Read the forum code of contact
By: - 24th May 2015 at 16:32 Permalink - Edited 1st January 1970 at 01:00
It has been established orthodoxy since the 1930's that one cannot design in a sub 90 degree angle between wing upper surface and adjoining fuselage side, as you immediately get an enormous drag penalty..............
I suspect you are over-stating the case. Adverse pressure gradients caused by the particular circumstances of a wing/body join can lead to a drag penalty of varying severity, but it does rather depend on the circumstances. The worst case might come with a circular fuselage and a thick-section wing set at a large incidence, while other combinations might not give rise to a detectable problem. The Whirlwind seems to have some "unusual" design features, and it may be that these simply happen to not give rise to a significant interference problem. There is no reason to suppose that those features could be usefully transferred to other designs.
By: - 24th May 2015 at 19:05 Permalink - Edited 1st January 1970 at 01:00
However, on the Whirlwind the rulebook appears to have been ignored by the young Petter.
Ignored by R.J.Mitchell too on the Supermarine S6 series. Oval section fuselage, low wing and no fairing.
By: - 24th May 2015 at 19:12 Permalink - Edited 1st January 1970 at 01:00
Ignored by R.J.Mitchell too on the Supermarine S6 series. Oval section fuselage, low wing and no fairing.
A 1928 design based on a 1926 design, so perhaps it would have been even faster with a suitable fairing
By: - 24th May 2015 at 20:25 Permalink - Edited 1st January 1970 at 01:00
Or perhaps 88 degs is so close to 90 degs that the drag increment is insignificant. Similarly the S6 fuselage may be oval but the local angle may have been close to 90 degs - I don't know the true angle. In the end, all designs are compromises. Making changes to the design for the benefit of a few degs interference may have been more penalising than the interference.
By: - 25th May 2015 at 07:39 Permalink - Edited 1st January 1970 at 01:00
I have read somewhere that this was supposedly one of the reasons for the cranked wing on the F4U Corsair, is it so?
By: - 25th May 2015 at 10:08 Permalink - Edited 1st January 1970 at 01:00
Thanks all. I think I have answered my own question, with the help of NACA report 540 from 1935 and a line from NASA report 7434469 from 1974. The first clue was the reduction in interference drag caused by reduction in wing incidence demonstrated by 540 and mentioned by HP111 above. Then, when rummaging around the NASA server, I found a report on fuselage camber related to US work on their SST. This mentioned, in passing, that a design that reduced the RELATIVE incidence between wing and fuselage over the trailing portion by changing fuselage camber) also reduced interference drag on this low-wing, tubular design.
It may be I was right to point to the localised widening of the fuselage here. When coupled with the angle off the vertical, the orientation of the plane thus created effectively changes the local camber and reduces the relative incidence of fuselage to wing trailing upper surface to near zero. Possibly not deliberate, but it seems to have worked. Any actual aerodynamicists out there want to comment?
By: - 25th May 2015 at 10:39 Permalink - Edited 1st January 1970 at 01:00
Well, I was an aerodynamicist, but long out of touch with such matters. But yes, that is why the Corsair (and Ju87 before it) had the inverted gull wing, and why the Hellcat and Hurricane had flat deep fuselage sides over the wing and a fairly sharp lower fuselage corner at the trailing edge. The latter more so on the Hurricane. I'm pretty sure that this localised widening of the Whirlwind's fuselage at the trailing edge will not have been accidental. It was designed at a time when doubt was been held about the value of the large fuselage fairings - e.g. DC 3, He 70, Spitfire - which took considerable effort to get right, with increased manufacturing cost and skin friction drag, if nothing else.
By: - 25th May 2015 at 12:24 Permalink - Edited 1st January 1970 at 01:00
Thanks Graham. I am just checking with Matt Painter's scanned data from the wind tunnel models of successive design stages to see whether the widening coincided with the introduction of the acute angle (it wasn't acute in the earliest phase we have data for, if anything slightly obtuse).
By: - 13th June 2015 at 13:41 Permalink - Edited 1st January 1970 at 01:00
So the next production design that I can find to have the locally flat section over the wing root, progressing back to a curved section aft of the trailing edge as a deliberate design feature, instead of any kind of fairing, was the F-86D.. a redesign of the earlier curved/faired F-86A fuselage. This puts the Whirlwind 10 years ahead of the thinking here. Interestingly, it might have been the 22 year old John Frost rather than Petter who did this. Certainly someone at Westland was keeping abreast of the absolute latest coming out of NACA in the US. The nacelles follow a theoretical shape described in a 1937 research note, while the report into the theoretical potential of locally flat sides (and straight junctions) rather than fairings on low wing designs appeared in early 38.
Does anyone have any more on this, especially John Frost - beyond the wiki article?
By: - 13th June 2015 at 16:57 Permalink - Edited 1st January 1970 at 01:00
If by "a locally flat section progressing back to a curved section aft of the trailing edge" you mean a 90 deg. angle between the wing and the fuselage, then you don't have to wait ten years. for you will find this on the Hellcat. Or, indeed, on the Ju88, and arguably on the Hurricane - although this had something of a sharp corner to the aft fuselage. What about the Defiant? I don't think we are short of other examples: Westland was not the only manufacturer receiving copies of NACA documents.
By: - 13th June 2015 at 20:19 Permalink - Edited 1st January 1970 at 01:00
No, I don't mean 90 degrees to the wing. Not at all. Look at the Whirlwind - now look at the Pilatus P-3. See the slightly awkward looking transition aft of the wing root that they have in common? This is a result of the designer aiming for as straight as possible a junction between wing and fuselage when viewed in plan.
This is done by locally making the fuselage flat sided, so there is no curve to interact with the aerofoil to make for a bent line. Hard to explain, but easy to understand if you look at a few examples. When done properly (on the WW the designer even corrected for the angle of the fuse side to the plane of the aerofoil by introducing the necessary fractional and imperceptible twist to keep that line straight) you don't need a fairing of any kind. Not like the Defiant et al. This is the theory put forward in the tech note. I'll dig out the links later.
By: - 13th June 2015 at 21:37 Permalink - Edited 1st January 1970 at 01:00
The older theory was that the angle was critical, so if you went less than 90 you hit interference problems. This was because low-wing, tubular fuselaged aircraft needed a fairing to ease the angle out that would otherwise exist. It was noted that the L-10 Electra didn't need a fairing, and a generation of designs ( Hurricane, Defiant, Hellcat) went along with the idea that this was because the flat sides avoided an acute angle, and the angle was the point. They all ended up needing various sizes and forms of fairing anyway. The 1938 NACA paper in question refined this slightly but crucially in terms of what I was referring to, placing the emphasis on a straight join in plan form arrived at, in this case, by local alteration of section profile to give flat sides
This subtlety was passed over generally until the very late forties. And consequently everything continued to need at least a little fairing, be it Mustang, Ta152 or even first-generation Sabre. However, if you look at the WW, there is NO fairing ( ignore the plastic kits), and a completely straight-in-plan longitudinal wing/fuselage junction aft of the mainspar.
That is what I am talking about, not the general shape of a fuselage or the angle to the wing. And interestingly it's in fact at an acute angle on the WW - 88 degrees - indicating that the designer knew it wasn't so much about that angle as the root form.
A modern low-wing airliner's wing root has growths and bulges that make that join straighter in plan. Very conventional now to straighten that line. Not in 1938.
Sorry I am probably not using the precise words an aerodynamicist such as yourself would. I hope you can stop laughing long enough to give me a chance, nevertheless. I have been looking at this in some depth.
By: - 13th June 2015 at 22:21 Permalink - Edited 1st January 1970 at 01:00
Ted Smith was big on reducing this interference drag, which I guess was a lesson he learned from Ed Heinemann while participating in the design of the A-20 and A-26. The Smith Aerostar--which we used to call the Air Roaster--was the ultimate example.
By: - 14th June 2015 at 10:49 Permalink - Edited 1st January 1970 at 01:00
But yes, that is why the Corsair (and Ju87 before it) had the inverted gull wing,
I think that there is a bit more to it than merely interference drag considerations. In both cases there was also a requirement for a high fuselage ground clearance .
In the case of the Corsair the comparatively large propeller diameter ( including an 18 inch ground clearance requirement) was problematical and was was solved by using an inverted gull wing, which also kept the undercarriage to a reasonable length. ( For example, Eric Brown refers to this in his Air International May 1979 article on the Corsair).
The Ju 87 had similar considerations, also associated with the ease of bomb loading . Examination of this aeroplane does not exactly indicate interference drag concerns as having being overriding !.
By: - 14th June 2015 at 12:02 Permalink - Edited 1st January 1970 at 01:00
Here's the NACA report: http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19930081386.pdf
Check out the photos of the wind tunnel model used, showing the junctions to be straight. The report concludes "The modification of a round fuselage to provide straight junctures appeared very effective aerodynamically. The resulting low-wing combibation possessed practically the same characteristics as the filleted condition"
This is not the same as the thinking that gave the contemporary generation of designs their shape, and their fillets. that was my point - hopefully you will see it now too, Graham?
By: - 15th June 2015 at 11:35 Permalink - Edited 1st January 1970 at 01:00
OK, I'm with you: you are meaning a straight line in plan view, not just in section. (Yes, you did say this in post 12.) And not just straight, but parallel sided. Not, I feel, a point made particularly strongly in the report when read in isolation from others of the period that (presumably) had already made the point for a 90 deg intersection in the front view. It may well have been clearer at the time. I would also query whether the extension of the flat parallel sides beyond the wing root was effecting this, something not discussed or even visible. However maintaining a constant (and rather wide) section is likely to be difficult on a small fighter where tapering to the tail has to begin early - or pay the penalty of extra drag elsewhere. Modern extruded tubes (Thank you Roy Brabrook) have this somewhat easier.
The difference between 88 and 90 degs is slight, and likely to be due to some design compromise rather than ignorance or a deliberate flouting.
Yes, the inverted gull wing was used for other purposes by the Ju87 and F4U - or indeed in Pulawski and Polikarpov's gull wing fighters. That doesn't diminish in any way the value of the 90 deg join, just points out that all aircraft designs have to meet a number of different requirements, and if these can be met in combination so much the better.
By: - 15th June 2015 at 12:07 Permalink - Edited 1st January 1970 at 01:00
:)
There was an earlier iteration of the WW design that showed 90 degree sides, so some other consideration does seem to have pushed that compromise. It might have been a reduction of the drag penalty you mentioned incurred transitioning from flat and parallel to curved and tapered. And that's an interesting point about the area behind the wing. Are there any publicly available primary source references apart from NACA out there on all of this?
By: - 15th June 2015 at 14:06 Permalink - Edited 1st January 1970 at 01:00
To answer my question - This is interesting (to some, perhaps); http://aaac.larc.nasa.gov/tsab/cfdlarc/aiaa-dpw/Workshop3/AIAA.2005-4730.copyrighted.pdf
Shows a fairing (FX-1) devised in 2006, according to the paper, to reduce the 'included' angle. This gave significant drag reduction.. but what that fairing is in fact doing is near-straightening the root (note the rear view, Fig.9). Not completely though. There was still a little bit of flow separation in the corner, so a fairing extending beyond the trailing edge (FX-3) then improved the situation a little further.. Talk about re-inventing the wheel.
Edit - actually a touch unfair - that fairing is a 'bulge' intended to keep to flow lines. But I wonder whether they would be divergent in the first place if that line was properly straight. But surely it can't be as simple as that.. can it? Graham?
By: - 16th June 2015 at 23:46 Permalink - Edited 1st January 1970 at 01:00
Maybe it can. I finally found someone who has spotted this parallel, straight sided thing 'in the modern era'. https://www.google.co.uk/url?sa=t&source=web&cd=17&ved=0CCoQFjAGOApqFQoTCPnRvcGilcYCFaFp2wodLUIAzw&url=http%3A%2F%2Fcontrails.free.fr%2Ftemp%2FInverse%2520pressure%2520gradient%2520matching.pdf&ei=HaSAVbnPD6HT7QathIH4DA&usg=AFQjCNED_9aZXigA8lTXXQ2erFY0DneXvA
Buried in this article is the same re-discovery.. the author's earliest example of this specific feature done apparently deliberately is the Bearcat (he says 'Grumman knew something'). He can't have noticed the Whirlwind over this side of the pond. Interestingly, it is known that Grumman did get a chance to study the one that went to the US.. but now that IS speculation!
Posts: 3,447
By: Beermat - 24th May 2015 at 12:22
It has been established orthodoxy since the 1930's that one cannot design in a sub 90 degree angle between wing upper surface and adjoining fuselage side, as you immediately get an enormous drag penalty as soon as you go beyond the right angle. Hence the large fairing on the Spitfire, where the low wing on an ovoid fuselage would otherwise create this situation.
However, on the Whirlwind the rulebook appears to have been ignored by the young Petter, where the angle aft of the mainspar appears a clear 88 degrees, and is totally un-faired. This should not have worked, and the RAE probably thought so too when they tested this specific area in a full-size wind-tunnel test.. and found the interference drag negligible (about 2 lbs).
So my 'question to the panel' is: what mitigating features might have been deliberately or accidentally introduced to perform this trick? Locally flat sides? An 'area rule' widening of the fuselage (about 2.5 inches, aft of mainspar)?
It certainly doesn't look right. It has taken some convincing on the part of our design guy Gunnar to make me accept that this design no-no was even perpetrated. But it does appear so.