Improving fuselage shape

[Toobuilder]

Funny how often the Tailwind bubbles to the top in these discussions.

 

[Riggerrob]

.

Without attribution John Thorp ... told me it was easier to control wing-fuselage separation by coordinating the widest portion of the fuselage at the wing trailing edge. But he went further, emphasizing his "poor man's area rule" - an integration of the various shapes to minimize the transverse movement of air about the fuselage as it flows from nose to tail. Arnold recognizes this as well but explains it differently than Thorp.

Finally, referencing the Thorp and Wittman designs, both employing "boxy" fuselages. If you understand that both designs were initiated in the early years of "legal" homebuilding, flat fuselage surfaces were easier for home "building" and, as John explained, the higher drag of a square flat plate vs. a circular flat plate can reduced if the "corners" coordinate adjacent pressure distribution curves so as to "shed" air equally, avoiding transverse flow.

FWIW !

Phil Bolger said pretty much the same thing with regards to (stitch-and-glue) boats made of plywood panels. As long as curvature is similar on both sides of the "corner" there will be little flow around the corner, ergo minimum drag.
IOW if the corner is in line with long streamlines, little air will try to "turn the corner" producing minimal drag.

 

[stanislavz]

Find original post Some design musings

 

[Vigilant1]

Phil Bolger said pretty much the same thing with regards to (stitch-and-glue) boats made of plywood panels. As long as curvature is similar on both sides of the "corner" there will be little flow around the corner, ergo minimum drag.
IOW if the corner is in line with long streamlines, little air will try to "turn the corner" producing minimal drag.

It would seem likely that the penalty for a flat or square cornered fuselage would be relatively higher at lower speeds.
1) Higher wing CL results in a greater amount of flow over the top and onto the top of the fuselage behind the wing, not aligned with the fuselage longitudinal axis. So, more drag on top and the turbulence behind the sharp edges underneath.
2) Higher AoA of the wing brings a higher deck angle for the fuselage. Again, more difference of alignment between the fuselage (and those corners) with the relative wind.

Aerodynamically, and at these conditions, a square fuselage might be better if the corners were at the top and bottom? If the wing junctions were at the side corners, the obtuse angles at the fuselage/ wing junction would presumably be a lot better than the to-be-avoided acute angles that sometimes crop up with high or low wing designs, and no need for additional fillets.
But, not great for shoulder room.

 

[J Galt]

Thank you for the info and reference. I hope to find that "Master Lines Manual" and pick it up but there's nothing available right now. Most of my knowledge is regarding the Spitfire, but occasionally the Mustang gets referenced in there. I have read a couple places that the Mustang produced net thrust via cooling, also that the wing did not support much laminar flow (in real life), points that Lednicer opposes. His CDmin is significantly higher for the Spit than other RAE papers and my own calculations would show, his numbers are higher for the Mustang than you see also so I don't know what to think on that.

Anyway, sorry for the slightly off topic but it does further the fuse shape discussion, thanks for the info.
Justin

Billski stated the practical approach and desirable results above.

That said, The guiding principles behind Mustang fuselage design were a.) steady velocity gradient from nose to wing via Projective Geometry which embodied conic development in the design layout, and b.) minimizing surface protrusions ahead of the wing/lifting line. Major reason the radiator/aftercooling duct intake was behind the lifting line on the P-51.

Projective Geometry as applied to airframe design, was extracted from the theorems of Pascal and Brianchon (ref NAA Master Lines Manual) and began at NAA in 1940 with the design of NA-73 (Mustang I/XP-51) under Schmued (design) and Weebe - Lines Group. If you Passed Descriptive Geometry and Engineering Drafting, find a copy of NAA Master Lines Manual, by Robert Kurt Weebe. It was passed to Douglas and Northrup and Lockheed during WWII and used at NAA al the way through the F-107 and X-15 and B-70.

Weebe was a little known Giant in airframe design principles.

Justin is correct on the fundamental design differences between the Spit and Mustang and Ledicer's papers on Drag comparisons is the best out there. In addition, all Spifires had a higher CDp for the fuselage.

I did (and still do) disagree slightly with his conclusions that the cooling system of the Mustang did not provided positive 'jet thrust' as his boundary conditions were an Assumed exit temp of 170 degrees in his CFD model. For a radiator temp of 240 degrees F and plenum exit reduced to minimum, NAA tests indicated 195 degrees - and a positive net thrust over the parasite and pressure drag of the 'system'. An extrapolation of Lednicer temp plot beyond 170 shows net Thrust at ~ 180 degrees. Additionally, Ed Horkey stated explicitly that slight net thrust over cooling drag was achieved.

 

[stanislavz]

Phil Bolger said pretty much the same thing with regards to (stitch-and-glue) boats made of plywood panels. As long as curvature is similar on both sides of the "corner" there will be little flow around the corner, ergo minimum drag.
IOW if the corner is in line with long streamlines, little air will try to "turn the corner" producing minimal drag.

Same was found here : Fuselage tail section Angle

And - all fast and boxy airplanes have 2d shaped windows, not curvy like all cub derivatives. And then you may get this airflow over the fuselage :

http://acversailles.free.fr/documentation/08~Documentation_Generale_M_Suire/Aerodynamique/Couche_limite/Flight_Measured_Aerodynamics_of_Wittman_s_Tailwind.pdf

 

[rslryl]

Lateral stability.

 

[Vigilant1]

Same was found here : Fuselage tail section Angle

And - all fast and boxy airplanes have 2d shaped windows, not curvy like all cub derivatives. And then you may get this airflow over the fuselage :

View attachment 105170
View attachment 105171

http://acversailles.free.fr/documentation/08~Documentation_Generale_M_Suire/Aerodynamique/Couche_limite/Flight_Measured_Aerodynamics_of_Wittman_s_Tailwind.pdf

I can't argue with those tufts. Looks darn good.

 

[stanislavz]

Phil Bolger said pretty much the same thing with regards to (stitch-and-glue) boats made of plywood panels. As long as curvature is similar on both sides of the "corner" there will be little flow around the corner, ergo minimum drag.
IOW if the corner is in line with long streamlines, little air will try to "turn the corner" producing minimal drag.

Same was found here : Fuselage tail section Angle

And - all fast and boxy airplanes have 2d shaped windows, not curvy like all cub derivatives.

1) Higher wing CL results in a greater amount of flow over the top and onto the top of the fuselage behind the wing, not aligned with the fuselage longitudinal axis. So, more drag on top and the turbulence behind the sharp edges underneath.
2) Higher AoA of the wing brings a higher deck angle for the fuselage. Again, more difference of alignment between the fuselage (and those corners) with the relative wind.

In real world - they are not isolated. Just my opinion, but if all taken to extremum - ie real life hauler with 40"x40" cross section for minimum accommodation of passengers - you still have square box at the wings. it may have rounded corners, and some convex - but it is still a box. And doing transition from circular to boxy, then again to circular - will make plug making for composite kind of funny.. Analysis of its cd / cl at all angles - close to impossible.

And my short investigation on fuselage shapes, shows, that you want that extra volume in tail. Good example Aeroprakt A22 vs A32 :


Same wing, same airfoil, just fatter aft. fuselage and some small improvements, make airplanes from 160 kmh cruise to 200 cruiser. I did read its creation story, guys make some homework in CFD, but later they did a lot of live tests with tuffs Problematic area was wing aft to fuselage and fuselage section after main whells. Similar could be seen in some czechs ul - from cockpit + boom tail like cora/yora, to fatter rear section.

And if you want to go extremum - Wren was able to fly on 8 hp only. Absolutely airfoil shaped fuselage. Landing gear in fuselage


Same , or even better for skyvan - clean nose, due to engines on wings..

 

[ragflyer]

Found it. It was on some forum, not pdf.


It is interesting to compare the performance of the Pipistrel virus SW to the wittman tailwind. The tailwind was the pinnacle of tube and fabric performance from yesteryears while the Virus SW is touted today as one of the most efficient 2 seater airplanes around. So how do these airplanes compare? Fortunately we have some good objective data for the tailwind from CAFE testing. The Virus SW data is from Pipistrel found at:

Pipistrel Aircraft Virus SW | Pipistrel
the CAFE tailwind data can be found at: http://cafefoundation.org/v2/pdf_cafe_apr/WittTail.pdf
Rotax 912 data can be found at: 912Sperf - Rotax Service - Your Automotive Authority

The tailwind data is at 1425 lbs Gross weight; the virus data is take at 1320 lbs (600 KG ); In an ideal world it would have been better to compare at the exact same gross weight. The difference is not much though and if anything goes against the tailwind

1. The flat plate drag area of both airplanes are very close at about 2.0 ft^2
2. The wetted area drag coefficient of tailwind is a little less than the virus sw given that the virus sw has less wetted area (despite 10^2 ft more in wing area) due to a pinched sailplane type fuselage versus the cruciform fuselage of the tailwind. My estimation is that the tailwind has about 15% less wetted area- tailwind 370 ft^2 versus about ~ 315 ft^2 for virus SW .
3. Given 1 the two airplanes are virtually identical when it comes to max speed at a given power
4. The virus has much lower span loading and higher aspect ratio than the tailwind; hence the virus has lower induced drag than the tailwind
5. Given 4 it is no surprise that the Virus has a higher best L/D than the tailwind (17 to 12.7); the Virus is about 25% more efficient at best glide speed for each.
6. The virus has better structural efficiency ( payload/grossweight fraction) than the tailwind- 0.39 versus 0.51; this is primarily due to the lighter rotax power plant in the virus SW
7. 5 and 6 means the the virus would exhibit better climb performance for a given HP
8. However in recreation flying the airplane is most often flown close to or at the cruise speed; at this speed the the effect of 4 (lower induced drag) is minimal and flat plate drag dominates
9. If the virus sw is flown at the design cruise of 165mph it consumes 18 l/hr per Pipistrel data; working from rotax specs for the 912 this results in about 85HP; working from the drag data in the CAFE testing at the same speed the tailwind will generate 165 lbs of drag at 165 mph requiring about 90 HP (assuming 80% prop efficiency)
10. In other words the Virus is at best just about 6% more efficient than the tailwind flying at 165 mph. At faster speeds than 165mph the difference will be even less.
11. Keep in mind the tailwind in the above comparison would have 562 lbs of payload versus 684 lbs for the Virus SW. However, this is virtually the difference in weight between a lycoming 360 and rotax 912.

In summary, the tailwind, if anything, is slightly cleaner aerodynamically than the virus SW. The SW compensates with lower wetted area resulting in identical flat plate drag area with the tailwind. The Virus SW biggest advantage comes from its higher AR wing (or lower span loading) and structural efficiency . This results in a 30% advantage at low speed and ~ 6% at cruise. And the low end is not a fair comparison as the tailwind was optimized for a higher speed. Put a higher AR wing on the tailwind and a 912 and you get virtually identical performance.

So much for 60 years of progress..... composites, laminar airfoils, FEA, CFD etc versus one man (albeit a very smart one) and his welding torch

Funny I read that and almost posted a like and then realized it sounded too familiar (particular the one man and welding torch) and realized I had written that post over 6 years ....
Some design musings

It is not a CAFE report but uses the CAFE report of Tailwind to compare with published (which tends to be optimistic) data of the Virus. Not perfect but tells you the biggest changes have been engines and looks. The aerodynamic (parasitic) and structural differences are minimal for our class of airplanes. In fact even if you compared the 707 with the 787 the biggest difference is engines (high bypass) and followed by the span efficiency and thinner wings for low mach effects.

 

[stanislavz]

Funny I read that and almost posted a like and then realized it sounded too familiar (particular the one man and welding torch) and realized I had written that post over 6 years ....
Some design musings

It is not a CAFE report but uses the CAFE report of Tailwind to compare with published (which tends to be optimistic) data of the Virus. Not perfect but tells you the biggest changes have been engines and looks. The aerodynamic (parasitic) and structural differences are minimal for our class of airplanes. In fact even if you compared the 707 with the 787 the biggest difference is engines (high bypass) and followed by the span efficiency and thinner wings for low mach effects.

I know. later found its source

Improving fuselage shape

Funny how often the Tailwind bubbles to the top in these discussions.

www.homebuiltairplanes.com

 

[mcrae0104]

Projective Geometry as applied to airframe design, was extracted from the theorems of Pascal and Brianchon (ref NAA Master Lines Manual) and began at NAA in 1940 with the design of NA-73 (Mustang I/XP-51) under Schmued (design) and Weebe - Lines Group. If you Passed Descriptive Geometry and Engineering Drafting, find a copy of NAA Master Lines Manual, by Robert Kurt Weebe.

That looks like a tough publication to find. Does it cover material not in Roy Liming's Practical Analytic Geometry?

 

[Speedboat100]

Supplementing the comments above, John Thorp explained fuselage, as an airfoil, could develop lift coefficients far above a wing, but at even greater drag coefficients -arguing against the approach.

Without attribution John Thorp, (prior to Arnold's articles), told me it was easier to control wing-fuselage separation by coordinating the widest portion of the fuselage at the wing trailing edge. But he went further, emphasizing his "poor man's area rule" - an integration of the various shapes to minimize the transverse movement of air about the fuselage as it flows from nose to tail. Arnold recognizes this as well but explains it differently than Thorp.

Finally, referencing the Thorp and Wittman designs, both employing "boxy" fuselages. If you understand that both designs were initiated in the early years of "legal" homebuilding, flat fuselage surfaces were easier for home "building" and, as John explained, the higher drag of a square flat plate vs. a circular flat plate can reduced if the "corners" coordinate adjacent pressure distribution curves so as to "shed" air equally, avoiding transverse flow.

FWIW !

Aren't they still today ?

 

[rv7charlie]

I can't argue with those tufts. Looks darn good.

You might want to read the paper & look a little closer. It's not immediately obvious, but flow over the top of the fuselage is quite disturbed on the left side. Though Raspet attributes cause to the windshield corner.

 

[stanislavz]

It's not immediately obvious, but flow over the top of the fuselage

It is only around windows frame. Bus isn't due to propeller flow ? Ultimate solution could be like this
:

 

[Lendo]

S, All good responses to your enquiry.

Most Texts books suggest lower than 90° between wing and Fuselage allows the pressures from the wing and Fuselage to interact and cause drag, I would assume over 90° is better than less than 90°.

Again Most texts books consider the wing lift is the same as if the fuselage wasn't there, however Sadraey suggests that only 45% of that Wing area Through the Fuselage is represented by the lift of the Fuselage. There is a whole rage of Fuselage shapes, so you be the judge.

The Vision has a Wasp-waist (taper) further aft of the Cockpit, but I think a nice taper after the Cockpit and Wing would look nicer. Roncz suggests, if this is done right, it may actually give some thrust, by creating a high pressure area - again you be the judge.

I have the Vision Plans and have discussed the Design with the designer Steve Rahm. The 6" radiused lower corners are for strength as sharper radiuses can cause stress risers, however I have seen much lower radiuses used to in designs, I assume to better line-up with the wing.

I would build the Vision in Carbon using a Carbon Spar (Now available through Scott ) or make your own using Jim Marske's Manual. I would also used Ribletts Airfoil and not taper the thickness Root to Tip. I would increase the Span slightly ( and Wing Area) and allow .5 AR from Root to Tip. all geared toward the LS stall of 45 Knots.

 

[rv7charlie]

S,
If you zoom in on the top-rear view, you can see flow separation starting 5 or 6 tufts forward of the wing trailing edge, and 2 or 3 on either side of the left wing root. Hard to tell in a 2D image, but given the angle of the tufts, they may be standing up off the fuselage. The one that's even with the flap is almost certainly standing up. Spiraling prop wash coming up over the left wing?

I think that image is at ~60 kts, so it's probably not that bad at cruise speeds.

 

[stanislavz]

S,
If you zoom in on the top-rear view, you can see flow separation starting 5 or 6 tufts forward of the wing trailing edge, and 2 or 3 on either side of the left wing root. Hard to tell in a 2D image, but given the angle of the tufts, they may be standing up off the fuselage. The one that's even with the flap is almost certainly standing up. Spiraling prop wash coming up over the left wing?

No idea. I was investigating this one many times. Have some others from virtual cfd sources.. Its stills looks better than many curved shapes, sad that on compsite you want as less area as possible. And flat panels.. But TW have battens on its sides, and one on top to make it more curved..

 

[Pops]

My Falconar F-12 had straight sides over the wings. Gullwing doors hinged at the middle of the top of the cockpit. There was a huge amount of lift over the top of the doors. Each door had a front and rear double over center latch at the bottom of the doors. One time I forgot to latch the rear latch over my left shoulder and took-off. Just after lift-off the rear of the door lifted up about 6" breaking the laminated door bow . I tried to hold it down as much as possible, and didn't move it down any. It had a center stick so I had my right hand on the stick, and setting sideways and shoving down on the bottom of the door. Had to take my hand off the stick to reach to the left to operate the throttle and back to the right for the stick, flap handle and the overhead center mounted trim wheel like on a Piper Cherokee. Runway to short to abort so had to go around the pattern and land. Airplane had an approach speed of 80 mph light and 85 mph heavy.
Made a believer out of me on fuselage lift.
Clean airframe-- lower the nose a little in cruise and get a large increase in speed. Easy to get to VNE if not paying attention to the airspeed.

 

[Sockmonkey]

Thanks for the vote of confidence.

There are folks, including the late Mike Arnold, who believe that perfect is actually a fuselage that goes slightly wider as you go aft through the wing. See the AR6. Not sure if I buy it, and the info out there on straight vs slightly expanding has not proven to be definitive yet. I do know that straight with vertical walls works pretty darned well in a bunch of airplanes. Your call.

Billski

They mean for it to start going wider once it passes the thickest part of the wing, yes?
My thinking is that they think doing that will have the same effect as trailing edge root fillets.
Putting something there to take up some of the space where you would otherwise get a bubble of low pressure causing drag.
You guys get what I'm saying? Dunno if I put it well.
YMMV