We see that there is quite a lot of flaws in the proposed strategy, for whose where no deductions can be made, and complementary studies are required. Above is presented a list of the main flaws, together with adapted solutions.

Calculations issues / solutions

• Comparison of wings with a regular sail set whose aspect-ratio and surface gives a very different aerodynamic behaviour / Add to the study another regular sail set, to be compared directly, include the analysis of this new regular sail set in the conclusions.

• Catamaran platform underpowered by all the studied configurations / Confirm what stability is really awaited for this platform, review this conclusions.

• 2D Viscous calculations have been made far out the limit of the software / Redo full LES or RANS Navier-Stokes calculations on the whole range. Review the calculations and the conclusions

• Overshoot and variation of the 2D1/2 lifting line / Correct interpolations model to clean the outputs from un-wanted behaviour, review the calculations and the conclusions

• Use of the ORC model for comparison / Review the methodology for establishing the regular sail set aerodynamic coefficients, propose new models.

Now let's propose some hints for answer :

A : What's the optimum naca profile(chord to width) for a wing sail? I'm thinking 20-25%?


That's depend on the AWA you will navigate. Regarding to speed you want to attain and the race course, a thinner profile will be more efficient at low AWA, while a thicker will be more efficient at medium AWA. The 25% seems to be a upperlimit, considering its wetted surface and the narrow band of interest of its propulsive force.

B : A thicker foil would stall at a greater angle of attack than a thinner foil, would have a greater range of useable angles of attack, would be more forgiving?


Clearly, the answer is no, on the contrary of what i've said previously in the post. My mistake. A (fixed) wing is tedious to trim. The high variations of the aerodynamic coefficients and the relatively small domain of AWA where these wing sails are efficient, do require from the sailor a more intense and quick trimming action, for this efficiency to be keept.

C : I thinner foil would allow you to sail closer to the wind?


Yes. see answer (A).

D : The thinner foil would have a better lift to drag ratio?


The Lift/Drag ratio of the foil vary with the AWA and the profile. Besides, we see that the question of the L/D ratio do is not pertinent enough for a choice to be made, among different foil profiles. The driving force and the heeling forces are a combinaisons of the lift and drag generated by the foil. A catamaran needs to fly one hull to give its best performances. In that sense, the drag of the sailplane in not always bad.-

F : Are there any real disadvantages of a wing sail from an aerodynamics perspective?


Bonus answer : Yes, as seen in the answer (B), the aerodynamic coefficients of a wing sail, composed of only on fixed part, has a limited range of efficiency. Which is an obvious disadvantage compared to a classical sail. Outside this range, it's aerodynamic capabilites just fall down. Even with wing composed of 2,3,4, or more ailerons, this fact won't change, and the trimming of such sails should be contiuuous, and executed with velocity and precision. From that perspective, a traditionnal sail set is more tolerant. One must also add that, the gain of using a wing sail upon a traditionnal sail are far to be proportionnal to the additionnal efforts required from the sailor.

G : Since the wing sail is a more efficient lift device can you get away with a little smaller sail or a lower aspect ratio for similar performance?


I would not recommend to lower the aspect-ratio of the wing. With or without aileron, the aspect ratio of each wing og the bi-plane should be at least equal to the aspect ratio of the equivalent mono-sail.

H : How much more efficient is the true foil than the standard sail, really?


Besides the results obtained for a fixed wing, we can anticipate that a pivoting mast can extend the range of AWA where the use of foils constitutes an advantage. But we have also see that such configurations should be trimmed more frenquently and with more accuracy than a traditionnal sail set. As we get close to the wind, or when the average boat speed is high enough, then a wing made of one single part could deliver +50% of additionnal driving force. When the boat speed is lower, then a traditionnal sail set will be the simplest option. Wings can also be efficient in that medium and high AWA range, but in that case, it should be asymetric, composed of multiple elements, also allowing a control of the twist. Which means a great increase of technologic issues to overcome. Not to mention, again, the intrisic difficulty to adapt, in real time, the wing shape, to the wind.

J : Is it more efficient to use a wing sail like a lift device or drag device downwind?


Bonus answer : I look at the downwind calculations I've made on the wings for the military drone I've studied. Considering also the data given in the ORC documentations, I would say that, downwind, anything that produces drag with the less frontal surface is efficient. A wing profile has been made for small angles of incidence. As Erwann said, only the linear portion of any wing profile is to be exploited.

References
[1] Wingsail on Wikipedia, the free encyclopedia, Wingsail - Wikipedia https://en.wikipedia.org/wiki/Wingsail

[2] Order of Magnitude, an history of the NACA and NASA, https://engineering.purdue.edu/~andrisan/Courses/AAE190_Fall_2001/Orders_of_Magnitude.pdf

[3] Thread by Dustman "Wing sail question", Wing sail questions https://www.boatdesign.net/threads/wing-sail-questions.67682/

[4] Little America's Cup Book by François Chevalier, Chevalier Taglang: LITTLE AMERICA'S CUP BOOK - four side stories - PART ONE http://chevaliertaglang.blogspot.com/2015/06/little-americas-cup-book-four-side.html

[5] XFoil software by Drela and Giles, https://web.mit.edu/drela/Public/web/xfoil/

[6] Airfoil Tools Website, Airfoil plotter (n63012a-il) http://airfoiltools.com/plotter/index?airfoil=n63012a-il

[7] Extended Lifting line theory for viscous effects, Jose Rodolfo Chreim, https://www.researchgate.net/publication/341668932_EXTENDED_LIFTING-LINE_THEORY_FOR_VISCOUS_EFFECTS

[8] Offshore Racing Congress, ORC VPP documentation 2021, Chapter 5. Aerodynamic Forces, https://www.orc.org/rules/ORC VPP documentation 2021.pdf

[9] Wind tunnel test data of NACA63012, https://www.chegg.com/homework-help...-following-lift-curve-slope-stall-s-q25502354

[10] Wind tunnel tests of a wing at all angles of attack, https://journals.sagepub.com/doi/full/10.1177/17568293221110931

[11] Drag Coefficient & Lifting Line Theory Aerospaceweb.org | Ask Us - Drag Coefficient & Lifting Line Theory https://aerospaceweb.org/question/aerodynamics/q0184.shtml

 

In supplement, I've made, using the same tool of mine, a comparison of a rectangular wing sail and an elliptical wing sail, both of the same span and same area. The induce drag and the surface repartition in height are thus the only parameters that differs from the two configurations. I let you draw your own conclusions by yourself, but to me, one thing is clear :
Due to the highest surface area on the top of the rectangular wing, its performance are not so far from an elliptical wing. If both are orientable, the maximum gain obtained by the elliptical configuration is about 5%, compared with the rectangular configuration. Not so important... Nevertheless, it is possible that this gain may become greater if configurations with higher aspect-ratios are studied. I will do this exercice in the next post.

Looking forward hearing from you, I give you down-below the bulk results, in graphical form and attached csv.

Cheers,





PS : the legend of the latest graphic has wrong names. _wing12 is to be read _rectangular and _wing25 is to be read _elliptical.

 

Thank you Alan for this amazing analysis. It looks like you put a good bit of work into it, and it is appreciated. Sorry for the delayed response, wasn't feeling good then got slammed with work. I'll try and give a good response this weekend.

 

Sorry Alan, cannot comment as I have not read it seriously, if each of your post is a PhD thesis, we must prepare some beers before to log in, so be indulgent.
Cheers
EK

 

Fact is regular sails, full battens, with high aspect ratio, together with a pair of backstay, and made out of very rigid material, are very efficient. So, questions posed by Dustman are quite interesting. Besides, if you check modern wing sails, mounted on sailing yachts ( not those mounted on drones ), you see that they have an higher aspect ratio than than the one of any regular sail clothes that could be put on this kind of boat. Composed in many parts, it appears that wing sails are to be quite optimized, in order to beat a efficient regular sail set. This is also what I find in this "thesis"...

 

I haven't forgotten about this. Things got really hectic, as is the nature of my work.

A couple things.

It seems an advantage with the wing sail is being able sail efficiently upwind. According to my research a proper wing can give you a good 5 to 10 degrees more to windward. Could make up for some of the compromises I will be forced to make due to budgetary constraints.

Quandary: Some articles I read made an argument for lower aspect ratios. Stability being a major component of that(you can carry more sail at higher wind speeds). I also got to thinking of the reynolds number the sails will be operating at related to chord. A greater chord length would seem to operate at higher reynolds numbers, so may be more effective at lower wind speeds?

This is not meant to be a racing catamaran, I am more concerned with never surpassing my max righting moment. My design goal is to never have a chance of capsize due to wind action even in a gust of double the mean wind speed. I think it will still perform sufficiently well at the proposed sail area. According to my observations, the primary reason catamarans capsize is due to excessive sail being carried in bad conditions.

Happy spring.

 

Hi Dustman,
family and professional ever go first, this is how it is.

Yes, what you've read is ture. Wings can deliver a driving force more windward than a conventionnal sail. The stability considerations about fractionnal rigs are also true. Although there is a trade-off to be verified for your project. The windward benefits from the wing configuration come with the boat being able to achieve sufficient speed in the sailing conditions for which she is designed. Using lower aspect-ratio wings will also produce an higher drag, in opposition of being able to sail close to the wind. On top of my previous study, you might want to make a VPP analysis to get an idea of this trade-off between stability and speed, perhaps fine-tuning the wings areas and aspect-ratios. From my experience, in your case, I would say that, with the given surfaces, the boat's weight has to be low, below 400lbs, with a minimal wetted surface area.

The Reynolds number characterize the ratio of inertial forces over viscous forces. The effectiveness of a given shape is not given directly by the Reynolds number, it represents only the ability of the flow to stay laminar in presence of a pressure gradient. Below a certain Re value, the flow is laminar. Above, it is turbulent. Re is proportionnal to the product of the flow speed to a characteristic lenght. If the flow speed is lowered, a greater characteristic lenght is to be given, to the shape, for the flow regime to stay the same. Also, if the flow speed is lowered, the lift produced by the shape is lowered. So, in order to deliver a sufficient lift, you also want your surface are (ie span or chord), to be increase. These are the two reasons why chord lenght may be increase, on sails at at lower wind speed.

Then, I agree that, with the intended wings configuration, It will be difficult for this boat to capsize. Let me admire a little bit the spring flowers popping in the garding before going quickly into the calculations I've made, and I will tell you the wind range.
Cheers,

 

Hi Alan,

Finally ready to respond to your phd thesis(as Erwan so eloquently put it).

I should start off by explaining more clearly the boats intended use and the rationale for the design.

This boat's purpose is to sail from Corpus Christi, TX to Miami, FL primarily via the intracoastal waterway with two modest open water crossings, then from Miami across the gulf stream to Bimini, Bahamas, then island hopping the entirety of the Bahamas. My longest crossing will be approximately 80 miles, while most other crossings will be in the 30 to 60 mile range. I'm hoping to have adequate performance to makes these crossings in daylight hours, avoiding overnight crossings if possible. Two of these crossings, the gulf stream and from Georgetown to Conception Island, concern me a bit given the size of the boat. Average wind speeds from March to June when I'll be sailing in the bahamas seem to be around 10kts, mostly varying from 5 to 15kts, and short periods of higher or very low winds. The winds are mostly from the east/southeast. I'll be sailing against the wind or upwind to some degree almost half the time, and making numerous course changes going in and out of anchorages, and exploring various small islands and estuaries.

Originally I was going to build a biplane cambered junk rig on free standing masts, for easy reefing, tacking, and minimal need to mess with sail trim, using only 2 lines per sail, a halyard and sheet. At the time I was looking into wing sails out of interest and realized that the cambered battens I was planning for the junk sail were already in the shape of a wing, why not just cover the outsides with sail cloth and make a soft wing sail; it would actually be simpler to build, and would maintain all the aforementioned attributes of the junk sail, but would be significantly more efficient. It also solved a lot of other more minor complications with the junk design, and gave me the ability to balance the sail better for lower sheet loads, as well as being able to balance the weight distribution of the battens better so they can be reefed or put up with minimal fuss. Anyway, in the span of few minutes I realized that wing sails would have a large number of advantages over a traditional sail or junk sail.

Here is my reasoning for a wider beam:
-More resistance to capsize when exposed to breaking waves on the beam.
-The ability to carry more sail without increasing capsize risk.
-The potential to use the differential thrust of the biplane layout to balance the helm; the wider beam giving a longer lever arm. I'm not actually sure how well this will work...
- More deck space.
-Lower interference drag between hulls.
-Greater comfort due to reduced roll angle changes from wave action, while underway or at anchor.
-Also, I will have dual electric motors and want to use differential thrust for steering, the wider beam will allow me to mount the motors farther apart, increasing their effectiveness for steering.

Honestly I don't see much of a downside for the wider beam except for structural considerations.

I will be helming from a central position, won't be using my weight as a counterbalance. My stability calculations thus far have taken this to be the case.

Are these calculations assuming a wind speed of 10 knots? By under-powered do you mean only that it can't attain it's righting moment, or that the boat will be slow? Does the "minimum surface" represent the amount of sail area it would take to equal the righting moment at the calculated wind speed? My calculations may be way off, but it seems I would be able to achieve 7kts in 10kts of wind on a beam reach with 128ft2 sail area. I have been considering increasing sail area since the sails will be so easy to reef, it would be nice to have better light air performance and greater area for downwind sailing. Speaking of which, do you know how to calculate downwind performance? I can find very little information. Most boats seem to do about half the wind speed dead downwind without pulling out the spinnaker.

My first thought is that a regular sail will rarely be trimmed optimally because of everything you have to fiddle with. With a wing sail all you have to do is sheet it in or out.

The 63012 seems to have too narrow of a range to be practical. It seems to me that something around 18% would smooth out the lift curve compared to 63012 and be more efficient at smaller wind angles than the 63025.

I asked on another sailing forum what the typical course deviations are for self steering systems. The responses seem to indicate a typical course deviation of about 5 degrees. To me this means that the wing sail would still maintain higher efficiency within these deviations. I would argue that constant trimming would only be necessary when trying to maximize performance. Again, my primary considerations for using the wing sail are not performance related. The enhanced windward ability will be a major benefit for it's application though.

According to the graphs I've looked at the l/d ratios improve drastically from an aspect ratio of 2:1 to 3:1, modestly from 3:1 to 4:1, and improvements are relatively small beyond 4:1. When compared against stability and structural considerations going beyond 4:1 aspect ratio doesn't seem prudent.

Thank you for analyzing the rectangular vs elliptical sails. As you noted the differences are minimal and doesn't seem worth the extra complexity of building the elliptical wing.

I'm curious why you have chosen to analyze the 63012-25 rather than 0012-25? Do they behave much differently?

Thanks again, Alan,

Dusty

 

Sorry for the headache, Dusty. It's hard to go through the infos I've given here. Thanks for sharing details about your project, and sorry for the crappy presentation. I've taken the naca63 as for an example, it was the first at take my hand on in three mouse click. Your idea is great. Having pictures of these batten could help precise things. Have you planned also your return trip or is it one way ?

 

You have given the most helpful and comprehensive response I've ever received on this or any other forum, though it did have a few quirks. I feel bad for not being able to converse about this on the same level as your presentation. My primary purpose in asking these questions was to confirm that the wing sail is a viable option for my purposes. Your analysis and responses have given me a lot of useful information and a greater understanding, and have confirmed that it is indeed viable, and what the potential drawbacks might be. I am in your debt, if you ever need help figuring out how to fix, build, or grow anything(electrical, plumbing, carpentry, drywall, tile, water harvesting, gardening, etc) I am at your disposal.

When work calms down in a few weeks I'll be drawing up some of the details of the overall design and posting them on this forum for scrutiny. By the fall I will have built a scaled down test version and will post pictures and some performance data.

The return trip would be in reverse order. If it seems safe enough I may continue on to the Dominican Republic, Puerto Rico and the Virgin Islands. I may sell the vessel upon return and use the experience gained to build a larger version for greater journeys.

Now to go roast in the Arizona sun fixing a roof.

Be well.