rivet tension strength 2

[rb1957]

yes i know, rivets in tension is not the best idea, but sometimes we have to (little clips and brackets) and the loads are small and not vibrating, etc !

my question is about bruhn's tables in section D1.26. i've been using these for quite a while and just realised they don't specify a rivet material, AD, D, DD. They only specify AN470 or AN426 which is just MS20470 and MS20426 (no?)

i assume they're good for AD. any comments ?

 

[Lcubed]

Flabel recommends 2/3 of the ultimate shear allowable for protruding head rivets, which seems a little daring. The use of the 10% rule, that is, 10 percent of shear allowable, seems to be conservative according to all references that I have checked. Hope this helps.

Regards.

 

[SuperStress]

rb1957,

That's a very interesting problem, and a little investigation has shown me that there's apparently quite a bit of disagreement about how to handle this situation.

One company manual I have (that's from the early 1970's) is indeed less conservative than Bruhn's numbers (as Bruhn claims), but also does not specify rivet material.

A newer company manual I have (different company) from the 80's gives allowables by rivet type. Bruhn's numbers are conservative compared to the company manual for protruding head AD fasteners, however Bruhn's contersunk allowables are unconservative for even DD rivets!! Go figure?!

Anyone else have some data they can share?

SuperStress

 

[rb1957]

thx guys,

Lcubed ... 10% seems very conservative (only 40lbs for a BJ4), 2/3 = 260 lbs for a BJ4 looks reasonable

maybe 1/3 for CSK (BB4 would be good for 130 lbs)

as i thought about it, bruhn's table is for an AN470, and i thought that back then maybe they only had AD rivets ... where's will taylor when you need him !

 

[Sparweb]

Bruhn is usually conscientious about fastener materials, in other sections. Instead of thinking that this is an oversight on his part, it may reflect a conclusion that the material of the rivet is not significant. The Young's modulus of the rivet is the same regardless of the aluminum alloy, and the tables do specify the alloy of the sheets fastened together, so I suspect that the shape of the rivet head has a bigger role to play, the sheet material plays a lesser role, and the rivet alloy least of all.

Bill McCombs might have a bit to say on the topic in his Supplement, and you can find more rivet-in-tension data on page C11.50 of Bruhn, Figures C11.37(a) and (b).




Steven Fahey, CET

 

[wes616]

This was pretty interesting. Not an answer to the posed question, but short.

http://naca.larc.nasa.gov/reports/1944/naca-tn-930/naca-tn-930.pdf

Wes C.

 

[rb1957]

sparweb,

bruhn does specify sheet material "24ST, or harder".

whilst head type is important for tension strength (intuitively more important than shear), but rivet material has to be as important for tension (presumably failure is the rivet head parting company with the rest of the rivet, but possibly it is more of a gapping failure, which would be more dependent on E).

thx for the lead on C11, i'll see how that compares with the D1 data.


wes616,

your link is the same as bruhn pg C11.50 (tho' bruhn credits NACA 2661); thx anyways (it's amazing what you can find on the internet !!)

 

[rb1957]

bruhn credits "Martin Co. Structures Manual" ...

does this ring any bells ?

looking at the data steve found (fig C11.37) i think the data (section D1.26) is consistent with AD rivets.

 

[wes616]

Sorry for the double up...

Wes C.

 

[rb1957]

wes616,

no problem

 

[BVF]

Bill McComb's Supplement to Bruhn does have a few things to say. On page A10, he states, "Solid and blind rivets and blind bolts are not allowed in primary tension applications. They are allowed in secondary tension cases such as skins with aerodynamic lift forces or internal pressure, attachment of shear or compression webs, and stringers to frames.

"*For secondary tension an effective tensile "pull-thru" allowable of 20% of the ultimate joint strength allowable (at the specific sheet thickness) is considered to be conservative. This approach is recommended for any fastener type *except* those having shear heads and reduced head features or if specific data has been developed. For NAS 1097 reduced shear head rivets use 10% of the particular ultimate joint strength allowable."

Then, on page A15, he has a table for MS20470AD (protruding head) rivets in 2024 and 7075 sheet, and a similar table for MS20426AD (flush). Values range from 140 to 416 lbs for the MS20470AD4, and from 230 to 375 lbs for the MS20426AD4.

It's not clear to me the source of the data.

McComb's Supplement is frequently a useful reference.

Brent

 

[rb1957]

thx brent,

McCombs table A20 (pg A15) is a little lower than bruhn D1.26, so i'll use it in future.

 

[crackman]

Hi All

The Grumman Structures Manual and also the handy little pocket size Grumman Structural Design Data booklet has tension allowables for the following:

Prot. Head Al Rivets in 2024-T3 (Ref. NASA-TN 930)
MS20426-DD Csnk in 7075-T6 (Ref. Grumman Test Rep. GE-148)
NAS1097-AD Csnk in 2024-T3 (Ref. Grumman Test Rep. GE-148)
AN427 Monel Csnk in 7075-T6 (Ref. GE-148)

James

 

[WKTaylor]

Guys... Taylor’s Rule-of-Thumb to using rivets in a shear-tension application…

1. Use of driven rivets in shear-tension applications has a checkered history of success and failure. Obviously Bolts***, with washers and substantial nuts, can perform reliably and repeatable in shear, tension and shear-tension joints.

2. How to use rivets in tension-shear applications begs an understanding of the variables.

2.1 The formed head has very reliable geometry including head size, transition radius and grain flow.

2.2 The bucked-tail ["driven head"] has extremely unreliable geometry. Tail protrusion determines potential bucked-tail dimensions (driven perfectly): a short tail will result in a minimum diameter/height bucked “head”; and a long Tail will result in a large diameter/high bucked “head”. If driven incorrectly, You can have a wide variety of bucked head imperfections, the worst of which include: too shallow a head: toe-nailed head, off-center head, under-bucked diameter, cracking/chipping, etc…

2.3 Hole geometry, including roundness, size, roughness, hole-edges (burrs, excessive chamfer, etc) and depth of material stack-up can have a profound affect on rivet installation [driving] and mechanical performance [both shear & tension].

2.4 See MIL-STD-403 and NASM47196 for allowable installation and bucked-tail geometry [acceptance] limits. Very unnerving as to what can happen… and what is acceptable.

3. NOW, can rivets be used successfully in tension or tension-shear joints??? I believe they can, but ONLY with installation precautions and conservative analysis, as follows. NOTE: it is up to the designer and analyst to put “numbers” on the following guidance.

3.1 Installation precautions.

3.1.1 Shall not apply to reduced-head [shear] rated rivets.

3.1.2 Specify that rivet-heads be on primary tension side [bracket/clip, fitting, side]. The head geometry is a reliable size/shape and can be analytically evaluated for tension.

3.1.3 Rivet installation shall be per MIL-STD-403 and NASM47196. NO EXCEPTIONS, especially in bucked tail dimensions.

3.1.4 Hole depth [material stack-up] should be at least 1 x D. The tension-side member and bucked-side thickness MUST be at least 0.33 x D. If total thickness is less than 1 X D, then back-up bucked-tail with a 0.032 or 0.063 thick washer or plate [aluminum or steel, deburred]. This insures that the rivet shank provides high interference friction to resist “pull-thru” and “loosening” in shear-tension loading… in addition to the cross-section of the deformed bucked tail.

3.1.5 A minimum of (2) rivets shall be in tension-shear. NOTE: all rivets loaded in tension-shear should experience “balanced loading”, IE: the centroid of load and centroid of the fastener pattern (in tension) should be closely aligned to minimize eccentricity [unbalanced fastener loading].

3.2 Analysis (ultimate load).

CAUTION: use rivets in tension for lightly-loaded structural purposes ONLY. If failure occurs, then safety of aircraft cannot be compromised.

3.2.1 P (tension-ult) = ((nominal rivet Dia/2)**2) x (Pi) x (nominal depth of head per spec) x (alloy shear strength from NASM5674)
Note: Fsu is used as opposed to Ftu since head will be shear critical thru the fillet radius “upward” [also isolates from rivet-shank-shear inter-action].

Example for MS20470AD4-X
P (tension-ult) = ((0.125-in / 2)**2) x (3.1416) x (0.054-in) x (26,000#/SqIn) = 17.3 pounds-force

Example for MS20470AD6-X
P (tension-ult) = ((0.187-in / 2)**2) x (3.1416) x (0.080-in) x (26,000#/SqIn) = 59 pounds-force

Example for MS20470E6-X
P (tension-ult) = ((0.187-in / 2)**2) x (3.1416) x (0.080-in) x (41,000#/SqIn) = 89.9 pounds-force

3.2.2 Analysis shall verify a minimum tension margin of safety for ultimate load as MS = 0.5 [equates to a SF= 1.5 + 0.5 over limit load] for P (tension-ult). This “conservative” loading will provide added margin for “maintenance abuse”. Note: this is in addition a +MS for the rivet in shear [small decrease in shear due to a tiny cross-over from the head-in-tension will likely be insignificant in the “noise” of the analysis… unless MS = 0.0]

4. Comments.

4.1 I strongly recommend using at least an AD6 rivet or larger “just because”.

4.2 The low performance of rivets suggests “why” industry rarely uses rivets in significant tension.

4.3 A typical AN3-X ALUMINUM Bolt has a P (tension-ult) rating = 1,100#, per specification (although *I* would not stress it above 66% of that number… especially in a fatigue environment!!!). Use an aluminum “AN” nut with matching tension capability for a light weight Installation. Note: a tension-head Aluminum Hi-Lok and mating collar would also have good tension strength! NOTE: for this high load capability, an elliptical load distribution envelope for Shear-Tension allowables, should be used.

Bolts*** = male fasteners with hex/pan/flat/countersunk/etc heads (full or reduced size), NOT including male fasteners with full-threaded shanks [screws]. Lock-bolts with pins and swaged or threaded collars also “fit” this definition.

Regards, Wil Taylor

 

[rb1957]

Wil,

thx very much for your thoughts. some very interesting practical installation thoughts. i think you may file the rest of this post as "well, you can lead a horse to water ...".

i'm surprised that your allowables are SO low, particularly if you then recommend using only 2/3rds of the allowable. i agree that a predictable failure mode would be shear of the head ... wouldn't the shear area be piDt (you've used piR^2t, which [pedantically] throws off your units). if this area was used, the AD4 allowable would become 17.3*32 = 554 lbs, and 2/3rds is 390 lbs, much closer to McCombs data.

another failure mode, which your rules of stack up aviods, would be pulling the head (or the tail) though sheet.

you could also get a tension failure of the shank, but this is probably higher than the shear failure of the head. thinking about bolts (we'll happily calculate the tension allowable based on the min. area), bolt head profiles must reasonably protect against this type of failure, except for CSK heads

 

[BVF]

(piDt)(Fsu) times an appropriate reduction factor makes sense here. I would still prefer to use values tabulated from empirical data.

Brent

 

[WKTaylor]

rb1957 & BVF...

Apologies to all for my initial "unacceptable reply" (above)...

The equations/concepts I presented above are fundamentally flawed as noted. Somehow a major order-of- magnitude error just didn’t register in my brain ["aw shit"]. I suppose the lesson here is never-ever write technical prose on the eve of surgery... especially when you have no chance for a logic-check on the results before publishing.

Here is what I ment to say, based on conversations with "gray-hair" types... [I mixed parts together from another issue...Duhhhhhh...]

Scratch out pqara 3.2 above and replace it with the following [someone please provide a math/logic ckeck on this... just-in-case].

3.2 Analysis (ultimate load).

CAUTION: use rivets in tension for lightly-loaded structural purposes ONLY. If failure occurs, then safety of aircraft cannot be compromised.

3.2.1 P (tension-MAX) = (nominal rivet Dia) x (Pi) x (nominal depth of head per spec) x (driving factor) x (FSu from NASM5674) x (Fty/Ftu from any reliable source) x (0.50)

Notes:
Driving factor = change in head height [crush-flattening] during riveting = roughly 0.80-to-0.95 (depending on rivet alloy/temper, “spring-back” and driving irregularities).

FSu is used as opposed to Ftu since head will be shear critical thru the fillet radius “upward” [also isolates from rivet-shank-shear inter-action]. I believe that this ratio should NOT be greater than 0.75, under any circumstances

Fty/Ftu is use to represent approximate ratio of shear yield/ultimate [pertinent for the alloy]. Once a riveted joint yields it should be repaired… or cyclic loading will destroy it

(0.50) ~ Factor for abusive loads and long-term durability

Caution Note: if operating at elevated temperatures (+180F), then include reduction factor for estimated temperature/time exposure.

AD = 2017-T4 [RT]
FSu = 26,000#/Sq-In
FTy = 32,000#/Sq-In
FTu = 55,000#/Sq-In
FTy/FTu = 32/55 = 0.581

E = 7050-T73 [RT]
FSu = 39,000#/Sq-In [41,000 driven)
FTy = 60,00#/Sq-In
FTu = 70,000#/Sq-In
FTy/FTu = 60/70 = 0.857 = 0.75 [arbitrary]

3.2.2 Examples

MS20470AD4-X
P (tension-Max) = (0.125-in) x (3.1416) x (0.054-in) x (26,000#/Sq-In) x (0.80) x (0.581) x (1/2) = 128#F

MS20470AD6-X
P (tension-Max) = (0.187-in) x (3.1416) x (0.080-in) x (26,000#/Sq-In) x (0.80) x (0.581) x (1/2) = 283.5#F

MS20470E6-X
P (tension-Max) = (0.187-in) x (3.1416) x (0.080-in) x (41,000#/Sq-In) x (0.90) x (0.75) x (1/2) = 482#F

>>These results make sense relative to real-world load capacity<<

3.2.2 [NA]

Regards, Wil Taylor

 

[BVF]

>>>P (tension-MAX) = (nominal rivet Dia) x (Pi) x (nominal depth of head per spec) x (driving factor) x (FSu from NASM5674) x (Fty/Ftu from any reliable source) x (0.50)<<<

Thanks for clarification, Wil. There's a lot of good thinking here. The 0.50 fudge factor on the end seems a bit conservative (not a bad thing for a calculated allowable).

I would add a cautionary note that the sheet thickness has a significant effect on the tensile strength of the joint. The sheet can deform, or the head could even pull through before the fastener itself fails. The tables in McComb's book, and those in Boeing Design Manual BDM 4301, have values dependent on sheet thickness. Wil's calculated values are at the low end of these tabulated values.

Brent