Lecture 11.5: Simple Connections for
Buildings
Simple connections are defined as joints between members that have
not
been designed with the intention that they transmit significant
moments. Their purpose is to transfer load from the supported member
into the supporting member in such a way that essentially only direct
forces are involved, e.g. vertical shear in a beam to column or beam to
beam connection, axial tension or compression in a lattice girder chord
splice, column base or column splice connection. They may, therefore,
only be used in situations where sufficient bracing is present that,
when the joints are assumed to function as pins, adequate overall
structural resistance is present. Popular arrangements include lattice
girders and bracing systems or connections between beams and columns in
rectangular frames in which lateral loadings are resisted by stiff
systems of shear walls, cores or braced bays.
Figures 1a and 1b illustrate multistorey frames in which simple
connections may be used for each of the 6 different requirements A-E
listed alongside Figure 1a. Thus the structural idealisations suitable
for determining the distribution of member forces will be as shown in
Figure 1c and 1d, with all lateral loading being resisted by the bracing
or shear wall. When considering the design of the frame to withstand
gravity loading, the assumption of pin connections makes the overall
structural analysis particularly straightforward, since loads can be
traced from floors into beams into columns and eventually into the
foundations using a simple statical process.
Simple joints also lead to easier fabrication and erection and as explained in
Lecture 11.1.1
are, therefore, likely to produce the most cost-effective steel frames.
Taking the example of a beam to column connection, the simple joint
must:
- transfer the beam reaction into the column in shear
- have sufficient flexibility not to transfer other than small moments
into the column, e.g. due to some small eccentricity in the lines of
force transfer
- possess sufficient rotation capacity to permit the beam to develop its "simple" deflected shape.
Thus, in terms of the classification system introduced in
Lecture 11.1.2.,
the connection should function as "nominally pinned" for both moment
capacity and rotational stiffness and the only form of load transfer
required will be the vertical shear illustrated in Figs. 9(2) and 11 of
that Lecture.
Simple connections will normally be either fully bolted, e.g. the arrangements using angle cleats of Fig. 10 of
Lecture 11.1.1,
or will involve a combination of shop welding and site bolting, e.g.
the fin plate and end plate arrangements of the same Figure. Except for
connections subject to vibration, e.g. in foundations for moving
machinery or in crane support structures, untorqued bolts in clearance
holes should be used.
This lecture discusses the structural design of several examples of
each of the 6 connection arrangements listed in Figure 1. In doing this
it makes use of basic material on weld strength and bolt strength
presented in
Lectures 11.2 and
11.3 respectively, as well as the approach to the analysis of connections given in
Lecture 11.4.
Floor decks in buildings are usually supported by means of grids of
secondary beams and main girders simply connected to each other.
Some typical connections are illustrated in Figure 2. Types A and C,
which make use of web cleats bolted to both the girder and the beam, are
the most common forms. Type B with the cleats bolted to the girder and
welded to the beam, and types D and E where a flush end plate is
adopted, may cause lack-of-fit problems during erection due to the
dimensional tolerances. Connection types D and E possess some
predictable stiffness and strength, but their consequent partial
continuity is usually neglected in design.
As shown in types C and D, the beam end may be coped removing part of
one or both flanges, when the beam and girder flanges meet at the same
level. The beam is thus locally weakened. The appropriate checks must be
made as discussed below. Nevertheless, this solution is less expensive
than type E, which requires that a tee stiffener is welded to the
girder.
As a variant to A the web angles may be replaced by a fin plate, as
shown in Type F, a single plate which is shop welded to the primary beam
and site bolted to the secondary beam. A fin plate connection is
particularly simple to both fabricate and erect, but it requires careful
design if it is to function as a notional pin [1]. In particular, there
is a need to decide where the "hinge" is located as explained in
Section 3 of
Lecture 11.6.
For web cleated connections, it is good practice to place the angles
as close as possible to the upper flange of the girder in order to
minimise cracking of the concrete floor slab due to the beam rotation.
Bolts and welds in connections should be able to resist the beam
reaction and any relevant moment due to the eccentricity of the force to
the centerline of the connecting components as explained in Section 2
of
Lecture 11.4.3.
When a beam is coped, as in connection type C, it should be verified
that no failure may occur at the section that has been weakened (block
shear) as explained in Section 2 of
Lecture 11.4.3.
Several forms of simple beam-to-column connections are illustrated in Figure 3.
Type A, which is shown as fully bolted, may also be configured by
welding the cleats to the beam end. For lightly loaded beams, a single
sided cleat may be used but the additional eccentricities must then be
allowed for when checking bolt strength, etc.
The finplate Type B requires the same form of attention when deciding
on the design model as discussed in the previous section where its use
in beam to beam situations was discussed. It is one of the few
arrangements obviously suitable for use with SHS (either RHS or CHS)
columns as no bolting to the column is necessary.
Both types A and B provide some allowance for tolerance (through the
clearance in the beam web holes) on member length. Type B permits beams
to be lifted in from one side.
Types C and D require a more strict control of beam length and of
squareness of the cross-section at the end of the beam. The flush end
plate scheme of type D is sometimes preferred to the part depth end
plate (type C) in order to reduce the chances of damage during
transportation. Partial depth endplates should not normally be less than
about 0.6 times the beam depth or the end torsional restraint to the
beam may be reduced. Figure 4 illustrates how flexibility and rotation
capacity is provided. Depending on the details, the connection behaviour
of type D could change from a notational pin; it may be more
appropriate to acknowledge this semi-rigid behaviour (see
Lecture 11.7).
This may be avoided by keeping the endplate thickness down to a maximum
of 8-10 mm and making the bolt cross-centres as large as is practical
so as to ensure adequate flexibility and rotation capacity.
As for beam-to-beam connections, the bolts and the welds should be
able to resist the beam reactions and the relevant moment due to the
eccentricity of the force to the centreline of the connecting material
as explained in
Lecture 11.4.3.
Since this eccentricity is relatively small the column bending moment
for such a connection is much smaller than from a moment connection as
discussed in
Lecture 11.6.
Since the general approach to the design of all forms of simple
connections is essentially the same, it will be sufficient to consider
only one type in some detail. Figure 5 illustrates the 6 possible
failure modes for a finplate connection; the load carrying capacity for
each must be calculated and the lowest value compared with the design
requirements. Methods for doing this have already been presented in
Lectures 11.4.
It is also necessary to ensure - usually by means of appropriate
detailing - that the connection will function in the manner intended,
i.e. will not be too stiff and will possess adequate rotation capacity.
This may be achieved by:
- ensuring that strength is governed by a ductile mode of failure.
Bearing of the bolts in either the finplate or the beam web is
usually arranged to form the governing condition. When performing the
structural checks it is necessary to be consistent in the assumption of
the location of the line of shear transfer, i.e. the "hinge" line. One
approach (1) that removes the need for a decision is to design
both
the bolt group and the welds for the combination of shear and
eccentricity moment. Alternatively, the location can be chosen as the
bolt group for the stiff support arrangement illustrated in Figure 5 or
the weld if the support is more flexible as would be the case, for
example, if a RHS column were used (due to bending of the column face as
a plate).
In simple frames columns are predominantly stressed in compression.
In theory no splice connection is required, since the compression force
is transmittable by direct bearing. Due to the presence of geometric
imperfections (lack of straightness of the column) as well as of
unavoidable eccentricities, and to the fact that even carefully machined
surfaces will never assure full contact, connections have to be
provided. They should be designed to resist the internal forces (other
than compression) determined in the column at the point where they are
located.
Even when the column is subject to simple compression, and full
contact in bearing is assumed, codes specify stiffness and strength
requirements to be fulfilled. Eurocode 3 prescribes that the splice
should provide continuity of flexural stiffness about both axes, and
should be able to carry a force, acting at the abutting ends in any
direction perpendicular to the axis of the member, not less than 2,5% of
the compression.
The location of the splice should be selected so that any adverse
effect on column stability is avoided, i.e. the distance of the
connection from the floor level should be kept as low as possible. A
limit of 1/5th of the storey height is usually accepted. If this
requirement cannot be fulfilled, account should be taken of the (second
order) moment induced by member imperfections.
More significant bending resistance may be required in splices when
columns are subject to primary moments, as in a frame model assuming
hinges at, or outside, the column outer face. In addition, in columns
acting as chords of cantilever bracing trusses, tensile forces may arise
(uplift) in some loading conditions, which must be transmitted by
splices.
Typical column splices suitable for use in simply designed frames are
shown in Figure 6. They are of two basic types: A, B and C all transmit
the whole of the force through the cover plates, whilst D-G rely on
direct bearing.
When a bolted solution is adopted (types A, B and C), both flanges
and the web are usually connected. Type A uses a double cover plate,
whilst type C uses single cover plates for the flanges. These may be
positioned on the outside faces of the flanges so as to reduce the plan
area occupied by the splice. Forces are distributed among the connecting
plates in proportion to the stress resultant in the cross-sectional
elements, e.g. for simple compression in proportion to the areas of the
flanges and of the web. Differences in column flange thickness may be
accommodated by the use of packs.
When the surfaces of the end cross-sections of the two column shapes
are sawn and considered to be flat, and squareness between these
surfaces and the member axis is guaranteed, the axial force may be
assumed to be transmitted by bearing. Fillet welds (type D) or light
cover plates (type E) are provided to resist possible secondary shear
force and bending moment when the upper and lower columns differ in
serial size. A plate may be interposed, and welded to both column
sections as in connection type F, or, alternatively, two welded plates
bolted to each other may be used (type G). Plates are flattened by
presses in the range of thicknesses up to 50m, and machined by planing
for thicknesses greater than 100mm. For intermediate thicknesses either
working process may be selected.
Where there is a significant variation of cross-sectional dimensions
in the arrangement of type F, the plate(s) must be checked for bending
resistance. A possible conservative model assumes the plate is a
cantilever of breadth equal to the width of, and clamped to, the upper
column flange. The axial force, which is transmitted between the
corresponding column flanges, is applied as an external load at the mean
plane of the flange of the lower column.
Full details of this approach are presented in ref. 2, from which it
is clear that if plate thicknesses are to remain reasonable, then only
moderate offsets of the order of the column flange thickness are
possible. For larger differences in column size, a short vertical
stiffener may be located directly below the flange(s) of the upper
column to directly assist in transferring the locally high force.
Connections within the bracing system or between the bracing system
and the main framing have to transfer forces between a number of
differently oriented members. Since the triangulated bracing arrangement
will have been designed on the basis that each member carries only
axial forces (apart from any relatively small bending effects due to
non-coincidence of centroidal axes), the design requirement for the
bracing connections is essentially the transfer of direct forces between
a number of differently oriented members.
Two basic arrangements are illustrated in Figure 7: Type A attaches
the bracing to the main framing, Type B is an internal bracing
connection. Types C and D combine both functions by making the beams
part of the bracing system. Details of the design considerations and the
calculations necessary to effect these have already been provided in
Section 1.3 of
Lecture 11.4.3.
A column base connection always consists of a plate welded to the
foot of the column and bolted down to the foundations. A second, usually
rather thicker, steel plate is normally incorporated into the top of
the foundation, as illustrated in Figure 8. It helps both to locate the
foot of the column accurately and in spreading the load into the weaker
(concrete or masonry) foundation material.
Baseplate connections in simple construction are generally modelled
as pins, and designed to transfer either concentric force (compression
or tension) or a combination of axial and shear force (usually when the
column is part of the bracing system (Figure 8c)). In some instances
they may, however, be designed to transmit also bending moments due to
moderate load eccentricity, or for erection stability.
The plate is always attached to the column by means of fillet welds.
However, if the column carries only compression loads, direct bearing
may be assumed, provided that the contact surfaces are machined or can
be considered to be flat. No verification of the welds is then required.
Machining may be omitted if loads are relatively small.
Where there are moderate tension forces or no net tension the holding
down bolts are usually cast into the foundation (Figure 9). They anchor
the baseplate by bonding (Figure 9a), by bonding and bearing (Figure 9
b, c), or by bearing (Figure 9d).
When tensile forces are significant, it is necessary to provide
appropriate anchorage to the bolts. For example threaded bolts may be
used in conjunction with channel sections embedded in the concrete.
In tension connections the baseplate thickness is often dictated by
the bending moments produced by the holding down bolts. The bending
moments may require the use of stiffeners (Figures 8c and 8d). Such an
arrangement significantly increases the fabrication content and
therefore the cost of the column base as compared with the "simple"
case.
In high-rise buildings it may be convenient to combine the steel
structure resisting gravity loads with a concrete core resisting
horizontal forces.
Attaching the steel frame to a concrete core is mainly a practical
problem, since the two systems are built with dimensional tolerances of a
different order of magnitude. Special care should be taken to account
for the relative sequence of erection of the concrete and steel system,
the method of construction of the core (on which concrete tolerances
also depend), as well as the feasibility of compensating for
misalignments.
The connection should be able to transfer to the core vertical
forces, V, due to loads applied to the beam, and horizontal forces, H,
due to wind and frame geometrical imperfections (lack of verticality).
Some connection types are illustrated in Figure 10. It is important to
stress that the details in the concrete wall must be suitably designed
to disperse connection forces safely. In particular the details are
especially important when deep beams are required to transmit high
vertical loads.
The type shown in Figure 10a, with pockets in the wall, is convenient
for ease of adjustment, but complex in terms of core erection. Types
illustrated in Figures 10b to 10h where part of the connection is
encased in the core wall during concrete pouring, may be preferable.
The steel plate may be flush with the wall surface, as in types b-f,
or extended outwards as in types g and h. In the first case, which is
usually the more convenient because the steel plate can be supported on
the inside face of the formwork, a single web plate is welded on site to
which the steel beam is then attached. In the latter case the beam can
be connected directly to the encased plate. Reinforcing bars (rebars)
and/or headed studs can be used in order to transmit both components of
the beam action. Full penetration welds are preferred when the rebars
are connected directly to the flush plate (Figure 10d), so that
eccentricity of the force with respect to the weldment is avoided
(Figure10c).
Checking of the various components within the connection should be
conducted in a consistent manner, ensuring that the principles of
connection design, e.g. the assumed distribution of forces satisfies
equilibrium, are observed. As an illustration of this, consider the
structural requirements for the arrangement of Figure 10h. Assuming that
the shear transfer plane, ie. the "hinge" location of the simple
connection, is the mid-plane of the wall, then the set of headed studs
must resist only shear. Alternatively, if the "hinge" is assumed as the
wall face, then the studs should be designed to resist a combination of
shear and moment. This general requirement for a consistent approach to
modelling the force transfers is further explained in Figure 11, which
details the load transfer for the arrangement of case 10e. The shear
force V is assumed to be resisted entirely by the shear studs, whilst
the moment M is carried by a couple consisting of tension in the upper
rebars and compression transmitted by contact stresses between the
concrete and the steel plate. Whichever arrangement is adopted, however,
the main requirement is to ensure a proper dispersion of forces into
the core wall.
