Rotating Wing Mast – theoretical discussion

[MalSmith]

Originally Posted by national

If you take the mast and the sail as seperate entities they both have a drag and lift force assosiated to them. Together they create a total drag a lift force which is the important factor for a design. When a mast is rotated the force on a mast can actually contribute significantly to the Total lift. Thi sis especially true when an elongated mast is used as it creates significant lift which can actually act to drive the mast forward rather than backwards.

With a rotated round section the force on the sail is similar to the elngated section, the main difference is the force on the mast. This change is significant.

I have spent so long thinking about this I have forgotten what I was writing about! But i think this makes things clearer!

Taken in isolation, it is true that an elongated mast will povide some lift if it is set at an angle if incidence to the airstream. A perfectly round mast will not provide any lift no matter what "angle of incidence" it is set at. But a sail attached to the mast affects the airflow around the mast due to upwash created upstream of the sail. The stagnation point is shifted towards the windward side of the mast and even a round mast will be providing some of the lift. In fact one of the problems of a circular section mast, as Tom Speer mentioned earlier, is that it's less than ideal leading edge geometry develops a sharp suction peak (local high lift) which leads to early flow separation. When using a round section, it should be over-rotated far enough so that the still attached flow is "peeled off" the mast by the sail before it can separate of it's own accord. I maintain that where the total sail area (mast/sail combination) is limited, the benefits of an elongated mast over an over-rotated round mast are not that significant. If axtra sail area can be gained under the rule by using an elongated mast, then it's the extra area that is beneficial over and above any aerodynamic considerations.

Mal.

[national]

Originally Posted by tspeer

It would be interesting to compare the XFOIL results with, say, tose from a Navier Stokes code. Do I smell a MS thesis topic here?

Tspeer - you seem to know a bit about this XFOIL, is it easy to use?
I also think it would make an interesting study to compare viscous CFD with a panel type, has anyon ever done it for a similar flow type?
It seems to be set up to run airfoils, is it easy enough to run sails and or masts?

[brian eiland]

I was searching for some info and photos of the Beirig's camberspar jib arrangement and this subject thread was referenced. However I did not find it. Anyone know of reference sites??

[Erwan]

It is a long time Tom Speer paper is a kind of benchmark for my research.
XFOIL is stressed to its limits and the Max lift and Min drag could be a bit different than in the real world. So the question to the experts is
How much reasonnable discount on the Lift Coef ?
How much reasonnable increase on the Drag coef ?

In 2D of course with about 1 000 000 reynolds

In fact I would like to reconciliate real world an theory: If you consider a A-Cat windward, a 4.5 meter/ second windspeed is enough to be full trapeze, the apparent windspeed is.... the righting moment is ..the average AoA is... and the IMPLIED Max Lift Coef hardly achieves 1.2.

Of course it is a 3 D result, induced drag accounts for a lot in the global drag.

Can we consider that from 2D to 3D induced drag could decrease max lift from 1.7 (theory) and 1.2 (A-Cat observation) ?

Or other factors must be considered as well ?

Thanks to the experts

EK

[tspeer]

Originally Posted by Erwan

...Can we consider that from 2D to 3D induced drag could decrease max lift from 1.7 (theory) and 1.2 (A-Cat observation) ?

Induced drag, per se, does not affect maximum lift. In principle, the angle of attack can be increased to compensate for the induced velocity ("downwash"), restoring the lift coefficient. The span loading for minimum induced drag will produce a uniform induced velocity across the span. So that just amounts to a bias in the local angle of attack compared to the reference angle of attack.

Or other factors must be considered as well ?

Where 3D effects really affect maximum lift is in the distribution of induced velocity across the span, along with camber shape and twist up and down the sail. If some portions of the sail reach their stall angle of attack before other sections, then the lift will not reach its full potential. If you stop sheeting in when the first section starts to separate, then other sections will be operating below their maximum lift. If you continue to sheet in, some portions will be stalled and others still operating below maximum lift. If you sheet in until the last section reaches its maximum lift, other sections will be deeply stalled.

The maximum lift will be a compromise where the loss of lift due to stalled sections starts to exceed the increase in lift of the un-stalled sections. But in any case, it will be lower than having all the sections performing at their maximum. The local section loading will depend on the section shape (zero lift angle of attack and maximum lift angle of attack), twist (which determines the local section incidence angle), chord length (planform shape) and induced velocity from all the other sections. Twist is probably the most useful parameter to vary to harmonize the span loading for maximum lift.

It may be that a lift coefficient of 1.2 isn't maximum lift for the A-cat either. One can really only operate at maximum lift if both hulls are in the water. If the windward hull is flying then the lift is limited by stability, not aerodynamic stall. CL=1.2 may be the best value for performance if the drag increase above that reduces performance more than the increased lift improves it.

If induced drag is the largest contributor to aerodynamic drag, then it's likely the lift coefficient for maximum aerodynamic lift/drag ratio occurs below maximum lift. The lift on the hull depends on the aerodynamic loading from the rig. So it makes sense to operate past the point of best aerodynamic L/D if the increased lift required from the hull improves the hydrodynamic L/D.

Where these break-even points are will depend on the windage, hull form drag, wave drag, and board-foil profile drag as well as the induced drag of the rig and hull.

[Erwan]

Tom,
Thank you very much to have taken time to provide me with this academic approach. My question was a bit stupid as I had the basic equation to translate from 2D to 3D using Aspect Ratio.
Of course, the implied 1.2 Lift Coefficient is an average Lift Coef, and therefore we could find higher 2D lift coef for this wing section.
Your remark about the break-even points is very relevant and prompt me to read again the Norwood book.
I feel so lucky to have this opportunity to communicate with you on these issues, that I cannot resist to ask you another question about your workpaper: "TearDrop Mast:
You warn that XFOIL is stressed to its limits,so even without experimental datas, how much would you discount XFOIL Lift Coef, and/or increase XFOIL Drag Coef, in order to get closer to real life ?

I am looking for this information because I would like to investigate the "theorical" interest to have a thick asymetric wing section, regardless of weight constraint, section shape adjustments, trimming issue ..., just theorical advantage in 2D between Teardrop mast rig and asymetric thick wing section.

I promise it will be the last question, I guess many cat sailors harass you about these issues.

Thanks in adavnce & best regards

EK

[tspeer]

Originally Posted by Erwan

...another question about your workpaper: "TearDrop Mast:
You warn that XFOIL is stressed to its limits,so even without experimental datas, how much would you discount XFOIL Lift Coef, and/or increase XFOIL Drag Coef, in order to get closer to real life ?...

Ah, that's the rub, isn't it? Without empirical data, or a more powerful CFD method (like a Navier Stokes code), there's no way to know.

All I can do is what the cartographers of old did with uncharted waters - place a label that says, "Here be dragons."

[Erwan]

Tom
Thank you for your this remark, in fact my question was quite similar to those of my friends who ask me where will be the Dow Jones 3 years from now.
(I was mutual fund manager)

[PI Design]

Hi National,

I've just read your paper in the RINA Journal of Maritime Engineering. I have a few questions if you are still around.

[petereng]

Does Xfoil include the mast section? Or is it only the sail surface?
Peter S

[tspeer]

Originally Posted by petereng

Does Xfoil include the mast section? Or is it only the sail surface

When I've used XFOIL, I include both the mast and the sail. I represent the sail as a two-sided surface that is approximately 0.4 % thick. That makes the mast & sail a single section with a kink on the windward side.

When doing inverse design, it's unlikely you'll be able to maintain the required symmetry of the mast. I use a spreadsheet to reflect the lee side of the mast to the windward side. So it's an iterative process - start with a wingmast & sail, design a new section with XFOIL, fix up the coordinates with a spreadsheet, and go back to XFOIL to analyze the result.

[petereng]

So Xfoil predicts the lee and windward side separation bubbles? If so... its a good tool to put mast shape and mast /sail interaction issues to rest?? Peter S

[tspeer]

Yes, XFOIL does predict the separation bubbles on both sides. XFOIL uses an inverse boundary layer code that handles modest amounts of separation.

However, what XFOIL cannot handle the flow through the gap between mast and sail. So if the sail is attached with the bolt rope in a groove, XFOIL probably does a pretty good job. If there's a track with slides or cars, the question still remains.

Another thing XFOIL cannot handle are large backwards facing steps, such as you find with a Bethwaite-style wingmast, ala the Tasar. I've tried fairing the step in with the idea that if the separation still occurs at the shoulder of the step, the results might not depend very much on the shape of what is submerged inside the separation bubble. But I wasn't very successful with that approach. I had to sooth things over too much just to get a solution.

[tspeer]

Originally Posted by petereng

So Xfoil predicts the lee and windward side separation bubbles? If so... its a good tool to put mast shape and mast /sail interaction issues to rest?? Peter S

Here is an example of a mast rotation study using XFOIL, and you can see the change in the separation bubbles on each side.

[C Jenkinson]

I thought I'd keep my question in this thread, though I'm not (specificallly) following on from where Tom Speer finished off. The issue is to do with the configuration of pivot points for a rotating wing mast. I've attached a very rough sketch below of a proposed design of wing mast and have put the mast pivot at 45mm from the nose, which is 30% of the 150mm chord. Is this an appropriate position? Secondly, since the boom must pivot, in the horizontal plane, relative to the mast, where should that pivot point be for the boom; should it be inline with the luff groove bolt rope where the sail tension acts on the mast (i.e. position B in sketch) or should it be set to the side nearer the mast pivot (position A - it just happens to be in the centre of and underneath the halyard sheave in the sketch). Basically, how is mast rotation angle relative to the sail otherwise controlled - does anyone have any diagrams I can look at? I promise, I can sketch better than this, but hopefully it will do for now.