Static Fastener Strength Evaluations for Doublers 3

[joekm]

Lately, I've been playing around with trying to develop a computational method to solve arbitrary doubler installations. In the course of investigating this, I've found myself questioning a very common method that is typically used to verify static strength of the fasteners that attach the doubler. Specifically, you distribute the ultimate strength of the material removed among the fasteners on the appropriate half (or section - depending on repair geometry). Usually, you wind up with a pretty comfortable margin of safety.

This works because of the underlying assumption that plastic deformation is expected and this will drive the load distribution to be "more-or-less" equitable among the rivets. That's fine for finding compliance with ultimate loading, but limit loading is a different story. The requirement there is "no detrimental permanent deformation". I've done some playing around with this and it appears that this method does not show compliance with limit loading requirements.

So, I've been working on a way to modify the method to include showing compliance with limit load requirements. I'm pretty close to having a method at this point, but I'd like to ask a few questions before I do anything rash like notify the FAA.

1) Am I re-inventing the wheel here? Has someone else already discovered this and is there already a published method that I simply haven't seen yet?

2) Is there something I've missed in the underlying assumptions of the traditional method that would imply compliance with limit loading requirements?

3) Any thoughts, ideas?

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[jetmaker]

Joekm,

Could you elaborate on how it does not comply with limit loading requirements. I'm not too clear on exactly what "method" you are refering to.

Thanks,

jetmaker

 

[joekm]

A readily available example of the method I refer to could be found in the RAPIDC analysis methods document, section 2.0. They also do example calculations in the appendix.

located here:

http://aar400.tc.faa.gov/Programs/AgingAircraft/Commuter/RAPID/

This is the way I was taught to do it years ago and the way I've typically seen it done.

It's not exactly intuitive. The problem is that the underlying assumption that makes this method valid doesn't really apply when you are still in Hookian range.

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[joekm]

Just an update here...

I've been in informal contact with a researcher scientist at the W. J. Hughes tech center over this. Obviously, I went into more detail than I did here. I was just notified that they are sending my questions on to Boeing for further comment.




--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[royalra]

Joe:

Other than fatigue what purpose would be served by analysis of the doubler installation at limit load? Local deformation with loss of sealing or perhaps electrical grounding?

Russ Royal
Aeromec Inc.

 

[joekm]

royalra,

Mainly the requirements of 14 CFR 25.305(a), I.E. no detrimental pernament deformation at limit load. Although it's arguable that this need not be bi-axial - at least not to limit load in both directions.

Your question has merit, after all, to add a doubler to aircraft skin is to weaken it. The question remains though, "can a doubler install design be made to comply with 25.305(a)?" (and by extension similar regs for other aircraft types).

My initial calculations suggest that it is possible, but the limit loading condition becomes critical. However, I'm "working without a net". That is to say, I'd really like to do some actual testing to confirm that my approach is valid.

....meanwhile, still waiting to hear back....

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[crackman]

Hi Joe

Not really sure this helps but:

Hart-Smith of Douglas generated several codes for analyzing doublers and splices for both isotropic and anisotropic materials. It was a load balance method I believe rather than based on displacement. Anyways, his method took into account ramber osgood coefficients and allowed the user to input elastic and inelastic fastener flexibility coeffiecients to determine actual fastener load distribution. I have used it several times for both static and fatigue analyses although its been a while. The only real source of good inelastic fastener flexibilities is of course from Boeing (book 2 I believe?). The code is named A4EJ and has several version for various applications, it was funded by the USAF.

Anyways, good luck

James Burd

 

[joekm]

You might be on to something there crackman:

With a little intuitive thought, a load balance approach could bring the matrix size down quite a bit. My goal is to find something that is "field applicable", say, on the order of a Hardy-Cross implementation in complexity. I'll have to look into this.

<rant>
Sometimes, it's hard not to get annoyed at Boeing's tendancy to make every little piece of data they have proprietary. How is it that the USAF funds their development and they get to lock up the methodology? It's like the old adage: "Buy a man a fish,and he eats for a day. Teach a man to fish, then sue him for theft of intellectual property." The whole thing makes me want to write a work-alike code and release it under the GPL.
</rant>

Then again, the above rant sould be taken with the consideration that I haven't had my morning coffee yet.

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[Sparweb]

Joe,

Did you use something like Swift's paper as a starting point, or did you use first principles to create your matrix?

It's the first fastener load that kills your limit load analysis, (from a FAR 25.305 point of view) isn't it?


Steven Fahey, CET
"Simplicate, and add more lightness" - Bill Stout

 

[joekm]

Sparweb,

I did use Swift's paper as a starting point and, yup, first fastener load is the problem. Crackman has given me another way to ponder this but, he did mention Ramberg-Osgood, which is transitional yield behavior. Therefore, the issue of compliance with 25.305(a) is still up for grabs.

I'm playing with tweaks to Swift's equations but I'm not yet convinced that they are necessary. In any event, preliminary numbers suggest that sizing for yield stress behavior is possible.

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[Sparweb]

By making a doubler with a suitable limit stress behaviour, do you think that the fatigue behaviour will automatically fall in line? Or is that phase-2 of your project?

I've played around with Swift's equations, too, using MathCAD. I've made some results that match the graphs that Swift includes with his paper, but to do so requires some assumptions that Swift never stated.

The pitch and spacing of the rivet rows have a significant impact on the first-fastener-loads; far greater than the effect of a rivet size up or down. Have you observed this effect too?


Steven Fahey, CET
"Simplicate, and add more lightness" - Bill Stout

 

[joekm]

SparWeb:

First, fatigue is another project, this is just about the static analysis. Although, it was while research fatigue and damage tolerance methods that I ran across this issue.

As for your findings, yea, I get similar results.

For his example problem, I match his peak load pretty well (he gets 187.2 lbs and I get 187.1 lbs). Based on his figures, I'm in the ballpark of his other loads but it's hard to tell given the quality of the plots. There is also a typo on Figure 4. I'm pretty sure that should be rows of 8 rivets, not seven (kinda hard to explain that middle rivet otherwise).

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[jetmaker]

joekm,

Your initial question has me puzzled. Why are you interested in complying with CFR 25.305(a), I.E. no detrimental permanent deformation at limit load? Yielding of fasteners and doublers IMO do not consititute detrimental deformation. I can not see how the yielding of a doubler would preclude continued safe flight as long as seal integrity was maintained on a pressurized skin.

jetmaker

 

[joekm]

Jetmaker:

Compliance with 14 CFR 25.305 would be mandated for a doubler installation on a commercial aircraft.

As you your opinions; OK, but I'd still like to see some test data, or perhaps a supplimental analysis.

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[crackman]

Hi Joe

Just a little more info on A4EJ. In lieu of Boeing's data, one way is to just apply a very high flexibility at the given threshold (ie when limit is reached - if I recall correctly is set by the user by inputting a particular load). What the program then does is redistributes the load among the remaining fasteners to maintain a load balance. I will email you the report number (i have it at work). If things have not changed, I believe that you can request a copy of the report and the program from Wright Patterson thru an outfit called ASIAC (management of data is outsourced to this company although this was several years ago). I believe the only caviat is that you must be a US citizen. I think it is in the freedom of information area and therefore open to the public.

Just as a personal note, I did like the program however the input required is quite extensive. A graphical user input would be nice and I am sure there is a young engineer with either java or viz basic experience who could easily do it.

Will email you info tomorrow.

James Burd

 

[joekm]

That would be a big help crackman, thanks.

Yea, appliying a high flexibility would work, but the very method implies that yielding is expected. This suggest to me that the phenomena is well known and somebody, somewhere, has accounted for it either via testing or more detailed analysis.

Now that I've seen it, I'm obligated to understand it lest I find myself playing a game of "The Emperor's New Clothes".

BTW - I don't do either java or VB, but I do play with Python and have started looking at MS Visual C++. If I can bang something together, I'll certainly send it to you.

Incidentally, Python is highly recommended as an introduction to object oriented programming. It's a very clean, elegant language and is growing in popularity in the scientific community with several decent mathematical/scientific libraries available. For Windows, the best distribution is:

http://www.activestate.com/Python.plex?_x=1

The freeware version comes with it's own IDE. Also, plug-ins for visual studio as well as it's own visual development environment are available. I'm partial to XEmacs though.

--
Joseph K. Mooney
Director, Airframe Structures - FAA DER
Delta Engineering Corporation

 

[crackman]

joekm

Sorry for the long delay, been tied up. Here is the last known address and phone number for ASIAC who handles the release of computer programs and methods from Wright Patterson:

WL/FIBAD/ASIAC
Bldg 45
2130 8th Street, Suite 1
Wright Patterson Air Force Base, OH 45433-7542
Tel: (513) 255-6688
email: asiac@fltvc1.flight.wpafb.af.mil

ASIAC Manager: Gordon Negaard (not sure if current anymore)

There is a report with a listing of all available software, it is WL-TR-94-3127 "Guide to ASIAC Computer Programs". There are FEM codes, crack codes, and joint codes listed.

Hope this helps

James Burd
FAA DER - Fatigue and Damage Tolerance