Crane girders
The design of crane girders
Recent correspondence in Verulam¹ suggested that there were no decent examples of crane
girder design to the Eurocodes. David Brown of the SCI rises to the challenge…
The problem
According to the contribution in Verulam, a number of problems exist
with the design of a mono-symmetric member (a plate welded to the top
flange of a UB) and destabilising loads:
• BS 5950 examples have ‘mysteriously disappeared’ from the
equivalent Eurocode publications.
• The only way to design the member is to use ‘a piece of software
from a French website’.
• There is no way of checking the result (from the French software).
• Gantry girders would have to be doubly symmetric, or have the top
flange fully restrained.
What are the options?
Looking back at the BS 5950 examples in the SCI library, most are monosymmetric
plain plate welded to the top flange is presented in early editions of the
‘Red Book’2.
Some of the examples calculate the section properties of the
compound section – not a precise task, (especially before channels
had parallel flanges) and verify the fabricated member on that basis.
Alternative examples adopt the traditional and simpler approach of
assuming that the additional plate (or channel) carries the horizontal
loads, and the rolled section carries the vertical loads.
If one held the pessimistic expectation that the Eurocodes always
adopt the most complex approach, one might be pleasantly surprised to
find that the simple approach is allowed in clause 5.6.2(4) of EN 1993-6,
v
e
16 NSC
with a channel welded to the top flange. An example with a
e
h
Technical Digest 2019
which is the Standard covering
the design of crane supporting
structures. According to this clause,
lateral loads are resisted by the
top flange, and vertical loads are
resisted by the main beam under
the rail. This simple approach will
be familiar, and facilitates the use
of mono-symmetric sections.
Following this simple approach,
torsional moments are resisted by
a couple acting horizontally on
the top and bottom flange. As an
alternative, torsion may be treated
rigorously.
Lateral-torsional buckling
Gantry girders are unrestrained,
and have lateral loads applied
at the top flange level (or
above). As the beam buckles, the
vertical loads may be eccentric
to the shear centre, so there are
additional torsions on the section,
as indicated in Figure 1. Clause
6.3.2.1 of EN 1993-6 insists (quite
properly) that these torsions must
be accounted for. The designer
again has options, according to clause 6.3.2.3.
The first option is to simply consider the top flange and part of the
web acting entirely alone, and check it as a simple strut. Safe, certainly,
but conservative. The second option is to assess the member for the
combined effects of lateral-torsional buckling, minor axis moment and
torsion, using the interaction expression presented in Annex A of the
Standard. The UK National Annex endorses the use of this alternative.
Of course, the interaction expression looks complicated:
My,Ed
LTMy,Rk/M1
CMzMz,Ed
Mz,Rk/M1
+ 1
+
kwkzwkBEd
BRk/M1
A numerical worked example would help, as the correspondence in
Verulam notes. Fortunately there is a full worked example in P3853, which
is SCI’s publication on the design of steel beams in torsion. Example 2
is precisely the case under consideration – a gantry girder, except the
selected member is a UB with no plate. Because this comprehensive
numerical example exists, no further attention is paid to the interaction
expression in this article.
Destabilising loads
Loads that move with the buckling compression flange are classed as
destabilising. As the correspondence in Verulam indicates, one would
normally assume that gantry girders are subject to destabilising loads.
EN 1993-6 offers an interesting twist (no pun intended) to the
classification of destabilising loads. Clause 6.3.2.2 suggests that if the
crane rail is fixed directly to the runway beam, the applied vertical load
can be considered as stabilising. This unexpected conclusion is because,
as shown in Figure 2, as the runway beam starts to twist, the application
of load moves to the ‘high’ side of the rail, which is actually on the
‘restoring’ side of the shear centre. Thus the load is stabilising and in
these circumstances the Standard notes that it may be assumed that the
loads are applied at the shear centre.
If the rail is supported on a flexible elastomeric pad, the loads are
destabilising and the Standard notes that the loads should be assumed to
be applied at the top of the flange.
In BS 5950, destabilising loads were treated by multiplying the system
Figure 1 Torsions on a gantry girder
No elastomeric bearing
Stabilising eect as
the beam buckles
Elastomeric bearing
Destabilising load
Figure 2 Influence of crane rail on load classification