Wing Box Design 1

[admiral007]

I am reviewing a design of a wing. This is a very large wing, and for manufacturing purposes, it has been split into several parts (leading edge, control surface and the wing box).

The designers also split up the wing box. My question has to do with how they split up the wing box. They split the box into three sections, but the breaks are parallel to the root of the wing (in between ribs) and not perpendicular to the root of the wing (in between spars).

Also, the current design connects the wing box sections (which are laminate) together using NAS bolts. There is no structure to provide a load path between the leading or trailing edges of each box section.

I've attached a simplified sketch. DESIGN 1 is what the design currently calls for (breaks parallel to the root). DESIGN 2 is what I would have thought would be a better way to break up the wing (breaks perpendicular to the root).

It's important to note that because of the size of this wing that there are cables attached to the tip of the wing to help stabilize it in flight.

Any thoughts on either of these designs? Which way is better to have the breaks and why?

I think that the current design (DESIGN 1) would see high bending moments that would cause the structure to fail.

Any thoughts or comments will be appreciated.

Thanks!

--Erik

 

[SWComposites]

what is "large"? C172 or B747?

why break up the wing at all?

Design1 looks like a very bad idea

Design2 looks like a typical metal multispar box with separate skin panels between the spars; if there are no spars at the fastener lines, then that also looks like a bad design.

 

[rb1957]

it would have been nice to know the scale of the "project".

design1 is more typical of airplane design, as wing planks are limited in length. the splices between the planks tend to be heavy, particularly the lower splice (why?). another real world reason for rib-wise splices is that the loads increase towards the wingroot, so the panel is machined from a thicker billet, and the breaks allow for natural changes in billet thickness (and hopefully machining time can be "optimised". hopefully you don't mean to use tension bolts to splice the joints together ... shear fasteners are much more efficient.

design2 is somewhat representative of an integrally machined panel designed wingskin. the panel cross-section will change significantly from root to tip. the splices have the advantage the they are not carrying wing bending loads and so can be much lighter than design1. but the splices (on a typical wing) would be much longer than design1 ... sort of "swings and round-abouts".

you're right about wing deflections ... this is something that takes careful design, analytical predictions, and final ly testing to verify.

i'd suggest getting hold of an airplane design book ... Raymer, Niu, Torenbeek, ...

 

[rb1957]

2nd thoughts ...

clearly both designs will have the same bending moments in them. design1 is complicated by the splice joints that need careful design. I'd expect to see multiple rows of fasteners, of course depending on the size of the plane and the stresses in the wing skin. personally, i'd model the wing as a continuous structure, and design the joints separately, knowing the loads that the wing wants to apply to the joint.

if this is a "large" wing, then you probably can't buy wing skin planks to go from root to tip.

is damage tolerance a design consideration ? design2 can be more damage tolerant than design1, but not necessarily. and there is nothing necessarily wrong with design1; both designs need to be designed correctly to achieve the design goals at something like minimum weight.

 

[admiral007]

I can't go into details about the program I'm working on, so I'm sorry to be vague, but this is not a true "wing" and it's aspect ratio is small. This wing needed to be broken up, from what I understand, to be easily shipped.

Why is DESIGN 1 bad, specifically? I know it's bad, and I've tried to tell the designers this, but they just brush it off and say that their FEM shows that it's okay to be built this way. Is there a quick hand calculation I can perform to show them why this is not a good design?

I agree, DESIGN 2 will have to have multiple spars, and the fasteners will be on those spars, but this is common design practice for a wing, correct?

I've been researching and looking through other drawings and design books for a while now, and can't find any wing built like DESIGN 1, so I'm having trouble proving that it is a bad idea.

 

[admiral007]

To give you an idea of the size, the span is about 6.5 meters and the chord is about 13.5 meters

 

[rb1957]

design1 is not inherently "bad" ... it has features that could make the design heavy, but ribwise splices are quite common for the two real world considerations (size of plank (max typically is 12', <4m) and thickness required (design2 would require all the planks to be the maximum thickness, and so have the maximum machining time).

multiple spars is generally a very heavy design, but more damage tolerant If it's designed to be. But multi-spar wings tend to be more common in military planes, which is where i guess you are.

bottom line ... neither approach is Wrong, both need to be designed carefully.

where are you in the organisation ? you sound a little like a project engineer (in a strongly functional organisation, ie the power is on the functional side of the matrix, and project engineers are the "fleas on the dog's back"). when you "the desingers have run an FEA" do you mean designers or stress specialists ? (in my experience, Never Ever trust a designer's FEA ... generally they don't have the experience or the tools to do the job properly). Have design decisions already been made ? Are you revisiting a previous decision? and hence getting resistance.

 

[rb1957]

2nd thoughts (again) ...

in your design, short span, long chord, chordwise splices Could be more efficient. draw out the splice lengths ... shorter will be lighter. is this a composite wing ? (which would throw out the discussion of machining time)

again, with your wing planform, running spanwise splices along %age chord lines will probably be impractical at the tips. but maybe one plank parallel to the front spar, one parallel to the rear spar, and a middle triangle ? but you could run 2, 3, maybe 4 spars at %age chord lines.

 

[admiral007]

@ rb1957: I'm an contract aero engineer trying to validate this design for use on our program. This design was created by one of our suppliers (a first of type for them). They designed it and created the FEM as validation. They have someone who has worked in FEMAP for several years, but I don't know if he is a stress analyst or not. Before I dig too deeply into validating their FEM, I want to make sure that this is a viable design. They haven't entirely convinced yet that this design is the way to go. So far as I can tell, they chose this design because it will be easier to ship. Granted, that's an important aspect to take into consideration in design, but it should't be the only one, which is what it sounds like. They refuse to even consider the other design.

 

[rb1957]

arrhhh "partners" ... the bane of our existance.

shipping is a rather odd justification ... the splices will either be line drilled at their facility and disassembled for transport or line drilled on final installation ... neither is stunningly good ! better to ship the wing in one piece ... depending on where it'll be stuffed (systems).

they'll say thay can meet the weight required, untill they have to say they can't (and try to blame you guys !).

Either design can work. The load transfer at the joints in design1 is critical, particularly on the lower surface splices.

If you've defined the loads, and they're applying them correctly, and meeting all their requirements, then how they do the detail design is pretty much up to them ... so long as they do it Properly !

good luck ... you're somewhat skrewed. If they won't consider an alternate design ('cause they've invested a considerable amount of time in this one) then all you can do is check that they applying the loads right, using the right allowables, and haven't overlooked anything (DT, fatigue, ... corrosion, ...)

 

[cpinz]

Hi all,

just curious.....which airplane has a wing with a span of 6.5 m and a chord of 13.5 m ?!!

It seems to be a "big" airplane in a chordwise direction but not in the spanwise one

 

[debodine]

The old F-104 Starfighter? )

 

[debodine]

That was silly, I was thinking in feet, not meters. Too American, I am!

 

[rb1957]

i figured it's a delta wing, which would probably make it non-commercial

 

[Aerodesign]

Admiral007,
Is it a surface skimmer - like the Russians make? Or the version to go to Mars?
Your client sounds like a nightmare - if you are being paid as a consultant to validate 'a design' then the approach you should take is to question the root assumtions and the entire concept to certification route.
A particular design is often a compromise dictated by project specific factors - transportability may be a high priority in this case, or manufacturability.
In validating a particular design you would need to consider not just structural aspects.
If your job is to check the design they have conceptualised then that is a different (and more managable) process. This process starts by checking the root assumptions of loading, adherance to design guidelines (company or industry based), design calcs, FEM work, compliance to regs, etc etc etc.
Unless the company you are working for are dumb they must realise that to optimise a design will require a number of iterations (design loops). Maybe they did that already? or thought they did that already!
Possibly the approach to take to prove to yourself the benefit of changing from structural Design A to B or C... and then present that to them. This could be achieved using some calcs or FEM work.

 

[admiral007]

@aerodesign
i can't go into any more detail about the specifics of the design. i can tell you that it does have to be the size it is, but so far as i can tell from our supplier, they didn't do any design iterations - they just chose one design and went with it. my job is to check their design and their analysis. right now, in my project, though, i can't afford to spend the time to go through and fix their design/analysis. the burden of providing a viable design lies on them. if they can't do it, my company should not accept their services any longer and be looking for another supplier who can.

 

[MikeHalloran]

It sure sounds like an ekranoplan.

What an odd supplier relationship.



Mike Halloran
Pembroke Pines, FL, USA

 

[admiral007]

@MikeHalloran
yes - it's a very odd supplier relationship. i'm new to the program, and i'm still a junior engineer, but i know this is a weird (bad) relationship.

 

[rb1957]

nah, it's the normal supplier relationship, at least normal for my experience in the field !

if they're meeting their weight target, and have applied the loads properly, and have considered everything (static and DT), then the design's acceptable. probably not optimal, but then nothing is (or at best it's optimal for right now, or better some time ago when the data collected for the trade studies was valid).

it's not your job to ensure they give you the best design. there are no doubt many things that are not in the design spec that would improve the wing (maybe in-service maintainability, design elongivity (ie gross weight increases)) but that's not valuable in today's world.

the real "fly in the ointment" (raisin in the bread pudding ?) is they'll insist they can meet the weight target and when they can't plead their way out of penalties. and if you make them change the design, well then Everything that happens from here on is Your fault.

and you probably don't have time to look into everything they've done either ! Just skim the surface ? the splices concern you the most, look into them; i'd look into the fatigue/DT. do the wing root reactions look reasonable ?? where's your boss in this ? does he think it's ok and you're "just" doing a sanity check ?? again, you probably don't have much "face time" so make the most of it.

remember "chicken little" and "the boy who cried wolf".

maybe it won't turn out as bad as it could ... rest assured you won't get any credit, but if things go "pear shaped", guess who's to blame ?

good luck, you'll need it

 

[gutboy17]

Is this company willing to fund testing for validation. There are plenty of resources that will test the product that can be used for validating the FEA and the other performed analysis.