Huth Formula for Fastener Flexibility 6

[737eng]

I asked this in another forum as well. But I thought a discussion on different methods of determining fastener flexibility may begin here. I have been using Tate and Rosenfeld and recently read the paper by Huth and felt that since his formula was determining lower compliance that the increased stiffness would produce more conservative results. However, unlike the T&R, it appears that the Huth is dependent on which element you make T1 or T2 (for single shear joints).
My question from the other forum:
In the Huth formula for fastener flexibility (Huth on Influence of Fastener Flexibility). If you have a single shear joint with two different sheet thicknesses (t1 and t2), which sheet thickness do you utilize for t1 and t2(i.e. the thin sheet for t1 and the thicker sheet for t2 or vice versa). Due to the 1/2nt2E3 factor in the right hand side bracket, this multiply by two in the denominator can make a difference depending on which thickness you utilize. Also, since this equation is also good for double shear with n=2 in lieu of 1, I am wondering if this 2 is a typo. My example: I have a 0.056" thk skin doubler on a 0.050" thk skin, I am obtaining my fatigue stress from the original skin which I am making t1, the doubler, which is thicker, is then t2.

Now I understand that in the grand scheme of things this difference of 0.050 and 0.056 is going to have an insignificant affect on the calculation. Additionally, for most joint designs, the two parts should be close in thickness which would also provide an insignificant change. However, it would be nice to now which way the formula was inteneded.

 

[rb1957]

T&R (which I assume is NACA TN1051) is pretty limited in its application (maybe it's enough for what you need); or is there another report with more material combinations ?

 

[737eng]

Tate and Rosenfeld NACA TN1051 is what I was referring to. I don't have any other references for T&R. I have seen it referenced in other papers, but nothing that expanded upon it. It typically meets our needs. If required we do utilize other methods which most are aware of or we will just model the joint using beam elements.

 

[nveneno]

The Huth formula it’s not the only way to get fasteners stiffness, there are also some different approaches to the problems with different results also.

-Swift (Douglas)
-Tate & Rosenfeld
-Huth, Schwarmann (Airbus)
-Vought
-Grumman
-Delft University

Which one gives the right answer? That’s the question

Take a look to the attached document to see the details of every formula (I hoppe it works)

Veneno

 

[40818]

nveneno:

They all give the right answer, at least for the data they were based upon. Thats the key really. So the method you should use is the one that is agreed with the powers that be.

 

[rb1957]

yeah, there's no "correct" formula, taking the "legalistic" interpretation that one is "right" and all the others are (by default) "wrong".

there are an enormous number of varibles that affect different fastener geometries differently, and each author has assumed the shape of the interaction and which factors are going to be most significant. if you look at the different solutions they work in contradictory ways and give wildly different results; at best each matches a set of data and one or another is preferred by whoever is looking over your work (as 40818 says, "the powers that be").

i think the important point is that whichever formula you use the results doesn't change in a really significant way. if you have the time use a couple of formulas, see how the results vary ... if it is something that is really critical and you're getting questionable results, i'd then suggest a test.

 

[nveneno]

Hi, thanks 40818 & Rb1957

A few years ago, in my internship, I made some test in Nastran simulating fasteners joining two sheets, with RBE (rigid elements), CBUSH (6 dof springs), CBAR (beam elements), and CFAST (connector elements with 6dof springs), I also use different model to get the fasteners stiffness (only for CBUSH & CFAST).

In general terms, the stresses in the shell are quite similar in every model, but the stress in the surrounding areas of the fastener and the loads carried by the fasteners itself varies widely. I didn’t make any experimental test to validate the model

So that’s why I ask again, which one gives the right result?

veneno

 

[rb1957]

all and none ... each has been matched to a specific set of data (so each gives the correct answer for a limited set of geometry) but none successfully account for all the variables in the design (so there is at least some designs for each formula where the fastener model doesn't work).

i have modelled joints with CELAST springs; i chose the douglas model 'cause it was conveniently cloe to hand, and was happy in my work ... i'm not going to say that the douglas model is superior to the others, but it is an industry standard so it has some "street cred."; any of the other standard models would have worked out equally well.