Testing wing loadings

[Deskpilot]

As my design progresses, albeit at a very slow rate, my mind has turned to the load testing of the wings. Looking at various postings and images I see that it is quite normal to carry out this most important of all tests with the wing assembly supported in the middle, with or without the fuselage connected. Top loading the wings for -ve 'g' and inverting the wing for +ve 'g' loadings, (correct me if I'm wrong). Now this is all well and good BUT. Shouldn't we be testing the wing attachments at the same time? By this I mean, shouldn't the fuselage be attached and held in suitable cradle but with no support under or over(inverted) the wing itself? This would then test the complete assembly rather than the wing alone. As I see it, it's no good testing the wing to say +6g only to find that the wing/fuselage connections can only handle +3g and the wing departs the fuselage(strut-less, high wing config) or -3g(low wing config) Some systems apparently only need to be tested in one direction due to the symmetrical design of the wing!!???

 

Am I right, or am I right? Your comments and or corrections please.

 

 

[Head in the clouds]

Yes, you're right Doug, best way is to test the complete assembly. Here's a link to a thread I posted about the first testing of the Jabiru, you can see the structure is supported inverted from the cabin floor so the wings, struts and all their attachments, and the cabin-top compression carry-through are being tested at the same time. HIH, Alan

 

 

[kgwilson]

I can't see how you could accurately test wing loadings without the wing being attached to the fuselage. If you set some sort of support up rather than the fuselage it would require identical load/stress points as the fuselage attachment. The fuselage design may be that the wing load is spread through a very strong subframe or across a larger part of the fuselage structure. You need to spread the load very evenly across the whole wing as well. That is my amateur view anyway.

 

 

[djpacro]

Depends how much of the structure you want to demonstrate by means other than testing. E.g. Truss fuselages are fairly amenable to analysis .......

 

Yes, a symmetrical structure reduces the number of load cases although the attachments and struts need consideration as well as fuselage if that is being tested at the same time - fuselages rarely symmetrical.

 

 

[facthunter]

Testing the wing by sandbagging it ignores the fact that most of the lift is obtained from the upper surface. In this test the lower surface takes a larger than required load and the upper one is not loaded at all. The spar (s) are loaded though, but the ribs are not proven. As for the attach/ spar mount fittings I understand that the calculations done correctly for the materials used are sufficient. How do you cope with the movement of the Cof Lift. Calculations? from data relating to the airfoil. Nev

 

 

[Yenn]

If the upper wing surface provides so much lift, why does the surface form a concave shape between the ribs aft of the spar? Forward of the spar is ply, so not concavity.

 

 

[Deskpilot]

If the upper wing surface provides so much lift, why does the surface form a concave shape between the ribs aft of the spar? Forward of the spar is ply, so not concavity.

Are you picturing the wing surface on the land or in flight? I would imaging it goes from concave to convex as lift increases.

 

 

[djpacro]

Basic stucture of the ribs are proven as they are beams so mainly interested in shear and bending near either end.

 

Movement of centre of lift: low angle of attack (max load factor at design dive speed) and high angle of attack (max load factor at design speed) must both be considered with their different chordwise load distributions. Negative load factors if required. Maximum torsional load? Aileron and/or flap deflected so consider which combination in conjunction with load factor/angle of attack.

 

A number of different loadings to be applied so to work out the one that may break the biggest bits and do that last. Of course, nothing may break as the test is only required to ultimate design load rather than to failure.

 

 

[Head in the clouds]

How do you cope with the movement of the Cof Lift. Calculations? from data relating to the airfoil. Nev

When the wings are sandbagged the airframe is set with the nose down (inverted - equivalent to nose-up in flight) by about 15* to simulate the stall angle and the torsional loads applied by the forward shift of the Cp.

 

If the upper wing surface provides so much lift, why does the surface form a concave shape between the ribs aft of the spar? Forward of the spar is ply, so not concavity.

I would imaging it goes from concave to convex as lift increases.

Nev is correct about the lift of the top surface, and therefore so are both of your comments, the fabric does become convex in flight. That is the main reason the Foxbat now has a metal skinned top surface rather than the fabric it had in the previous version, the fabric was changing the foil shape too much at high alpha.

 

Tubular spars and metal top skins are another thing altogether, the flex of the spars causing oilcanning and eventual failure of the top skins, but that's another matter.

 

The need to check wing stresses by loading the upper surface is not a normal practice but the fact that around twice the lift inflight is produced from the top surface, compared with the bottom surface, is a very good pointer to why you need to stitch the fabric to the ribs, and why stitching right through the wing from top surface to bottom surface is far better than simply stitching to the top chord of the rib.

 

 

[Deskpilot]

This is the sort of testing that perplexes me. Testing a Sonex wing.

 

 

 

[M61A1]

Bill Whitney covers this subject very well in his light aircraft design notes. If you can dig up a copy, or even get them from Bill, I think you will find them very helpful.

 

 

[cscotthendry]

WRT to testing the wing and attach points; I imagine the difficulties of loading a wing to +6G (is that 6 * MTOW per wing or 6 * MTOW /2?) attached to the fuselage. If you test one wing then the fuselage is going to be subjected to an asymetric 6G loading and therefore must withstand that amount of twisting to the fuse structure. Is that something designers cater for? Alternately if you put two wings on and each is loaded to +6G then the fuselage will be carrying the equivalent of +12G at the point(s) at which it is suspended. That would be a potential failure point that the designer probably never allowed for, nor needed to.

 

Edit: I just realised that each wing would only be loaded to +3 * MTOW, but the question about the twisting and suspension points on the fuselage still apply.

 

Just my uneducated opinion though.

 

 

[Head in the clouds]

.... I just realised that each wing would only be loaded to +3 * MTOW....

Actually it's not even as bad as that.

 

First you deduct the weight of the wings themselves because they carry themselves in the air at any dynamic (G) acceleration, so you can knock around 50kg off your MTOW. Then - if the fuel tanks are in the wings you can knock off the weight of fuel that you can carry, say 70kg/100lt, so you're already down to 480kg for a 600kg MTOW plane.

 

Now multiply by 6 for 6g = 2880kg, which is how much weight you must put on the wings in sandbags, but you load it in an elliptical pattern on the wings, to represent the approximately elliptical lift profile that your wing generates due to tip losses (plus any taper), so the CG of the applied weights is a fair bit inboard of the centre of each wing's span.

 

Note that the rectangular loading on the Sonex wing posted earlier by Deskpilot is incorrect, the air doesn't load a wing uniformly like that, however that test rig does test the wing-spar very adequately, which is the object presumably.

 

 

[facthunter]

It does test the spar, but not correctly. Early designs they sat the wingtips on trestles and supported the plane, thatw ay and added weight. The spar has to cope with bending and shear loads which vary all the way along the spar. The outboard section of the wing can be quite lightly built if you examine the exact situation applying and that includes the lesser aerodynamic loads as the air spills over the outer tip from the bottom and into the low pressure area on the top surface of the wing, generating the wing tip vortices obvious on highly loaded wings. Nev

 

 

[Deskpilot]

Bill Whitney covers this subject very well in his light aircraft design notes. If you can dig up a copy, or even get them from Bill, I think you will find them very helpful.

Got them and will watch the DVD's again.

 

 

[metalman]

It does test the spar, but not correctly. Early designs they sat the wingtips on trestles and supported the plane, thatw ay and added weight. The spar has to cope with bending and shear loads which vary all the way along the spar. The outboard section of the wing can be quite lightly built if you examine the exact situation applying and that includes the lesser aerodynamic loads as the air spills over the outer tip from the bottom and into the low pressure area on the top surface of the wing, generating the wing tip vortices obvious on highly loaded wings. Nev

The wings on my RV6 are very substantial at the root and steadily get lighter till they're just a open C section at the tip, I figure there's buggerall load at the tips

 

 

[facthunter]

Exactly.. Except for the "hangar rash" types who don't watch where they are shoving their "monster". Nev

 

 

[Kitplanes]

First you deduct the weight of the wings themselves because they carry themselves in the air at any dynamic (G) acceleration, so you can knock around 50kg off your MTOW. Then - if the fuel tanks are in the wings you can knock off the weight of fuel that you can carry, say 70kg/100lt, so you're already down to 480kg for a 600kg MTOW plane.Now multiply by 6 for 6g = 2880kg, which is how much weight you must put on the wings in sandbags, but you load it in an elliptical pattern on the wings, to represent the approximately elliptical lift profile that your wing generates due to tip losses (plus any taper), so the CG of the applied weights is a fair bit inboard of the centre of each wing's span.

To add to this, you can also deduct the lift created by the roof as well, if it's a high wing plane. Alpha is set at the manuevring speed, normally around 12-13degr when setting up the test rig. The weight is not distributed evenly when testing. Lift is normally eliptical as mentioned before and easily calculated by using schrenks method but the weight distribution differers when its a strut braced plane or a cantelever wing arrangement. With a strut braced plane the shear and moment normally peaks at the strut attachment and then tapers off to the tip and inboard to the cabin attachments. If the aircraft has a cabin then other factors such as compression of the roof and other trusses must also be factored in. This can easily be calculated using Eulers formula. Remember to use a design safety factor when doing the calcs. Add to this the ultimate load factor required for production of 1.5 normally and you should be well on your way. This will ensure that you have some"fat" build into your design. There is nothing worse than having a showpiece that breaks during testing.

The wing and fuselage testing is but a small part designing a safe aircraft. The best course of action will be to do your design calculations by refering to a certification standard like FAR 23 or ASTM F2245-10 for LSA or CS-VLA or the Australian equivelant. Most of these documents also have the formula's required for your design file which should include a aerodynamic , structural and performance analysis.

 

Enjoy your design and remember the old saying " If it looks right, it will fly right" Leave experimental designs to the experts like NASA, it's cheaper and safer....

 

 

[Guest Maj Millard]

Always an interesting subject,...but the correct procedures are well know and available in many publications. It is important to test the wings attached to the rest of the structure, as it all has a part to play in the air. The photo of the Sonex wing being loaded may be just to test the capability of that one component alone. I'm sure other complete tests were also done on that aircraft. Yes upside down with the fuselage supported and the whole win assy attached and rigged as ready for flight. Manoeuvring angle of attach, and yes load I reduced span wise from the root to the tip. Once loaded, leave for 24 hours, and attach weighted strings hanging from the ceiling to measure any droop or wing twist under load.

 

Every item is under test, struts, attach brackets, in-wing bracing, columns, struts, control cables,....the works. Look for the weakest link, cables that are too tight, brackets that are deformed, struts or tubes that are starting to bend. Then check your weighted string lines. A bit of 'droop' at the tips should be expected when loaded but not a lot. Them remove the weights gently, and again check strings lines. The structure should go back to it's original. If the wing is to be fabric covered, load test can be done with out the fabric attached to observe the internal structure, cables etc. if it all passes with flying (sic) colors then when fabric covered it'll be even stronger. Additional weight tests may be done when covered but Probabily won't be necessary . Obviously if the wings are stressed metal covered (ala Sonex) then the metal skins are very much a structural part of the wing as a whole. Wing root attach fittings, and wing strut attaches much be especially scrutinised . Depending what your structure is to be used for, the load figures applicable to the FAA 'utility' category are generally sufficient.....................Maj....

 

 

[Head in the clouds]

To add to this, you can also deduct the lift created by the roof as well, if it's a high wing plane. Alpha is set at the manuevring speed, normally around 12-13degr when setting up the test rig. The weight is not distributed evenly when testing. Lift is normally eliptical as mentioned before and easily calculated by using schrenks method but the weight distribution differers when its a strut braced plane or a cantelever wing arrangement. With a strut braced plane the shear and moment normally peaks at the strut attachment and then tapers off to the tip and inboard to the cabin attachments. If the aircraft has a cabin then other factors such as compression of the roof and other trusses must also be factored in. This can easily be calculated using Eulers formula. Remember to use a design safety factor when doing the calcs. Add to this the ultimate load factor required for production of 1.5 normally and you should be well on your way. This will ensure that you have some"fat" build into your design. There is nothing worse than having a showpiece that breaks during testing.The wing and fuselage testing is but a small part designing a safe aircraft. The best course of action will be to do your design calculations by refering to a certification standard like FAR 23 or ASTM F2245-10 for LSA or CS-VLA or the Australian equivelant. Most of these documents also have the formula's required for your design file which should include a aerodynamic , structural and performance analysis.

 

Enjoy your design and remember the old saying " If it looks right, it will fly right" Leave experimental designs to the experts like NASA, it's cheaper and safer....

I'd agree with all you say except the second and last sentences KP.

 

Alpha should be set at the angle at which the CP moves to its forward-most point, since the purpose is to test the anti-drag capability of the wing/fuse structure. That may be at manoeuvring alpha (which is normally about 8-9* for 1.3Vsi rather than 12-13* I would have thought) or anywhere between that and the just pre-stall alpha (It will be at a different angle for different aircraft, according to the characteristics of the airfoil section that is employed, the CP doesn't move significantly with a symmetrical section so in theory you might then be better loading the wing with the tail down to simulate the drag loads rather than the anti-drag loads).

 

As for leaving anything to the 'experts like NASA' it's the homebuilders that have almost exclusively provided the developments in aviation from the earliest days to the present, and I am sure will continue to do so. Without meaning any offence, it may well be the cheaper and safer way to go, and manufacturing a lookalike of other 'breeds' of aircraft may well be the commercially less risky way to proceed , but unless we actively encourage the home designer/builder to step away from the 'tried-and-tested' types we'll all continue to be flying Cub lookalikes for another half century, and whilst they may be very capable utility bush aircraft there's nothing exciting about them at all.

 

NASA are not experts at 'experimental designs' at all, in fact they haven't produced anything notable or of use to recreational flying ever (a big statement - it might stir up a few data miners to liven up the thread), anything of interest to us that has progressed through the halls of NASA has always begun in the homebuilding environment and been accepted by NASA for further research, funding and development.

 

Homebuilders who have pushed the thought envelope and then gone on to do something with their dreams have brought us the Kitplane for one thing (Jim Bede BD4), moldless composite construction (all of the Rutan offerings) the unusual (Ligetti Stratos for example - with more time and a breakaway from some locked-in problems it would probably have evolved into something useful - and NASA remarked that it had achieved something that they had been trying to succeed at for a long time), Synergy is another good example of something that NASA should have been working on for a long time, but haven't. And considering that electric propulsion with VTOL/transitional flight are the perceived ways of the future where is NASA's contribution to that? Fortunately Joby are making inroads. NASA are, as usual since their transformation from NACA, still sitting on their hands, forgetting their portfolio in Aeronautics and bemoaning the GFC for interrupting their Space Agency while others show them the way forward.

 

 

[metalman]

I like the quote

 

amateurs built the ark, experts built the Titanic!

 

We just have to look at the experimental scene with the advances in glass panels, even the big names like Garmin are getting in on the act with great top end gear at bottom end prices,

 

 

[cherk]

Got them and will watch the DVD's again.

Any-one know where copies of Bill Whitney's book(s) are available? He previously advertised in the "rag" ; no longer.......and I've lost his contact details .

 

 

[M61A1]

Any-one know where copies of Bill Whitney's book(s) are available? He previously advertised in the "rag" ; no longer.......and I've lost his contact details .

I have a copy, but I believe that they are copyrighted. He is in the whitepages under C. W. Whitney at Graceville in Brisbane.

[Kitplanes]

NASA was just an example, and a bad one at that, but I do agree with you. A lone amateur build the ark and a group of professionals build the titanic and look how that turned out:laugh:. My point is that a little bit of knowledge is more dangerous than no knowledge, so the homebuilder must be careful to just"modify" this and that whithout keeping basic aeronautical principals in consideration.