1 9 Design of Joints Loaded in Shear
19.1 AN OVERVIEW
Eighteen chapters ago I described two kinds of bolted joints: those loaded in tension and
those loaded in shear. With the exception of Chapter 12, the discussion since then has been
focused on tension joints because they're more common, their behavior is more complex, and
analyzing them is more difficult. In this last chapter, however, we're going to take another
look at the shear joint. To be specific, we'll study the design of such joints and will see the
ways in which the design process is the same as that for tension joints, and the many ways in
which the two differ.
‘There are many different types of shear joint, but most can be defined as either a lap joint
or a butt joint, as shown in Figure 19.1. Historically, joints of either type were further
classified as “friction type” or “bearing type.” The structural steel industry has now aban-
doned the friction and bearing classifications, as we'll see, but the distinction is still handy for
a preliminary review of shear joint design, and so I'll continue to use it.
Shear joints are most commonly encountered in structures, such as airframes, buildings,
and bridges. Most of the bolted joints found in structures, in fact, are shear joints. In part this
is because of the way loads are applied to structures, but I suspect that part of the reason is
that shear joints are more forgiving of assembly errors or preload scatter; they can operate
successfully under a much wider range of clamping force than can many tension joints. One of
the main reasons for this is that shear loads don’t change bolt tension or clamping force the
way tension loads do.
‘One new problem we do have to be concerned about, however, when dealing with shear
joints, isthe possible mechanical failure of the joint members themselves. Tension joint failure
can usually be blamed on the bolts; either they have created the wrong clamping force or they
hhave themselves failed. Improper clamp can cause a shear joint to loosen under vibration, but
‘most shear joint failures involve the rupture of the joint members.
We're going to start our study of shear joint design with our old friends, the VDI
equations. We'll see what they have to teach us about shear joints, and will find that it's
useful but not enough. So we'll go on to look at the way the bolts and joint members see
and resist shear loads. All of this will be pertinent for the design of shear joints in general. As
we go along we'll take an occasional look at some of the codified design procedures, which
have been developed by the structural steel industry. Those procedures are described and
explained in detail in the definitive text, Guide to the Design of Bolted and Riveted Joints by
Kulak et al. [7]. This complete text is now available, free, on the Research Council on
Structural Connections (RCSC) Web site. Structural steel designers should rely on this text
and on the bolt specs also written by the RCSC [8,13] rather than on my text. The bolt spec is
also available on the RCSC and AISC Web sites. I'm not going to repeat the procedures
described in those documents here. But some comments are certainly in order.
443444 Introduction to the Design and Behavior of Bolted Joints
FIGURE 19.1 Two basic types of shear joint. The upper is called a lap joint; wo joint members are
bolted to each other. The lower is called a butt joint; the joint members are connected by upper and
lower splice plates.
19.2 THE VDI PROCEDURE APPLIED TO SHEAR JOINTS
Let’s return, for a final time, to the VDI joint design equations first encountered in Chapter
18 [II]. These raise two important issues we must address when designing any joint: the
minimum clamping force we can expect to see in the joint and the maximum tension the bolts
will have to support. Since clamping force is assumed to equal bolt tension, these equations
are expressed in terms of clamp force (F}) but really define bolt preload limits. I repeat those
equations here, for your convenience.
Min Fj — AF) | Fou | MF (18.5)
Since AF; — (I~ @k,)Lx, this ean also be expressed as
Min Fy — (1 ~ xa) + Figs + SF
We also have
Max Fy = agl(l —@bxa)Lx + Fings | AFoI 8.6)
where
AFp __~ loss of preload during or soon after assembly (Ib, N)
Max F; — maximum anticipated bolt tension created during assembly; equals the clamping
force that bolt applies to the joint before the joint is put in service (Ib, N)
Min F — minimum anticipated bolt tension and clamping force created during assembly,
before the joint is put in service (Ib, N)
Fog — minimum ‘preload (and per-bolt clamping force) required to prevent joint
failure (Ib, N)
Ix external tensile load applied to the joint (Ib, N)
an the scatter in preload caused by the assembly tools and procedures
aq — Max Fy/Min Fy
xq — load factor for a concentric joint, loaded internally at loading planesDesign of Joints Loaded in Shear 445
gy — AFy/L, where Fy ~ the change in tension created in the bolts by the external load
This is the load factor we used when discussing tension joints in general
AF, ~ the change in clamping force (or bolt tension) ereated by an external tensile load
on the joint ib, N)
Let's look at each of these terms and see how they apply to shear joints,
AF;: Since there is no tensile load on a joint loaded only in shear, AF, ~ 0.
AF: When designing a tensile joint we had to consider four ways in which initial preload
ight be lost during or after assembly, giving us this expression:
AFp = AFem | AF | MFcu AF as.)
where the subscripts
‘em — embedment relaxation
EI — elastic interaction loss
CR — creep loss
‘TH ~ gain or loss because of thermally induced differenti
expansion,
Each of these effects can cause a loss of assembly preload in a shear joint as well as in a
tension joint, but the effects are usually smaller, with the exception of embedment. Let’s look
at each effect as it applies to a shear joint
Enbedment: We can expect to see a typical embedment loss, perhaps 5%-10%.
Elastic interactions: Bibel tells us that the average elastic interaction loss in an ungasketed,
‘metal-on-metal joint is 18%, well below the values for gasketed joints 10}, This figure, however,
is based on limited tests on 24 in. diameter, raised-face, pressure vessel joints, where the bolts
are unsupported by metal-to-metal contact. I would expect to see less oss in most shear joints:
but that's just a guess.
Creep: Some structural steel joints are given a thin coat of paint, to control interface
friction and to provide some corrosion protection, and so we might see some creep loss after
assembly. I would expect the loss to be negligible, however, certainly nothing like the major
loss created by a gasket
Thermak In most of the (few) structural steel joints I've studied the bolts and joint
members have both been made of steel, presumably with similar coefficients of expansion.
Unlike, say, a pressure vessel joint, both bolts and joint members would experience the
same change in temperature at the same time. And the temperature change would be modest,
again unlike pressure vessel applications, where temperatures of 1000°F or more are
not uncommon. All of which suggests that differential expansion can be ignored in structural
steel joints.
Airframe structures, on the other hand, often involve several materials, including
aluminum joint members and bolts of ferrous metals or exotic alloys. Temperature changes,
however, are still modest. Grip lengths, furthermore, tend to be small in a structure which
must be light enough to fly. I'm sure that airframe designers take thermal change into
account, but I doubt if differential expansion is large enough to be a concern. All of which
suggests that in most shear joints, AF) is probably going to be less than 30% of initial preload,
perhaps much less.
‘iq: Shear joint bolts will be subject to the same preload scatter as tension joint bolts. If'a
torque wrench is used on unlubricated bolts, for example, scatter might be +30”. Following
VDI’s lead, this would create an a4 of (I. 0.3),(1—0.3) or 1.86. There are some factors that
make tool Scatter different when we're dealing with a shear joint, however. First ofall, preload
control is less important in most shear joints, so scatter doesn’t matter as much. Second, in