In essence he says the advantages in induced drag reduction is eaten up completely or almost completely by increased wetted area and weight on most real world shorthaul intra-continent flights. As you never get to the flight levels where the induced drag is a real problem (FL above 390) your negatives of slower climb, higher weight, more wetted area brings the real world gains down to about 0.5% of fuel burn advantage. You further have negatives on landing with a frame that floats more. So your 737 and 320 winglets are questionable depending on your operations.
For mid, long and ultra long haul winglets make sense.
CM From , joined Dec 1969, posts, RR: Reply 1, posted (1 year 1 month 2 weeks 8 hours ago) and read 2102 times:
Thanks Ferpe! Interesting read, but I think in may ways Mr. Arvai is doing exactly what he accuses Aviation Partners and Airbus of doing... he is unrealistically skewing the numbers in order to strengthen his point.
The issue with vertical winglets is the twisting effect they have on the wing (This is why the blended-raked tip Boeing is using is more efficient than winglets - less twisting effect on the wing). In order to strengthen the wing against twisting, you have to add weight which acts against the aero benefit of the winglet. Boeing strengthened the 737NG wing box at around line 800, enabling all and-on 737 wings to accept winglets. Airbus is strengthening the A320 wing this year. ICAS published a very detailed paper on this issue in 2006 http://www.icas.org/ICAS_ARCHIVE_CD1998-2010/ICAS2006/PAPERS/581.PDF
Still, I think Mr Arvai is missing the point. To begin, he deliberately skews his data. For example, he says the 27,000' is the altitude at which winglet efficiency "kicks in" and says "many flights on the US east and west coast... never reach the altitude at which the efficiency of winglets kicks in". I am certain there are many flights which don't get above FL270 (I haven't been on one in dozens of US domestic flights this year), but I'd wager the majority of US domestic flights fly much higher than that, and that's the point. An airline does not need to gain efficiency from winglets on every flight, they need to gain efficiency from winglets on enough flights for them to provide a net positive business case - something Mr Arvai hints will only happen if an aircraft gets routed above FL390 or a speeds below .71 Mach. This is easily refuted by numerous academic studies on the topic.
The coup de grâce of Mr. Arvai's of his piece is this:
"So why is everyone installing winglets on their aircraft? Hype, and not fully analyzing the advantages and disadvantages associated with them.";
This is pure rubbish.
The EU's M-DAW research into winglets is highly regarded and has been widely cited in describing the advantage of modern winglets. A quick look at the summary material from M-DAW easily refutes Mr. Arvai's assertions. Mr. Arvai would want to read the report titled "Numerical Analysis of Transport Aircraft Using Different Wing Tip Devices", and "The M-DAW Project - Investigations in Novel Wing Tip Device Design" which can be found here http://www.etw.de/cms/uploads/pdf/JQ/AIAA2005_Mann.pdf
These publications will tell Mr. Arvai far more about winglets than he can learn from a Russian who decided not to use them.
ferpe From France, joined Nov 2010, 1785 posts, RR: 57 Reply 2, posted (1 year 1 month 2 weeks 4 hours ago) and read 2008 times:
Thanks CM for this very factual answer (thought I could trigger one of you in the know ).
I will now read all the referenced material . Should be nice with a more fact based discussion around this subject, I am playing around with parasitic and induced drag on all modern aircraft but to get things reasonably sorted I need to understand how to handle winglets / swept wingtips.
The thing that triggered me was the spec on 777-9X with it's low power/weight ratio (halfway explained by now), now I am trying to go further and dissect the difference in wing design philosophy between Boeing and Airbus. It will be a thread in due time but first I must understand winglets .
mandala499 From Indonesia, joined Aug 2001, 6181 posts, RR: 74 Reply 3, posted (1 year 1 month 2 weeks 2 hours ago) and read 1950 times:
The best way to check this is to get the FPPM of an NG with no winglet and one with... or the Performance Dispatch part of the FCOM for those. Unfortunately I don't have the correct data set (I got 900ER in kgs and 900 in lbs, so the rounding gets all screwed up!)...
But out of curiousity, I made some comparisons... 200nm, 400nm, 1000nm, and 2000nm trips at 60tons landing weight... for the NGs... (picked 60T because it was convenient on the table)
739 vs 739ER (prone to the rounding errors),
at 200nm and whatever the altitude the 900ER burns about 1.3% more.
at 400nm, the 900ER burns 2% less than the 900 at FL290, but diminishes as you go higher (zero or more at FL350)
at 1000nm, 900ER saves 1.5% at FL290 and diminishes to 0.5% at FL350.
at 2000nm, 900ER saves 1.75% at FL290 and FL350, goes down to 1.03% at FL330
Odd???? Well, perhaps the rounding is causing the problem...
But then, I decided to look at the holding fuel burn at 60tons at FL250 and FL350... and the savings for the winglets are about 3% at FL350, and 2.5% at FL250...
So, with the crappy and limited data I have, there's some truth to the scam... but the savings from the winglets are there... However, it's no brainer than for very short routes, the winglets will basically cost you more fuel burn... so I wonder what Mr. Arvai is fussing about???? 5% belongs to the 800nm and above regime, and because of the higher optimum altitude capability with the winglet... at the same altitudes, 3% is about right... but for long haul and talking optimum altitudes... 10% can be attained on the widebodies given the right circumstances of "the right wingtip aerodynamic gadget".
But we all know marketing is almost always too good to be true. Cut the marketing claims by 50% and we're probably at roughly the right number..
Now, if someone has the numbers that can really be used, it'll be appreciated!
Quoting CM (Reply 1): The coup de grâce of Mr. Arvai's of his piece is this:
Unfortunately, it sounds ridiculous, but I've talked to some airline management who seriously wanted winglets on his jets because he totally believes the hype and also mentions "because it looks cool" as well! So while I may disagree with most of what he wrote on that article, I gotta agree in a way with his so-called coup de grâce!!!
When losing situational awareness, pray Cumulus Granitus isn't nearby !
ferpe From France, joined Nov 2010, 1785 posts, RR: 57 Reply 5, posted (1 year 1 month 1 week 6 days 17 hours ago) and read 1805 times:
Thanks Manadala, I guess one must use the increase in flightlevel possible to get the true saving with winglets. If the winglet equipped aircraft can go to the higher FL earlier then the non equipped it must be included in the diff calculation.
To complement the arcticle lineup there are 2 very good articles from the 2005 Boeing Performance and Flight Operations Engineering Conference available at the SmartCockpit site:
"Blended Winglets" where SouthWest presents their operational evaluation of winglets on the 737-700
"Wintip devices" where Doug McLean from Boeing does a very good review of all devices and why they work and how. He also gives some good rules of thumb for their efficiency vs weight etc.
A very brief comparision is done in this Boeing article: Wingtip devices, Aero magazine no 30 which has been published here before.
Together with the CM link to M-DAW where the blended winglet is compared to the A wingfence one can now compare them all in terms of efficiency and their installations effects. I will conclude these rules of thumb and post them, they might be good to have for our discussions (and for my drag calculations ).
ferpe From France, joined Nov 2010, 1785 posts, RR: 57 Reply 6, posted (1 year 1 month 1 week 5 days 16 hours ago) and read 1629 times:
Quoting CM (Reply 1): The issue with vertical winglets is the twisting effect they have on the wing (This is why the blended-raked tip Boeing is using is more efficient than winglets - less twisting effect on the wing). In order to strengthen the wing against twisting, you have to add weight which acts against the aero benefit of the winglet. Boeing strengthened the 737NG wing box at around line 800, enabling all and-on 737 wings to accept winglets.
Reading through the documents this statement seems to be contradicted by the best B document, Wingtip devices by McLean (which I BTW find as one of the best as it does a very neutral and broad survey of wingtip devices and their trades):
"For retrofit applications such as the 737NG and BBJ, the required structural
modifications to the existing wing inboard of the tip device depend strongly on the
amount of structural margin available in the baseline wing. For the API winglet on
the BBJ, the baseline structural margin was such that only minimal changes were
required. Also, although winglets and raked tip extensions have roughly equal
impacts on bending loads for a given amount of drag reduction, winglets have less
impact on shear loads. Beefing up shear webs does not generally add much weight,
but it is expensive in retrofit applications, and this is a factor that will favor winglets
over raked tips in some cases."
Regarding the wing bending moments from the RWT resulting in similar structural impacts as the twisting effect of Blended Winglets; in my experience from working the designs on Boeing widebody wings, that's just not the case. At least not if you are designing a vertical winglet with equivalent aero (L/Dmax) benefit (examples of winglets with low structural impact and reduced aero benefit are 747-400 and A330). If a vertical winglet and RWT were equal in all aspects (L/D and weight), you would see Boeing opting for vertical winglets rather than RWTs on their widebody aircraft today, as vertical winglets preserve span and accommodate more gates.
As for the weight gain from adding winglets, there is persistent rumor in engineering circles that the wing modification for N636JB (the JetBlue sharklet demonstrator aircraft) resulted in 600kg added weight, including the winglets themselves. This is a bit more than the NG gained from strengthening the wing and adding the APB winglets. There is no doubt some of this was to address WBM, but I believe a good part of it was to stiffen the A320 wing against counterproductive wing twist.
Somewhere on A.net is someone who has worked wing/winglet design on single aisle aircraft who can more knowledgeably talk about twist verses bending moment than I can. Until then, my widebody experience says the RWT can achieve the same L/D improvement at lower total weight.
I think we have to differentiate between retrofit winglets and designed in winglets from start. The McLaen paper points out there is quite a difference, on an existing wing you can only work/change the spanload by working with the twist of the winglet, not the wing. IIRC you twist the retrofit winglet for less agressive air displacement = less backward force = less wing twist.
Fore a clean sheet design you work the twist of the wing and the winglet to get a better spanload profile and maximum overall efficiency, thus you load the parts harder. This is the situation you have been working on and there the RWT creates less twist which would be an essential element of it being lighter for the same L/D.
But only my speculation, lets see if some "let" wizard can come and save us .
ferpe From France, joined Nov 2010, 1785 posts, RR: 57 Reply 9, posted (1 year 1 month 1 week 5 days 2 hours ago) and read 1421 times:
So I have played around with the figures a bit, here my findings:
Caveat; these values are valid for a back of the envelope calculation of induced and parasitic drag according to http://adg.stanford.edu/aa241/AircraftDesign.html Kro and Shevell with some wisdom from Torenbeek as well . For Cls below 0.6 @ cruise Mach one does not have to consider compressibility drag, it is below 5% of total drag (and the inaccuracy of this method is way above that anyway )
To include the effects of wingtip devices over and above a normal wingtip one should ( Ferpes Home-areo cookbook )
(this is defined as running the wing around Cl 0.5 +- 0.1 which is the normal Cl area where one has min 1.2G to buffet for supercritical wings)
- For a raked wingtip (RWT) as for 767-400ER (10% of span length) use the non RTW span and add 85% of the RWT span in the effective aspect ratio calculation. Also add the extra wetted area of the RWT in the parasitic drag part. This leads to a 5.5% reduction of overall drag for the 767-400ER at FL350 and 73% of MTOW.
- For a normal Winglet as for 737 or 320 (14% of span length or 2.4m) use the normal wingtip span and add 45% of the Winglets span in the effective aspect ratio calculation. Also add the extra wetted area of the Winglet in the parasitic drag part. This leads to a 3.8% reduction of overall drag at FL350 for the 737-700W at 81% of MTOW.
- For the Wingfence of the 320 use the normal wingtip span and add 2% of the span in the effective aspect ratio calculation to cater for the fence effect. No extra wetted area needs to be added in the parasitic drag part. This leads to a 1.5% reduction of overall drag at FL350 for the A320-200 at 77% of MTOW.
At start and landing the RWT and Winglet are even more effective then above as the wingload and therefore Cl is higher (Cl of approx 1.3 and 1.8 respectively (more induced drag gets mitigated)). I have not tried to make any estimates of how to calculate these effects (yet ). As per the M-DAW the fence does not improve the low speed situation, is is then in full (delta wing) vortex flow and gives little help.
For climb to cruise altitude you are typically around Cl 0.5 and therefore the devices works as designed, I therefore do not understand Arvais argumentation re the climb .