Wing Structure Configuration

Discussion in 'Aircraft Design / Aerodynamics / New Technology' started by dannicoson, Mar 22, 2003.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Mar 22, 2003 #1

    dannicoson

    dannicoson

    dannicoson

    Member

    Joined:
    Feb 27, 2003
    Messages:
    15
    Likes Received:
    0
    Location:
    Round Rock, TX
    Orion,

    I've cut and pasted your last post to get this string started

    Dan

    Dan, to answer your question, you're right, fixing an aft spar would give you a stiffer wing, the question however is whether that is actually what you want. The drawback to this approach can be that you will get stresses in the wing where you may not necessarily want them.

    The Glasair is an excellent example. Due to their continuous structural design (one piece wing), the aft spar is configured for taking some of the bending load. For most flight this is OK, although there is a stress concetration at the aft spar in the area of the wheel well cutout. It turns out that adding material to take care of the stress concetration actualy stiffens that part of the wing and actually makes the stress case even worse.

    At high angles of attack however, this stress concentration becomes the highest stressed part of the wing. If we examine the Part 23 requirements for aerobatic flight and composite structures, by this definition the Glasair wing should not be considered aerobatic. The FARs require 6 Gs limit and a safety factor of 2.0 for composite structures (not 1.5). As such, the ultimate load needs to be 12 Gs. The FEA analysis I did for this condition resulted in stresses well over the failure criteria for fiberglass. This was quite a while ago but if I recall correctly, the ultimate strength of the wing at high aoa was less than 9 Gs.

    For most flight this is OK however, when you add in the long term effects of fatigue loading, Glasairs that are continuously used for aerobatic flight may develop weaknesses down the line.

    For this reason and a couple of others, I prefer to pin the aft spar and let it carry only the local loads of the ailerons and the flaps.

    As far as a stiffer structure is concerned, how much stiffer do you want it? You can design the skin laminate in such a way that it makes up for the lack of a fixed aft spar, and develp the same characteritics without the added complexity of loading up the aft part of your structure. But of course that's up to you - you may have different requirements or reasons than I use in my selections. A stiffer wing is good, especially considering controll crispness and flutter - all I'm saying is that there are different ways to skin a cat.
     
  2. Mar 22, 2003 #2

    dannicoson

    dannicoson

    dannicoson

    Member

    Joined:
    Feb 27, 2003
    Messages:
    15
    Likes Received:
    0
    Location:
    Round Rock, TX
    Orion,

    I have designed around a two-spar wing, the forward spar located about 25% chord (the forward most position of Cp movement for this airfoil), the aft spar is located at 70% chord simply because this allowed room for Fowler flaps and this location starts to give some depth for the spar structure.

    I had always planned on carrying the bending moment through the fuselage, not just pinning as you suggest. I guess I thought this is "how it should be done". Structurally it can work, as you say skin the cat more than one way...

    You have me thinking now. By carrying bending moment into the fuselage to the other side I have to build significant structure into the rear spar (this equals weight & cost). Might I be able to save weight if I leave most of that structure out of the rear spar and force the load to go through the ribs and skins to the forward spar?

    As I see it, the forward spar in this design has to be able to carry ALL of the bending moment anyway because at Va the wing is stalled which also puts the Cp right on the FWD spar. So the FWD spar has to have the structure no mater what.

    At very high speed or with flaps down where the Cp moves well aft (45% chord), the rear spar would normally start to pick up bending moment if it is stiff enough. But if I keep the sparcaps to a minimum the bending moment will transfer through the skin and ribs to the forward spar.

    I assume I would only need enough spar cap material in each cap to carry that amount of bending moment developed between each rib(?)

    Should the AFT shear web be sized to carry it's full share of the load and transfer that to the pin? Or does a good portion of the shear transfer forward to the forward spar too? I wonder if it ends up being plenty beefy just to allow attachments of flaps and all other local loads on the rear of the wing?

    I don't have the luxury of FEA to check my work. That's why I was planning on static load testing.

    Your thoughts and comments?

    Thanks,
     
  3. Mar 23, 2003 #3

    orion

    orion

    orion

    Well-Known Member

    Joined:
    Mar 3, 2003
    Messages:
    5,800
    Likes Received:
    135
    Location:
    Western Washington
    Hi Dan;

    As you say, the Cp does shift aft with speed and low CL but, when you load up the wing through a higher aoa, the Cp moves forward. Therefore, a high wing load is for all practical purposes centered around the main spar anyway, whether at Va or not.

    I prefer to put the spar at the thickest part of the wing and so the actual position varies, depending on the actual airfoil. I also like to put it at a position where the top skin is locally close to parallel to the bottom skin however at times this is impossible. I will admit though, sometimes other factors drive the configuration so it's not always where I would ideally like to put it.

    The 70% position for the aft web is about ideal. It allows plenty of space for a good size flap and it provides good structural height. I don't tend to use Fowler flaps - I prefer single slotted flaps as they are easier (and cheaper) to build and install. Performance wise, the two are virtually identical. The only benefit to Fowler flaps is the slight amount of added area the wing gains as the flap moves aft on its track. However this benefit to the CL is only minimal.

    Fixing the aft spar has a couple of problems. The first we discussed in the previous post, that being creating a stiffer area where you may not need nor want it. The resulting stress distribution may be more difficult to deal with than you want.

    The second part has to do with the twisting behavior of the wing section. The wing twists about the "shear center", best thought of as a center of stiffness. Making the aft part of the wing stronger and stiffer will move the shear center aft. If it moves too far aft, it can contribute to adverse motion as a function of loading. The result of this can lead to flutter or structural divergence.

    Leaving the aft spar pinned rather than fixed will save you weight, will simplify the fuselage installation and will maintain a benign motion in the wing structure under load.

    Also, remember that when you deploy the flaps, you are not likely to ever get anywhere near the ultimate loading of the wing. It is an interesting loading condition and worth examining but we tend to look at it only as a subset load in the FEA model. We never physically test for it.

    Nine times out of ten, the skin material you have and any flange that you have on the web, will carry more than enough strength to take care of any localized loads of the flaps or ailerons. Yes, this is partially a function of the materials and the structural configuration but for the most part, general aviation applications hit "minimum gauges" and these secondary loads take care of themselves.

    The aft web needs to carry the localized spanwise shear loads, the rib shear loads and the localized loads of the control surface attachments. The load on the fuselage attachment pin is the maximum anticipated wing moment divided by the distance between the main spar and the rear web. The forward spar attachment carries the primary wing shear loads, which are a function of the wing load.

    You may want to rethink having your wing design checked by FEA as it may save you quite a bit of time and cost in the long run. As it is you have to build the test wing panel and the wing itself. If the test panel fails, you'll be building another one. Once you finish testing, it's probably not a good idea to use it (if it doesn't fail that is) anyway.

    By the way, as a bit of advertising for myself, I do FEA as part of my company's services. My company is at www.oriontechnologies.net.

    Bill
     
    Last edited: Mar 23, 2003
  4. Mar 23, 2003 #4

    dannicoson

    dannicoson

    dannicoson

    Member

    Joined:
    Feb 27, 2003
    Messages:
    15
    Likes Received:
    0
    Location:
    Round Rock, TX
    Orion,

    I appreciate you input on the wing structure. I've been in this design stage a long time and am always learning something! That is one of the benefits of these discussion forums.

    Any chance you will be at Sun-N-Fun? I'll be there Saturday and Sunday and would like to meet you if you have time.

    Thanks,

    Dan Nicoson
     
  5. Mar 23, 2003 #5

    orion

    orion

    orion

    Well-Known Member

    Joined:
    Mar 3, 2003
    Messages:
    5,800
    Likes Received:
    135
    Location:
    Western Washington
    Hi Dan;

    No, I will not be at Sun 'n Fun. I think the only air shows that I'll make this year are the Abbotsford Show in BC Canada and possibly the Reno Air Races in September. I may make an appearance at the EAA fly-in in Arlington here in Washington but other than that I don't think I'll be able to get away for the other major events this year.

    If you want to discuss you wing further, you can call me next week - I should be around the office for most of the week.

    Bill
     
  6. Mar 25, 2003 #6

    Holden

    Holden

    Holden

    Banned

    Joined:
    Jan 19, 2003
    Messages:
    1,319
    Likes Received:
    139
    Location:
    USA
    Dan,

    I agree with Orion. Perhaps another way to say it is this:

    The pinned struture can be considered a "determinant" structure, whereas the aft spar structure is non-determinant, and requires FEA or very good guesses and lots of simultaneous equations.

    If the spar is put somewhere around the 35% C, or at the middle of the CP range, the wing box will be stressed plus and minus in shear, and serve to take loads as well as provide a surface for the wing. The spar takes the vertical loads and the box takes the torque. The typical "guage" needed for a wing skin is usually enough to do the job of torque, and the structure will therefore be lighter than an aft spar structure that does not stress the skins and requires an aft spar structure.

    Use a pinned structure. If you do, you can then test a spar as a separate structure and verify it before it is put in a wing box. The engineering is far less complicated. Instead of several FEA runs, you can do one or two because you can get very close. This will save a lot of money. A FEA run will require some iterations and is not cheap.

    Remember, if you make a change to the aft spar structure, you will need to do another FEA model, whereas with a pinned structure you may not.

    By the way, Orion, what would you charge to run a FEA for a wing?


    Holden.
     
  7. Mar 25, 2003 #7

    orion

    orion

    orion

    Well-Known Member

    Joined:
    Mar 3, 2003
    Messages:
    5,800
    Likes Received:
    135
    Location:
    Western Washington
    Hi Holden;

    The cost of an FEA run depends on the complexity of the geometry, the materials and of course the goal of the analysis. Generally a design of a composite wing panel primary structure (this is the main wing panel not including a carrythrough structure) runs in the neighborhood of $4,500 to $5,500, although we have done more complex designs for higher speed aircraft that came in substantially higher.

    For aluminum wings the cost is usually a bit lower.

    If on the other hand you are looking just to verify your choice of materials and so the analysis would need to do only one run with no optimization, the cost would most likely be around $4,000.

    The expensive (or time consuming) parts of an FEA analysis are the modeling of the geometry and mesh for the FEA runs and the documentation. Making changes in the material makeup or gauges (for aluminum) is relatively simple.

    A comparison between a fixed aft spar and a pinned aft spar for instance would not really require two geometry models (although a couple of modifications would have to be made), it can be done in one model with different loads and boundary conditions. The cost for this type of comparison then is usually small compared to the major tasks of creating the model and the subsequent report.

    Optimizing each condition (pinned versus fixed aft spar) would require several runs and several material combinations so this would of course raise the eventual cost.

    The difficult part (and limitation) of an FEA model is in trying to account for material property variations as well as those introduced by the builder. I try to be conservative in all our work and use historical data as a doublecheck for the computer derived gauges (or layers).

    The material properties I tend to use come from several organizations including Boeing, Ron Jones Marine and Ron Jones Composites, HITCO Aerospace Structures, and a number of miscellaneous sources and my material suppliers, all of whom tend to do a prety good job of predicting properties of materials commonly used in the homebuilt industry.

    Bill
     
  8. Mar 26, 2003 #8

    Holden

    Holden

    Holden

    Banned

    Joined:
    Jan 19, 2003
    Messages:
    1,319
    Likes Received:
    139
    Location:
    USA
    Orion,

    My guess was $5k. Your estimates sound about right.

    I got a quote a few years ago from a composites guy who just wanted to do the entire airplane. His quote was $50,000. That was a standard configuration with standard ideas. I had to provide the loads.

    To do a new and advanced configuration with engineering help would cost a lot, perhaps $250,000-$500,000 or more. Inventions, and creativity is extra. Coming up with a new idea takes a lot of time, thought and emotion. Sometimes a guy can work for months and come up with little to show for it.

    It is no wonder most engineers do small incremental improvements and believe firmly that what has been done is the best, deferring to the collective judgment of others. Mix that in with a desire for cash flow and a business plan, and very little new technology comes into play all at once.

    It really ends up as a labor of love. Most designers do it for the pride of design, like some ancient urge to make a mark on the wall of the cave. They get locked into a design concept and end up being driven by emotion and pride, and not logic and brutal analysis. Ironically, few do anything new and uniquely useful, ending up with a rehash of the same old themes. I hope I'm not in that group, but some days I wonder.

    The F-117 was a airplane that broke the mold and set the military on a new path. I wonder if GA will have it's F-117. It could soon be a reality with virtual VFR when the right airframe is combined with modern avionics. I guess it is more about vision than engineering when it comes right down to it.

    Holden

    PS Thanks for the use of the soap box...back to work.
     
  9. Mar 27, 2003 #9

    orion

    orion

    orion

    Well-Known Member

    Joined:
    Mar 3, 2003
    Messages:
    5,800
    Likes Received:
    135
    Location:
    Western Washington
    $50,000 to analyze an airframe? That sounds a bit steep. On average, we figure $4k to about $8k per major structure or assembly, although simple parts and/or details run much less. Even if we use an average of $5,000 per model and assume that we need to closely examine about six parts or assemblies per an average airframe, that comes in at about $30,000, give or take a bit.

    Your are however close on the overall costs. If we do all the work through the flight of the prototype, a new, simple airplane will cost about $250,000 to develop, although this number depends on a number of variables.

    A while back I responded to an RFQ for an airplane in the class of the turbine Malibu. An anlysis of the program showed that the complete development of the airframe through first flight will run between $550,000 and $650,000, or about half the cost of purchasing a single certified production model. Unfortunately this fraction does not work toward the lower end of the spectrum but the bottom line is that developing a new airframe does not cost as much as most people think.

    As far as a new airframe is conceerrned, I too feel that we must at least rethink the "mold". Maybe not break it but make it work better for the end user.

    In my opinion though, the emphasis should be on cost and utility, not necessarily on the avionics and systems. Many instructors and professional pilots I know already complain that too many people are taught to fly the electronics more so than the airplane. When something goes wrong, they tend to reach for the radio rather than figuring out what's wrong and flying the plane.

    Putting more gizmos in the cockpit does not necessarily improve the safety of the product. There is also a fine line between comprehensive situational awareness and pilot overload. Yes, modernized avionics and systems can improve the airplane over the current "steam gauge" technology, but only as long as the pilot can get all the information at a glance. Fiddling with screens, controls and trying to decipher crowded informational displays will add nothing to the flight experience.

    Well, enough of my own soap box. Happy flying!

    Bill
     
  10. Apr 20, 2003 #10

    Flying Boat

    Flying Boat

    Flying Boat

    Active Member

    Joined:
    Apr 19, 2003
    Messages:
    31
    Likes Received:
    0
    Location:
    Surrey B.C. Canada
    Orion,

    I would just like a little clarification on pinning the rear spar. Is the idea to have the axis of the pin longitudinal so that the rear of the wing could flex up and down but still transfer the pitching moment, in shear, to the airframe and tail section? I understand the need to keep the most rigid part of the wing ahead of the cp to avoid flutter, and the advantage of having the rear of the wing able to off-load some excess pressure by washing out but, I was wondering what effect pinning would have on the strength of a 2 spar "box" design.

    I have been out of the loop since DBOW went off line... its good to read your stuff again :)
     
  11. Apr 20, 2003 #11

    orion

    orion

    orion

    Well-Known Member

    Joined:
    Mar 3, 2003
    Messages:
    5,800
    Likes Received:
    135
    Location:
    Western Washington
    Hi;

    You're correct, the pin axis should be horizontal and aligned with the longitudinal axis of the airplane (actually, ideally the pin will be perpendicular to the motion of the aft spar so it acts as a simple hinge).

    The bending strength of the wing is primarily a function of the main spar; the shear strength is primarily a function of the skin strength and rigidity (panel stability). The latter drives the rib spacing along with the gauge of the skin, although the bending load also contributes to the skin panel instability criteria.

    Also keep in mind that usually the skin, being the outermost fiber of the wing beam in bending, is also under load and is stressed in conjuction with the main spar.

    A two spar box, which is pinned at the aft end, will of course have the bending stresses concetrated at the fixed part of the box.

    Bill
     

Share This Page

arrow_white