Xflir5 is telling me as chord length reduces and the reynold number drops drag goes up.

[Urepedese]

I'm hoping an experienced hand at foil design can pitch in here and check my theory. I'm currently pioneering a small hydrofoil that is working and I'm refining the foils to get lower drag as I refine my knowledge.

I'm entirely self taught and started this early December.

The issue I'm having is the xflir5 charts are telling me I could optimise the drag by increasing the angle of attack at the tips. Convention seems to be to reduce it. Along with increased AoA at the tips I only need an aspect ratio of about 5 for the amount of lift I need if I do an elliptical wing and this way.

2 other options I see from the data.
I could switch to the standard SG foil in brown at the tips as another option. This could also use a mild increase in AoA at the tips.
Instead of elliptical I could do a mildly tapered foil with a AoA around 9° at the root and taper to zero at the tip for an AR of 6.1

The small chord length (root 6-8cm) and low speed means a low reynolds number. The area needed, somewhere between 130 and 190cm2 tried so far at angles of attack around 4-8°.

I'm 3D printing these foils. I've even tried a box foil in the mix with an AR of 18, and it did work okay. It just hasn't given me the best speed to date.

My current focus that returned me from practical testing to Xflir5 was laminar flow separation and at moderate AoA the SG foils seem pretty good until I drop past 50mm. Am I overthinking this or missing something?

 

[Doug Halsey]

Could you possibly back up and explain your design some more. Nobody is going to be able to answer your detailed questions without knowing much much more what you're doing.

 

[tropostudio]

This risks jumping ahead of Doug Halsey's advice to providing more info on your project: the reason for decreasing AOA at the wingtips (washout) is to reduce the likelihood of tips stalling before the root. If the tips stall first, roll stability decreases. This generally applies to straight or aft-swept wings. Forward swept wings tend toward root stall, so tips could have increased AOA in some cases. Are you making assumptions based on 2D foil data that may not hold up for a 3d foil? You can model and analyse a 3D wing in XFLR5.

This thread from a few years back may be helpful:

C

Thread 'low-mid speed paddle board hydrofoil fences, vortilons, or nothing'

May 22, 2019

high aspect hydrofoils for surfing has come up in another forum but I figured here might be a better place to source some answers. basically the topic was on the handling performance of high aspect surf foils and their tendency for the wing on the inside of a turn to stall sharply due span wise flow and the poor low speed characteristics of high aspect foils.

all my questions are based on my experience with my home made foil- 1100mm span, 240mm chord at centre, 50mm of anhedral at the tips and roughly 35mm thick at the centre. take off speed is around 3-4kts in flat water while stand up...

• container
• Replies: 7
• Forum: Hydrodynamics and Aerodynamics

• 

 

[Skip Johnson]

C

Thread 'low-mid speed paddle board hydrofoil fences, vortilons, or nothing'

May 22, 2019

high aspect hydrofoils for surfing has come up in another forum but I figured here might be a better place to source some answers. basically the topic was on the handling performance of high aspect surf foils and their tendency for the wing on the inside of a turn to stall sharply due span wise flow and the poor low speed characteristics of high aspect foils.

all my questions are based on my experience with my home made foil- 1100mm span, 240mm chord at centre, 50mm of anhedral at the tips and roughly 35mm thick at the centre. take off speed is around 3-4kts in flat water while stand up...

• container
• Replies: 7
• Forum: Hydrodynamics and Aerodynamics

• 

Thanks for opening the door to a new unrelated rabbit trail.

 

[tropostudio]

Thanks for opening the door to a new unrelated rabbit trail.

Quite possibly related. The thread involves wing sweep, aspect ratio, DIY hydrofoils with feedback from Doug Halsey, Tom Speer, and the creater of XFOIL(code base for XFLR5). Maybe it's useful to the OP, but we probably won't know unless we get more info on their project!

 

[Urepedese]

Sorry for the tardy response. I might have found the answer and it is simple as Xflir5 CD is not total drag, only parasitic drag.

It would I could state my problem more succinctly in that the low reynolds number for lower aspect ratios is being equally matched by an increase in parasitic drag. So the general theory of increase AR I was working towards doesn't hold true in all situations.

I just did the calculation manually as I'm not getting XFoil to work yet for a wing, it keeps giving me faults. But an increase in aspect from 2.2 AR to 5.2 gives a 44% reduction in reduced drag but a 43% increase is shown by XFlir in parasitic drag. It looks to be laminar flow separation happens excessively once the chord gets much over

I've searched the problem in multiple ways and although obscure there is online content outlining this scenario for low reynolds numbers. Drone wings are mentioned in one reddit thread and it hinted at a low aspect wing matching it with higher aspect designs, exactly the scenario I've discovered.

I'm sorry I've refrained from sharing too much, I'm the first in the world to be doing what I've done and it will give me a competitive advantage in the next race. After that the gloves will be off for someone else to develop it. I'll come back and post photos when I'm past that point, which will be mid October.

I can tell you that is a partial lift situation and I'm lifting about 10kg at around 10km/h. I've used different wing areas from 14,000mm2 to 20,000mm2. The foil will be used on both fresh and seawater. I didn't know the amount I'd be lifting at the outset since I'm pioneering into a new area, I had to find it by trial and error. My first attempt was way undersized. Some early foils just had too much drag, foils supposedly okay for hydrofoils.

There is no pitching or rolling for this application. Just straight lift.

It would seem I'm getting close. A very simple foil 200mm wide and 90mm chord is working. Xflir5 tells me it has a L/D ratio of 130 at 6°AoA. I need to do more testing, spend a calm day on a lake to compare it other foils I have. I'll probably print a simple wing at 70mm. I'll design some fences for it and see if they give an advantage. I think they might reduce drag by 10%.

It would seem decreasing AoA at the wingtips, a higher AR, or an elliptical wing (I've got one I'm using with a 70mm center chord) may not be the best solution to reduce drag. Has anyone had this problem before?

 

[Doug Halsey]

It sounds like you think that you can reduce the induced drag by reducing the chord without changing the span. This is incorrect.

If that is indeed what you think, then you are probably being confused by the mixture of dimensional and nondimensional terms in the equations.

The most common expression involving induced drag is Cdi = Cl**2/ (Pi * aspect ratio). This gives the induced drag coefficient of an elliptically loaded foil, in terms of the lift coefficient and the aspect ratio (all dimensionless).

If you increase the aspect ratio by reducing the chord, you are also reducing the foil's planform area which is used in defining the lift and drag coefficients.

After taking account of those changes and converting the equation to involve actual forces instead of force coefficients, the above equation becomes
Induced drag = lift**2 / (Pi * dynamic pressure * span**2).

My apologies if this is not your problem

 

[tropostudio]

@Urepedese, sounds like you are figuring out what you need to know pretty well. Ignore the rest if it is too basic:

Your thread title is "Xflir5 is telling me as chord length reduces and the reynold number drops drag goes up." You are in XLFR5 2D foil design module, so drag calculations is limited to parasitic drag (the sum of viscous (skin friction) drag and form drag. Interference drag (such as at a foil/strut junction) and induced drag are not considered. Sounds like you have figured that out.

Here's a screen snip from Wickipedia:

Note that it is for 'steady flight,' so lift is constant. Velocity and drag can vary. There is one point - one velocity - at which total drag is minimized for a constant lift.

If you decrease velocity in order to reduce parasitic drag, you have to increase AOA to maintain lift. From both the screen snip and the formula for Cdi in Doug's previous post, this will mean an increase in induced drag, and also in total drag. Conversely, an increase in velocity to reduce induced drag means a decrease in AOA to maintain the same lift. Parasitic drag increases, as does total drag. One optimum - no free lunch!

One way to think of it is that for a fixed lift, at a high AOA and low velocity, a lower mass flow of fluid is diverted 'more downward' whereas at a low AOA and high velocity, a higher mass flow of fluid is diverted 'less downward.' Overly simplistic, but these screenshots from JavaFoil may help. Note that AOA changes from 2 to 8 degrees. Reynolds number is constant.

BTW, Martin Hepperle's JavaFoil is a great applet for 2D foil analysis: https://www.mh-aerotools.de/airfoils/javafoil.htm. Easier to use in many ways than XFLR5, or now XFLOW5.

 

[Urepedese]

It sounds like you think that you can reduce the induced drag by reducing the chord without changing the span. This is incorrect.

If that is indeed what you think, then you are probably being confused by the mixture of dimensional and nondimensional terms in the equations.

The most common expression involving induced drag is Cdi = Cl**2/ (Pi * aspect ratio). This gives the induced drag coefficient of an elliptically loaded foil, in terms of the lift coefficient and the aspect ratio (all dimensionless).

If you increase the aspect ratio by reducing the chord, you are also reducing the foil's planform area which is used in defining the lift and drag coefficients.

After taking account of those changes and converting the equation to involve actual forces instead of force coefficients, the above equation becomes
Induced drag = lift**2 / (Pi * dynamic pressure * span**2).

My apologies if this is not your problem

Perhaps what cuts to the heart of my problem is this study that shows the Oswald efficiency factor of a low AR wing might only be 0.4 under the conditions I'm designing for, whereas an AR of 2 might get me above 0.7 https://m-selig.ae.illinois.edu/pubs/AnandaSukumarSelig-2015-AeroSciTech-LowReWings.pdf

I've been using this to work out my actual lift. Hydrofoil Lift/Speed Calculator - Calculator Academy https://calculator.academy/hydrofoil-lift-calculator/ It is the drag that is in question though, in relation to the AR. Xflir5 does appear to be reflecting what the study shows in regards to the increase in parasitic drag at low reynolds numbers.

 

[Urepedese]

Thanks Tropostudio for you comments and the new tool (and thanks to Doug for pitching in).

Only issue perhaps with that new tool is it might not catch the problem, it says it doesn't deal with laminar separation bubble. Emphasis on the "might" part of that statement as I try to get my head around it.

There is a comment in the study I've linked in my answer to Doug that is rather interesting:

"A jump in e0 with taper ratio is observed in Fig. 24(a) for wings with an aspect ratio of 2 from a taper ratio of 0.5 and 0.75 to a taper ratio of 1. The results give credence to the non-intuitive notion that wing tip vortices aid in reducing drag by limiting the dominant pressure drag of flat-plate wings since rectangular wings have higher strength wing tip vortices in comparison with tapered wings."

I like your talk of mass. One interesting comment I've picked up with all my reading is someone saying talk of tip vortex is over stated and that it is all about the mass of air being deflected/directed and that the drag is caused when it remixes into the airflow. There is a doubling effect at play with this induced drag, so the trick is to distribute that mass of air to minimize the doubling effect.

 

[tropostudio]

@Urepedese - thanks for the link to the paper by Selig, et al. That is one serious test rig and thorough write-up on low Re flat-plate wings. You are right: the relationships between Oswald efficiency factor, Re, and taper ratio are very interesting.

 

[tropostudio]

Can't speak to the veracity of the results, but I was able to get XFLOW5 to perform a 3D analyis using the VLM2 method at a fixed velocity with varying AOA. I'd think XFLR5 should do the same.

Foil used was NACA 8413, with 2D analyses well above and below your target range. Ncrit was set to 3 and upper and lower trips at 3.0% of chord. No specific justifcation for that, other than rememberingTom Speer and Daquiri mentioned 1<Ncrit<3 and forced tripping for foils in water with XFOIL.

Simple 3D Wing based on your parameters:

3D analysis settings:

Results. As mentioned, no claims for veracity. Some methods work with some polar types but not others. Some methods seemed to fail for any polar type:

Here's a link to the XFLR 'Point out of Flight Envelope' error message PDF:
https://flow5.tech/xflr5/docs/Point_Out_Of_Flight_Envelope.pdf
And a warning!:

 

[Urepedese]

Tropostudio I gave up on trying to make the software work in 3D. But I often explain to people my brain works in 3D and I can stitch the 2D together to get a picture.

What I did was spent a day on a lake in mild conditions to test 3 foils, then a run without the foils. The result seems to back up the viscous drag problem/laminar detachment for low reynolds numbers. Over a 1 mile/1.6km course that was half a mile each way to account for the light wind three foils were almost identical. AR's of 5.2, 7.5 and a box foil at 17:1 and within 4 seconds of each other for an average speed of around 9.6km/h. Then a final run without a foil was 10km/h. I've been sulking since then.

Searched again but this time I asked how to design a foil for low reynolds numbers and I turned up more studies mostly pertaining to small drones. The problem is certainly known.

So a brain wave today was to use the NACA system to do something close to what I have in mind for a steel foil (printed ones keep breaking with the fatigue). I came up with NACA 6505 and popped it into Xflir5 alongside the SG6043 I've been using, and the same one modified.

The result seems to solve the problem, on paper at least. It even gets better with lower NCrit numbers. I'm not sure what is realistic for NCrit in my application.

Where it gets interesting is the Z shaped profile in the chart of Xtr vs AoA. the NACA 6505 has a very "Z" shaped curve where if I understand the chart correctly it shows laminar detachment at the leading edge of the foil for an AoA of 5° or over. It would appear this could work very well in a narrow range of AoA. Am I interpreting this correctly?

 

[tropostudio]

I think you are interpreting the XFLR results correctly from the standpoint of corrlation between graphs. As to whether the results will correlate to 'real world conditions,' I can't offer an informed opinion.

The abrupt change in XTr top at ~4 degrees AOA correlates to the abrupt increase in drag at the same AOA. Not sure what your 'Cambered Plate 6505-1' foil section looks like compared to a 6505, but a thin foil will typically attain higher Cl/Cd than a a thicker foil of the same family, but have a narrower 'drag bucket.' That is evident in your graphs. Screenshot below is of an SG6043 and a NACA 6505 as drawn by Flow5:

Here are wind tunnel test results for NACA 63-006 and NACA 63-021 foil sections from Abbott and von Doenhoff's Theory of Wing Sections, first published in 1949:

The 6X-XXX series were laminar flow foils developed using analytic methods, and and exhibit the typically flat 'drag buckets' over their designed Cl range. Thicker section means higher section drag and lower Cl/Cd mx, but a wider (more forgiving) drag bucket. FWIW, I remember seeing the composite AR-5 sport plane that set a speed record for sub-661 lb aircraft in the 1990's, built by Mike Arnold with guidance from aerodynamicist Bruce Carmichael. Every imaginable measure to reduce surface rougness, surface waviness, and appendage and juncture drag was used. Always scrupulously clean. Tons of work. No magic bullets.

 

[Urepedese]

Nothing new under the sun aye? (Well very little, I do have my name to a couple of patents).

Your response has been useful. I was seeing the first inflection point at 5° as the delamination point. That might still be an okay result, although I believe I want detachment much earlier so it reattaches. Maybe the SG at a higher angle of attack? More induced drag but offset by a much earlier delamination and hopeful reattach. The unmodified SG at 8°?

So on the bottom left of the image I attached is shows which line is which foil and the SG6043 is in there, along with the same foil I tweaked to get more lift at the lower speed. the -8. All are for Reynolds 100k to match the speed and dimensions for a higher AR foil. The NCrit is shown there also and is just at 5 for the two SG foils.

The modified SG6043-8 looked great in the first instance, but clearly there is an increase in drag with a smaller chord shown in my field testing. The field testing was with the SG6043-8 and the angle of attack around 4-6°. Foils I've tried and rejected were the Eppler 817 and NACA 8412. I need less drag, the SG6043-8 definitely gave me that over those earlier tested foils.

FYI the craft is 7m long so the angle of attack isn't getting a lot of change in use. The cambered plate at 5° is probably viable. I don't think I can escape a narrow drag bucket at this reynolds number.

Edit: I've been mixing up ideas in my head so my thinking above probably isn't so clear. I probably still isn't. I'm confusing two ideas I see as the path forward to reduce drag. One is improving the boundary layer energy and laminar flow so it doesn't detach too early and cause drag. The other is the strategy of invoking earlier separation so the laminar bubble is small and the flow re-attaches, which I think is as turbulent flow. The latter idea not as efficient but better than the drag caused with flow detachment and staying detached. Xflir5 seems to be pointing to a low AoA with the cambered plate (no modifications to that btw)