Shear tearout of bolt in structural standards (AS4100) 2

[Nereth1]

Hi all,

This comes from a mech eng trying to understand a structural engineering calc and why it contrasts so hard with mech engineering calcs. Typically when I analyse a bolt in a hole near the edge of a steel plate, based on sources such as Rourkes, Shigley's, and Machinery's handbook;

Mechanical practice - Bearing of the bolt against the hole edge:

Plate yield point = [bolt diameter]*[plate thickness]*[tensile yield limit]. Obviously a small amount of yielding will be required before this as well for the purposes of the bolt getting full contact with the hole edge.

Mechanical practice - shear tearout of the plate:

Plate yield point = [edge distance]*[plate thickness]*[2 shear planes]*[0.577*tensile yield].

By contrast when I look at AS4100-Steel structures:

Structural practice - Bearing of the bolt against the hole edge:

Bearing ply capacity = [0.8 capacity factor]*[3.2]*[bolt diameter]*[plate thickness]*[tensile strength]

I understand designing around ultimate strength rather than yield (structural engineering seems to be pretty plastic-deformation accepting in general). What I don't understand at all is the 3.2 factor. I can only speculate this is included as some kind of respect for triaxial stress states from compression of the bolt head, but not all bolts are tightened down particularly hard, no specific requirement for tightness is made, and even so, 3.2 seems like a massively excessive factor even for that. To me, by definition a ductile material (i.e. steel) has failed at 1*bolt diameter*plate thickness*tensile strength, if not, your tensile strength is wrong. I would really like that 3.2 factor explained to me.

Structural practice - shear tearout of the plate:

Tearout ply capacity = [0.8 capacity factor]*[thickness of ply]*[tensile strength of ply]*[edge distance]

- They are using tensile instead of shear properties - this I don't understand unless it relates to the next point
- They use a single shear plane - or perhaps they are just analysing for the plate 'popping' apart in a tensile manner?
- They are using ultimate rather than yield strengths (this I assume is just the general acceptance of yielding in a lot of structural engineering)

 

[Agent666]

See below for bearing and tearout derivation from the commentary to NZS3404 (NZS3404 is virtually identical to AS4100 apart from our seismic stuff and a few other minor differences, so check AS4100 commentary for similar):-

The only difference in NZ is the strength reduction factor = 0.9 vs the 0.8 you note for Australia.

Hopefully that answers your questions at the end if you follow the derivation.

I will note in NZ higher minimum edge distances apply for seismic load cases, in effect to eliminate the tearout failure mode, typically this is 2 times the bolt diameter in the direction of the load and 1.5 times the bolt diameter perpendicular to the load direction.

At these limit states there is significant yielding around the bolt holes as the material piles up or locally deforms, as structural engineers this is considered normal practice. Search the internet for examples of tests to the ultimate capacity of typical connections and the like to see what performance is expected in terms of deformation. As a mechanical engineer it probably quite a different approach to what you are used to where mechanical design generally intends to make sure the element doesn't even get close to yield.

 

[Harste]

Hello Nereth1,

Most of your queries are covered in the supplementary commentary to AS4100, and the associated reference documents from which AS4100 is based. Answers to your more intricate questions are too much to re-type in this forum, but if you can get your hands on a copy I think you'd find it to be very beneficial.

As for the 3.2 factor, the commentary says:
"Research generally indicates that shearing-tearing failure with considerable ‘piling-up’ of the ply material in front of a bolt, commonly referred to as a local bearing failure occurs at a nominal bearing stress within the range 4.5fyp to 4.9fyp. Hence, using the lower limit and the conventional nominal bearing area (df × tp) leads to—
Vbu = 4.5fyp.df.tp

where Vbu is the ultimate bearing capacity of a ply. For most structural steels, fyp ≈ 0.7fup so that this limit is equivalent to a nominal bearing capacity of a ply (Vb) given by—
Vb = 3.2fup.df.tp, as used in Clause 9.3.2.4."

Hope this helps.

 

[Nereth1]

Thanks guys - this helps a lot. Sounds like the basic reasoning for the 3.2 is "We tested it and that's reality". And they used UTS instead of yield strength because it was somehow more convenient, but then downrated it mathematically to yield strength.

I wander if the tests included tight bolted joints or if it's still true with a loose pin in a plate with no clamping pressure. I'll have to search for some myself.

 

[Agent666]

Nereth1, in NZS3404 & AS4100 there are different bearing criteria for the pin itself in bearing, the ply in bearing is the same as for bolts, refer to clause 9.5 (AS4100).

 

[Agent666]

FYI reading through NZS3404, it is actually different to AS4100 and that it notes that you design both the pin and the ply for the same criteria (as AS4100's pin design) due to some testing undertaken in 1999 that showed treating the ply the same as for a bolt was actually unconservative, resulting in the pin hole elongating (which is clearly undesirable for any dynamic loading. The suggested change was written into the standard as an amendment shortly afterwards. Just though I'd clarify this aspect as it suggests the AS4100 provisions have the same potentially non-conservative problem.

See attached on page 9 where this is discussed and some of the reasoning behind the changes.

 

[ajh1]

There is some current research ongoing in the cold-formed steel arena looking at this issue. One of the big questions that is intended to be addressed is that there appear to be two distinct forms of the failure depending upon I suspect the ratio of the width of the plate vs. the distance from the hole to the end of the plate. In the first case, the material tears in two reasonably parallel lines approximately the width of the hole apart and is primarily a shear failure. In the second case, the material fails in tension and the two sides of the plates tend to spread and separate between the hole and the end of the plate. What is yet to be determined is when does one occur vs. the other. The other fun question lately has been how do you measure the distance from the hole to the end of the plate in whatever equation you come up with. Is it based on: 1) center of hole, 2) edge of hole, 3) some value part-way between those two values.