Real world rectangular vs. Round fuselage cross section opinions

Discussion in 'Aircraft Design / Aerodynamics / New Technology' started by Jay Kempf, Apr 20, 2012.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Apr 20, 2012 #1

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,643
    Likes Received:
    916
    Location:
    Warren, VT USA
    A thread just popped up with flat vs. compound curved surfaces. For the purposes of focusing the discussion the specific configuration being high wing, tractor, conventional everything. The flight regime to inspect being only high speed cruise (not WOT).

    What is the penalty for having a completely rectangular cross section vs elliptical or highly rounded and specifically where does the increase in drag come from and how would you design around it.

    My opinion is if you have trim flaps, a well faired 2 and 3D loft that aligns with the direction of travel that you wouldn't have much if any penalty other than a slight increase in wetted area between two similar designs. Having a highly rounded structure and having to then transition to the wing junction area would create more trouble than it seems it is worth for some configurations.

    I realize that there are a lot of other compromises to go completely boxy like crosswind handling and off axis handling and maneuvering, but I am looking to discuss the on-axis condition.

    I have a reason for asking that has to do with prototyping methods.

    I have a feeling that the answer changes with V, but maybe not as it is a streamline thing in general, so maybe it varies with Rn or FPE...?
     
  2. Apr 20, 2012 #2

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    Well, there's the rub. Sure, we design to put the wing at a certain angle of incidence, putting the fuselage at precisely the right angle to the airflow to minimize drag at cruise, blah, blah, blah.

    And then the real-world airplane flies a degree off in cruise.

    It's just real-world tolerances. All of our design calculations, for all their depth, are just approximations. We don't really model the flow of every molecule of air around the vehicle with full Navier-Stokes solutions, so that we could know the forces and motions and angles of the aircraft in any situation as accurately as we modeled the air. Our stuff is just rougher than that. And then you have CG variations with different baggage weights, fuel loads, etc. The actual angle of incidence of a sportplane at a given airspeed changes during flight, in most cases, as you burn off fuel. So saying that you could somehow align the airflow well enough to get zero penalty isn't really possible in the real world.

    The "extra" drag of a sharp-cornered box comes from vortex-shedding at the corners, I believe. All the energy that it takes to spin up those little vortices of air at the sharp corners comes right out of your engine, ultimately. So unless you're perfectly accurate in estimating the cruise angle of attack of the fuselage and, even if you achieve that, perfectly ensure that the flow will always parallel the fuselage and will not be deflected by pesky things like wings and tails and propwashes, you're going to find two or all of those sharp corners misaligned to the flow, shedding vortices and sucking power that could otherwise be turned into speed.

    At really low speeds, like you said, the penalty is not as great. I saw a little single seater once that was make sharp-corner boxy. Thing was ugly as sin. But the pilot/owner flew to the air show while I got there in a car. It's all a tradeoff, just like the rest of this design nonsense. ;)
     
  3. Apr 20, 2012 #3

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,643
    Likes Received:
    916
    Location:
    Warren, VT USA
    Yup, ugly is a factor for those that care. And there is some small radius on the corner that keeps most vortices from forming at most small angles. Assuming the flat wrapping was going to be joined with some strip that provides that relatively small radius how does the analysis go for drag? The BD4 is pretty fast and not very draggy. Lots of data collected on that design. Very good numbers with that big prop blowing on all those sharp corners. At speed the flow straightens out some so the faster you go the less it matters me thinks as long as you are trimmed for that spot on the V curve.
     
  4. Apr 20, 2012 #4

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    It's all relative, which is why it's a tradeoff thing. Sure, the BD-4 took advantage of a sharp-edged, boxy fuselage for better ease-of-construction. Ease-of-construction was a higher "standard of goodness" for that design. The airplane would've been faster still with a beautiful rounded fuselage, but Jim Bede said "'X' is fast enough" and then used the "lost" performance to make the airplane far easier to build.

    Can you build a fast airplane with boxy corners? Sure. Either it'll have a larger engine than a same-speed "rounded" airplane, or it'll go slower with the same engine. "Slower" may still be perfectly fast enough, however.

    It's all a tradeoff. You puts your money on what's most important to you. Heck, look at the F-117. Boxy as all get-out. Because stealth was more important than absolute speed. They, like Bede, traded one thing for another.
     
  5. Apr 20, 2012 #5

    DaveK

    DaveK

    DaveK

    Well-Known Member

    Joined:
    Apr 22, 2007
    Messages:
    336
    Likes Received:
    14
    Location:
    Northern California
    As others have said its all a compromise. Think of the payload size and shape including the people. There is no point making big sweeping curves if it adds cross-sectional area. An example is the use of a circular cross-section for the fuselage for a single seater with relatively upright seating. Lots of extra surface area out to the sides that aren't needed for anything and as you said the wing junction can be a pain. Better to cut that extra volume out and maybe have more of a rectangular section with rounded corners. It fits the shape of the contents (you) pretty well and regarding structures, unless you are using construction like the BD4 or Tailwind there isn't much more work involved with rounding the corners for most sheet metal type aircraft. The simple curves improve stability of the structure too. With composites you could just as easily make a compound shape that really fits the contents with minimal wetted area as a less efficient shape so why not go with the efficient one?
     
  6. Apr 20, 2012 #6

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,643
    Likes Received:
    916
    Location:
    Warren, VT USA
    This is the crux. No you can't just as easily. It takes much more time to do compound curves, much more tooling... Cutting out flat panels is much faster, easier, can be self jigging, yadda. Things like compound canopies aren't easy. Joinery between non-flat things is not easy or just as fast.
     
  7. Apr 20, 2012 #7

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    So it sounds like, for you, ease of construction is worth the tradeoff for the ten or fifteen knots you might lose over an otherwise similar "curvy" ship at higher airspeeds. Nothing wrong with that at all.

    My point, to your original post, is that there really isn't any way to kill off the penalty completely. That's not realistic. It's simply a tradeoff you'll have to make, depending upon what's more important to you.
     
  8. Apr 20, 2012 #8

    Toobuilder

    Toobuilder

    Toobuilder

    Well-Known Member

    Joined:
    Jan 20, 2010
    Messages:
    4,384
    Likes Received:
    3,162
    Location:
    Mojave, Ca

    OK, you're never getting a ride in my airplane then! ;)
     
  9. Apr 20, 2012 #9

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,643
    Likes Received:
    916
    Location:
    Warren, VT USA
    That's not where I am going with this. What I am looking at is building a flying prototype of some design that could be drastically simplified to get it in the air. That would give you the ability to prove it out and work the kinks out of flight systems and collect a bunch of data while deciding to spend the money on the rest of the fancy curvy stuff. So, take the side and top view of a design and prototype the general design making every effort to avoid tooling and move the concept forward. That drives you to a test platform where you can cut and paste and modify and actually fly the thing, work the kinks out of your powerplant, instruments and controls, yadda... Then when you are satisfied you make a new fuselage shell and migrate all the expensive bits over to the new fancy shell. If you were faithful to hardpoints and made most of the installation modular such that you could move ALL parts over you wouldn't have sacrificed much in terms of redundant effort. That assumes that the flat wrapped scenario could be done as I am thinking. Make a canopy frame and flat wrap lexan. make a fuselage and bulkheads out of flat parts that can then be wrapped into final 2D lofted shapes... It's either that or a large scale RC model, also depending on size and complexity a redundant effort. I suppose you could fly a 20' model airplane mockup...1/1 before getting in the seat too. What I am looking at for construction it would dramatically shorten getting in the air. Understanding the sacrifice in drag would allow knowing if this mockup was adequately representing the final design. Not looking to kill off the penalty, looking to understand if it is predictable so that data taken on the mockup could be translated to a good guess on how the eventual final design would perform so you could make an informed decision on proceeding to the final design.
     
  10. Apr 20, 2012 #10

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    LOL! No, the little airplane I was talking about was ugly because of the nearly-unfaired VW conversion literally bolted to the nose (complete with old-style Vertex mag sticking straight up into the air), the constant-chord, totally unswept nature of all the flying surfaces, and the exposed bolt heads at every major structural attach. The full-bubble canopy looked like it was just stuck over a hole in the top of the pop-riveted fuselage as an afterthought. All the rivets were dome-head - none countersunk. Not to mention that the only paint on the entire airplane was the zinc chromate that was solely on the main gear spring legs. *scratches head* The thing was quite literally the "ugly duckling", and the swan part was that I'm sure the owner had a blast flying it around, while I was stuck on the ground.

    I'll take a ride in a HiperBipe any day! We need to get the SoCal HBA members together again at an airshow. It's been too long.
     
  11. Apr 20, 2012 #11

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    Oh. Ummmm... yeah. Quantifying that difference numerically somehow is going to be a bear, since you're talking about separated flow. Way out of my depth. You need to talk to Orion or one of the other more-knowledgable members about that.
     
  12. Apr 20, 2012 #12

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,643
    Likes Received:
    916
    Location:
    Warren, VT USA
    Maybe as simple as the boxy version wetted area, frontal areas and fineness ratios of the parts taken as fully turbulent vs. the same for the curvy areas taken as a percentage laminar estimated as a best case of performance increase. Those variables would give you a differing FPE and CD for the two designs taken at one point on the performance curve. Hmmmmm.
     
  13. Apr 20, 2012 #13

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    I don't think it's going to be at all that simple, but I suppose what you're describing would be better than nothing.
     
  14. Apr 20, 2012 #14

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    You know, I was thinking about this over lunch. I think you'd get a fair idea of the flying characteristics of of the "fancy" airplane this way, and certainly experience with the systems and some of the structural issues. Not so sure how you'd get much precision in flight performance prediction, though.

    If you're talking about rapid-prototyping one of your designs, say, with the split-inverted V-tail, I'm with you. One of the possible configurations for my motorglider shares that tail arrangement (the one I showed on the motorglider thread you started), and I have some concerns about the flight characteristics of that arrangement. No technical reasons why it shouldn't work well, but I'd sure feel better seeing a real-world test. I'm leaning on models, at least at first, but should I select that configuration for the final design, the urge to do an ultralight glider or something to test out the real-world flight characteristics will be very strong.
     
  15. Apr 20, 2012 #15

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,643
    Likes Received:
    916
    Location:
    Warren, VT USA
    That would be true if I didn't already have a spreadsheet with all that in it for the particular design I had in mind :)

    I have also been looking at how to collect real or at least better than impressions off of a large scale model. That requires telemetry, that requires probably an autopilot on board, that requires GPS onboard, that doesn't violate any FAA regs as long as it is within sight. I happen to have worked on a UAV project where I did all that stuff so pushing some data channels down to the ground is pretty straight forward. But I thought that collecting the data while flying might be a better approach. I suppose a black box sort of data collection component could be added to any model just to collect some parameters. After all that you have the fear of the model not scaling up in performance as predicted.

    For a full scale mockup I was thinking of all the things that you could actually do before spending money on CNC milled molds for a final config. You could work out cooling and cooling drag, you could work out flight controls and tweaking moments and forces, you could work out the prop and start just gear legs and then work into retracts. You could try out different wings, airfoils, yadda. A true test bed. You could start with some known engine to prove out the airframe concept and then migrate to the final choice...
     
  16. Apr 21, 2012 #16

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    I concur. Regardless of the amount of data sent down, nothing's going to give you the "feel" of flying the configuration without actually putting your butt in the seat and getting airborne. Nothing short of a full multi-axis simulator, that is, and I think perhaps that might be out of the scope for people like us. :gig:

    Yes. For something like that, I'd take your basic design and redo the structure for quick hand-layup, or some kind of flat-panel like you're talking about. The "cost" is having to do the structural design twice, of course. And you're essentially building two airplanes, even though you can gut the prototype for engine, instruments, some systems parts, wheels, etc. when you move to the final article.

    Me, the only major concern I have, if I end up selecting that configuration, is handling and "feel" in-flight. I picked up some balsa to make a simple "toss" model the other day, just as a confidence-builder. I'm not at the point of selecting the final configuration yet (I have three contenders), so at this point I don't want more from the little glider than just to watch it fly straight and true, not wandering or porpoising, as V-tails are reputed to do.
     
  17. Apr 21, 2012 #17

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,643
    Likes Received:
    916
    Location:
    Warren, VT USA
    V tails can wander but upside down ones are better than right side up ones. But I don't want to suggest that an anhedral tail might be stabilizing... :rolleyes:

    Adding some fixed vertical can always fix wandering and shallower than 90 degrees as a v tail should be can fix porpoising. Doesn't take much area, and nicely houses a tail wheel.
     
  18. Apr 21, 2012 #18

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,628
    Likes Received:
    5,279
    Location:
    Orange County, California
    I know, but, and please forgive me, it just doesn't suit my own sense of aesthetics. Dropping the lower portion vertically solves a multitude of sins, but I just don't care for the look as much. And I have nothing at all scientific or rational to base that opinion upon. It's purely an "eye of the beholder" thing.

    I've been sketching inverted-separated V-tail designs since the eighties. I love the look, and you and I both know the practical pros and cons. I'm willing to give up a little wetted area and make my tails larger if it means I can keep them as slab, un"bent" surfaces.
     
  19. Apr 21, 2012 #19

    karoliina.t.salminen

    karoliina.t.salminen

    karoliina.t.salminen

    Well-Known Member

    Joined:
    Oct 7, 2008
    Messages:
    407
    Likes Received:
    80
    Location:
    Finland
    Why dont you buy inexpensive CNC mill from Ebay (Chinese)? We purchased one which costed around 2000. Work area is 140 cm x 80 cm x 20 cm, good enough for cross section halves. With a little more expense you can get much bigger machine at price that is propably lower than getting your molds done 3rd party.

    The price value ratio is exceptional but you may find problems you may have to fix - one minor thing is that you have to expect there is not thread lock ( locktite) in any of the bolts. Open them up and apply yourself. It also helps to know about electrical engineering as not all Chinese electrical installations can be considered safe. With adjustments made by us, it is a fine machine that happily carves foam or MDF for me. It is great help even for small parts.
     
  20. Apr 21, 2012 #20

    topspeed100

    topspeed100

    topspeed100

    Banned

    Joined:
    May 4, 2009
    Messages:
    4,063
    Likes Received:
    63
    Location:
    Oulu/Finland

Share This Page

arrow_white