BL Transition Point - How to Record or Visualize It?

[green66]

Hi all,

Is there a practical way to record the location of boundary layer transition on a model's wing from actual flights? I've heard of oil+dye methods used on tunnel specimens but: a) would like to avoid making a mess on the plane, and b) think that this method may only work at higher Re.

I'm pondering: other oils, dyes, powders, thread/streamers, sensors, and on-board video to capture transition for those methods that don't provide a permanent 'imprint'.

I want to test performance of various turbulator locations; the analytical method is a start, but the real world is the bottom line, and knowing where transition actually occurs would cut through a lot of uncertainty and trial & error.

TIA

[rofujiyama]

Hi Green,
Awhile back,my friend and I were flying a glider with a schuman planform wing. It started to drizzle rain , one day and he noticed the rain had a slight airflow trace over the upper wing surface. He then got some WD-40 spray lube and applied a light mist on the wing panel upper surface. It was an electric, so he launched and did a short circuit around the field and landed. We noticed a definate flow pattern along each leading edge section, which ran approximately 90 degrees back from each section. Might be a while till you can get a day like that, but was a neat thing to see.
Then some , spray soap to clean off the oil, like gassers ,lol. Hope you can make use of his experiment.
Roland :)

[Ledbetter]

Short lengths of string in a wind tunnel.

[banktoturn]

green66,

I don't think that the oil/dye visualization is dependent on Re, as long
as the oil's viscosity is appropriate for the shear stress at the surface.
I think that capturing video or still images during the flight seems like
a promising idea. If you use oil on the surface, getting a good image
might be problematical, because of the lighting and the viewing angle
available to you. Maybe a strobe would help out with the lighting.

It would be interesting to apply some kind of oil to the leading edge and
look at how it spreads rearward during flight. This could be checked
after the flight, so you wouldn't have to generate good images in flight.
When it reaches the transition line, I would think that it's advance would
be affected. In theory, it would stop, but since the flow would be unsteady,
you might not be able to count on that. It might also make sense to put
spots of oil on the wing in a grid pattern. In theory, the ones just behind
the separation line should smear toward the leading edge. If this were to
work, you could gradually home in on the transition line by cleaning off the
wing and reapplying the spots nearer the likely location.

Some kind of streamers on the surface might be the easiest to get good
images of during flight. They could end up tripping transition, but that's a
possibility with oil as well. You could deal with that by placing the streamers
only near the transition line, and homing in, as described above.

You could also just apply temporary turbulators and measure airspeed to see
whether you're getting a drag reduction. They would have to be in the
correct range of thickness though.

Cool idea, good luck,

banktoturn

[green66]

Thx all, will consider all suggestions.

Here's some info that looks really applicable to what I'm trying to do; haven't read in detail yet, so not sure if workable at model scale:
Turbulators (http://www.standardcirrus.org/Turbulators.html)
Oil-Based Imaging (http://www.standardcirrus.org/OilFlows.html)
More detail on the green image (bottom of page) (http://www.ae.uiuc.edu/m-selig/uiuc_lsat.html)
Martin Hepperle's turbulator writeup (http://www.mh-aerotools.de/airfoils/turbulat.htm) This is the most relevant analytical info I've found for models, however no discussion on physically measuring the transition point.

[Batmanwpg]

I have flown a 2M glider for the last 2 seasons while experimenting with turb position. The turb positions predicted by computer sim (Profili-Xfoil) did not always prove correct but gave an idea of where to start for chord position turb placement. The airfoil was the HN1033 and turb positions tested where 50, 55 and 70%.
Keep in mind that the placement will always be a compromise and that not all airfoils will show benefit from using them! I selected as a final placement for best all around flight performance and handling characteristics. Things to watch for are slight increase in flight speed for normal cruise setting, reduced wing noise (very noticeable but this will depend on the type of wing construction material) moment co-efficient change and better stall control. I had flown the plane for a year before adding the turbs so I had a good feel for any changes. The turb thickness also needs to be sufficient. Doubling the tape thickness from .007 to .014 makes a big difference (.007 had little effect) for the turb positions I used. I have a friend that has tried up to .062".

[markdrela]

One way used in a wind tunnel which doesn't leave a mess is to dissolve naphtalene chunks (mothballs) in a solvent. I think trichloroethylene works, but I can't remember. Then spray a uniform layer of this on the wing using a high-quality spray gun. The solvent evaporates, leaving a whitish layer of napthalene. Then immediately turn on the wind, and watch the layer. It will evaporate first at the locations where the skin friction is largest, namely in the turbulent zone behind the transition line.

But I don't see how this is useful in determining turbulator position. The turbulator must be placed sufficiently far ahead of the intended transition point so that the unstable waves it generates are given enough time to develop and trigger transition. At low Reynolds numbers, the turbulator wants to be way forward. For example, F1A airfoils typically have turbulators at 5-10% chord, but the transition point is at 50% chord or farther. At higher Reynolds numbers, like on a full-size sailplane, the turbulator is placed just ahead of the intended transition line.

[markdrela]

Keep in mind that the placement will always be a compromise and that not all airfoils will show benefit from using them! I selected as a final placement for best all around flight performance and handling characteristics.
My view is that if the airfoil significantly benefits from a turbulator, then it is probably mismatched to the Reynolds number of that particular airplane. The only possible exception is F1A and F1B, which glide at a constant min-sink speed all the time after the launch. But in the vast majority of cases, especially if any sort of speed range is required, a turbulator just doesn't make sense. The problem is that the optimum transition location moves with speed and angle of attack (or with CL to be more precise), so one fixed turbulator will be a detriment at some parts of the operating range.

There is in fact an extremely effective turbulator device, which automatically changes height and chordwise position in response to AoA and speed changes -- it's called a weak separation bubble. And a weak bubble has essentially zero drag overhead, which is certainly not the case with a turbulator. So as long as the airfoil is designed to have weak separation bubbles in its intended operating Reynolds number, transition will occur close to the ideal location, and there's no need to mess with turbulators.

[markdrela]

I found these two postings on transition and turbulation from the bicycle-science news group, from about 10 years ago. The subject is streamlined-bike fairings, but it applies to airfoils as well.

[Batmanwpg]

<My view is that if the airfoil significantly benefits from a turbulator, then it is probably mismatched to the Reynolds number of that particular airplane. The only possible exception is F1A and F1B, which glide at a constant min-sink speed all the time after the launch. But in the vast majority of cases, especially if any sort of speed range is required, a turbulator just doesn't make sense. The problem is that the optimum transition location moves with speed and angle of attack (or with CL to be more precise), so one fixed turbulator will be a detriment at some parts of the operating range.>

True, but this might be become necessary when you buy a kit and want to improve its performance to bring it up to competitive standards of an Allegro Lite which I know was achieved! :D Side by side tests (and we did a lot under various conditions and time of day) with an Allegro shows a 2M with the turbulated HN1033 and similar wing loading to give nothing away to the Lite. All side by side flights were within + or – 10 seconds. The one thing that stands out is that my 2M is much faster in run mode which is a good benefit that I really like to have. So nothing was really lost by using the turb.

<There is in fact an extremely effective turbulator device, which automatically changes height and chordwise position in response to AoA and speed changes -- it's called a weak separation bubble. And a weak bubble has essentially zero drag overhead, which is certainly not the case with a turbulator. So as long as the airfoil is designed to have weak separation bubbles in its intended operating Reynolds number, transition will occur close to the ideal location, and there's no need to mess with turbulators.>

I have seen this phenomenon with Cp vs X graphs on Profili while testing the AG airfoils. So no turb needed on these. I think the AG airfoil series is great but it is not the only airfoil and sometimes a little creative work can reap great rewards.

[Ollie]

I think the AG airfoil series is great but it is not the only airfoil and sometimes a little creative work can reap great rewards."

Ha!

Great rewards? Maybe very small rewards.

[Batmanwpg]

Please explain yourself Ollie!

[Ollie]

I'm not sure. I'm etiher a bore or a boar. Your choice. ;)

[Batmanwpg]

I'll give you the fact that any performance increase might not be "Great" but any performance increase gained I'll take especially if the penalty is small or does not interfer with the envelope in which I fly my gliders. I feel the gain was worth it from 3 seasons of side by side test flying with an Allegro which impressed me the first time I saw it fly! I just decided to take another route to get to the same place. ;)

[green66]

But I don't see how this is useful in determining turbulator position. It's been my understanding that a trip should be placed just ahead of the transition point - "Just ahead" being 5-10% of chord, a range I guessumed as an uncertainty margin to stay ahead of the transition point. Obviously I wasn't aware of the need to allow distance for the trip waves to develop. I may have gotten this notion from writeups concerning full-scale planes, then blindly applied it to model scale.

My view is that if the airfoil significantly benefits from a turbulator, then it is probably mismatched to the Reynolds number of that particular airplane. Well that throws a demotivating wrench into my turbulation testing! I've had this notion that low-Re airfoils are, in general, especially amenable to turbulation, and that all but a handful of airfoils used on r/c gliders could be successfully turbulated..... if only one knew where transition was occurring.

So why all the hullabaloo and success stories about turbulating models having airfoils that I've thought were expressly developed for model-scale usage, e.g. E214, E387, etc, citing improved glide, tip stall resistance, etc?

Is the intent of tripping at low Re only to reduce drag at speed? The first plot below shows the E214 at Re=100k, plain and turbulated at 10 and 50% (the best of several chord placements). It looks like 50% turbulation would do a helluva good, but only at speed. This jibes with all my low-Re turbulation analyses, that no matter where the turbulator is placed, the resulting polars all show the same thing - that, again for low Re, I can't get an improvement near the upper knee of the polar. The second plot, however, of a 10% trip at various Re shows significantly improved L/D for Re > 250k.

So does this analytically confirm that, at sufficiently low Re, tripping simply cannot improve max L/D or min sink?

OK, now the real-world bottom-line - Let's say I want to test trips on my glider - no analyses, Re considerations, etc: Assume calm conditions, the plane is trimmed to fly dead straight, and a trip is applied to only one wing: If the trip provides a meaningful drag reduction, will it typically be enough to cause a noticeable roll to the non-tripped side? (and then, as a cross-check, move the trip to the other wing and re-test). That's the only simple method that I can conjure to truly prove the effectiveness of a turbulator; seems more sensitive / reliable than trying to measure glide slope changes.

Side question: What's with the jagginess of the non-turbulated curve? Is it just some computational instability associated with the separation bubble?.... and no effect on the interpretation of the curve?

[markdrela]

I've had this notion that low-Re airfoils are, in general, especially amenable to turbulation, and that all but a handful of airfoils used on r/c gliders could be successfully turbulated..... if only one knew where transition was occurring.
Xfoil will reliably predict the transition location on a clean airfoil, especially if there's a separation bubble. So it's not worth the effort trying to determine this via flow viz. A much more useful flow viz test would be to see if a turbulator causes transition too far forward.

Is the intent of tripping at low Re only to reduce drag at speed? The first plot below shows the E214 at Re=100k, plain and turbulated at 10 and 50% (the best of several chord placements). It looks like 50% turbulation would do a helluva good, but only at speed.
These plots are misleading. A smaller CL implies a higher speed and hence a higher Reynolds number. A Type-2 polar takes this into account.

Also, what you label "upper turb at ..." should really be "upper forced transition at..." .
The turbulator which achieves this transition location will have to be farther forward by some distance.

Finally, the E214 is much better suited for Re = 250K and higher. It's much too thick for 100K.

Side question: What's with the jagginess of the non-turbulated curve? Is it just some computational instability associated with the separation bubble?.... and no effect on the interpretation of the curve?
My guess is that the polars were computed with the original E214 coordinates, which have only 61 points which is very sparse. Re-paneling the airfoil with at least 160 panels will give smoother polars.

[green66]

These plots are misleading. A smaller CL implies a higher speed and hence a higher Reynolds number. A Type-2 polar takes this into account. In Profili (what the plots were made from), a "Type 2" polar simply considers several airfoils at the same Re - I take it that you're referring to something else. What is the Type 2 polar that you mean?.... and do you know if it can be created in Profili..... or a workaround based on Profili results?

[Ollie]

Your cart is before your horse. Try changing your point of view. Assume a context of weight, airspeed, wing platform, chords, wing loading and span to help you find the reynolds numbers for airfoils. Then when you find the airfoils' thickness' you can design wing strtucture. Then improve your assumed context of weight, airspeed, etc, etc. all over again till you are exhausted. It's lots of work but your not misleading yourself.

[Batmanwpg]

<Also, what you label "upper turb at ..." should really be "upper forced transition at..." .
The turbulator which achieves this transition location will have to be farther forward by some distance.>

Mark
"Upper forced transition at....." That was a great idea because now you don't have to consider the size of the turbulator. Just wish the program stated it that way. How do we determine how far forward the turb has to be up stream?
Also in this polar graph of the HN1033 with transition at 50, 55 and 70%, shows 70% to be good position which does not make sense? Actual testing I felt a turb at 50% was best and of course actual transition would occur at some point down stream from this, but 20% further down?

[green66]

Your cart is before your horse. Try changing your point of view. Assume a context of weight, airspeed, wing platform, chords, wing loading and span to help you find the reynolds numbers for airfoils. Then when you find the airfoils' thickness' you can design wing strtucture. Then improve your assumed context of weight, airspeed, etc, etc. all over again till you are exhausted. It's lots of work but your not misleading yourself. ???? Ollie - Please, take a break. Nothing in this thread suggests a design in progress. The thread is about turbulation. Try changing your point of view ;)

I only mention the E214 profile at Re=100k because it exists on one of my planes (Dodgson Windsong), which actually flies closer to Re=85k.

I am aware of the need to consider a complete picture when design involves multiple coupled variables. With that, the sequence of my design considerations typically goes:
- Intended application, flight modes, handling qualities, control config.
- Span, largely influenced by any fuse(s) I may have on hand
- Aspect ratio based on norms for span, considering Re, wing load, profile/induced drag, etc
- Area from span, AR
- Basic structure, materials

Iterate through the following:
- Wing loading
- Cl for min acceptable Re at about 3/4-span of the inner wing during a thermal turn
- Camber for desired Cl
- Thickness to meet strength reqmts
- Airfoil per desired camber, thickness, Re range
- Planform, airfoil transitions, washout, tip treatment for desired lift and Cl distribution, tip stall avoidance
- Tail areas for desired VCs
- EDA for desired spiral stability
- Stresses, lightening possibilities
- Adjust variables & re-calc, repeat until reasonably converged
* * * * * * * * *

[Ollie]

Good advise. I'll keep my yap shut.

[Batmanwpg]

Good advise. I'll keep my yap shut.


LOL :D

[histarter]

<Also, what you label "upper turb at ..." should really be "upper forced transition at..." .
The turbulator which achieves this transition location will have to be farther forward by some distance.>

Mark
"Upper forced transition at....." That was a great idea because now you don't have to consider the size of the turbulator. Just wish the program stated it that way. How do we determine how far forward the turb has to be up stream?
Also in this polar graph of the HN1033 with transition at 50, 55 and 70%, shows 70% to be good position which does not make sense? Actual testing I felt a turb at 50% was best and of course actual transition would occur at some point down stream from this, but 20% further down?

Sorry, I was in West Texas at the time this discussion was happening. My point of tripping (that statistical math validated) was for thermalling airspeeds primarily at lower altitude. I have learned to define those airspeeds as 13, 20, and 38 mph. This is contrary to the open ended philosophy that is normally applied, desiring airspeed definitions up to 200 mph! :eek:

I use 40 X 120 mil tripper as a one size fits all for enhancing 13 to 20 mph soaring - with a discovery of a bonus that it helps reduce sink all the way to 38 mph typically. ;)

Putting a 40 mil trip on the back of a wing for high alpha isn't noticed unless it is at the front of the profile demonstrating lower sink and airspeed due to increased alpha prior stall. Lowering alpha for more speed increases sink, but the 55% trip reduces intensity, and its effect seems to peak when CL is at 0.25 and velocity about 30 mph. Batman is working outside my universe and is doing a wonderful job relating the trip into the upper air for a larger world of soaring; however for universal application there would be no 'one size fits all' - meaning every AR, loading change, and pressure altitude would vary location and height of the trip somewhat to personally maximize its application to each pilot - that seeks a different nuance. :o :)

Every time I see a newbie duped into flying TD with an Oly while watching it fly with down trim for the breeze, and flounder downward for about a 3 minute flight that could have been a 5 minute flight (+ with ballast aboard and the trip) makes me quite sad. Putting a Trip on is not going to win contests; that takes skill! However, with the trip on, the less skilled will not lose so intensely for more encouragment - to go for the skill! :D