30 NSC February 2008

Technical
SCI Deputy Director David Brown examines the important issues of frame imperfections,
and frame stability. As the opportunity to design to the Eurocodes draws nearer, subsequent
articles will look at other aspects of design.
Whats new?
Despite the immanent arrival of the Eurocodes
(National Annex for BS EN 1993-1-1 is expected
around March) the underlying physics has not
changed, and BS 5950 is a mature Standard. We
would not expect much to change, although it might
be presented in a different way. That is certainly the
case with issues of frame imperfections and frame
stability. The same principles are being applied, and
both Standards are trying to ensure that designers
allow for the inevitable imperfections in real frames,
and to make sure that second order effects are
accounted for, if they are signicant.
Frame imperfections.
In BS 5950, we have Notional Horizontal Forces
(NHF) that the Standard says are to allow for the
effects of practical imperfections such as lack of
verticality. They are 0.5% of the factored vertical
load, and applied horizontally at oor and roof
levels. Strangely, BS 5950 says that they need
only to be considered in what is commonly known
as the gravity loadcase, i.e. with dead and
imposed loads. BS 5950 says they should not to
be combined with applied horizontal loads. The
philosophy is interesting, as it could be seen to
imply that frames that are imperfect under vertical
load only, but somehow become perfect when the
wind blows on the structure.
According to BS EN 1993-1-1 the frame may
be modelled as out of plumb as shown in Figure
1. It could be modelled like this, but thankfully the
Standard also allows the lack of verticality to be
modelled by Equivalent horizontal forces (EHF),
in exactly the same way as BS 5950. The basic
EHF are presented as
1
/200 of the factored vertical
load (exactly the 0.5% in BS 5950), and there are
reduction factors available, depending on:
The height of the columns (it cant be that bad
for the full height of a tall column)
The numbers of columns (they cant all be so far
out of plumb)
The change in EC3 is that the EHF are applied in
ALL load combinations. This will be quite new to UK
designers, but appears to have a reasonable logic.
Frame imperfections no longer depend on the wind
blowing, or not.
One important implication is that the bracing
system will experience higher loads than UK
designers are used to. Not only is the maximum
combination factor for wind 1.5, but the bracing
will carry the EHF in addition to the wind. In both
Standards, the EHF are related to the factored
vertical loading, which obviously changes
depending on the combination being considered.
Under the Eurocode loading system, if wind is the
leading variable action, then the other variable
actions have a reduced combination factor. Reduced
vertical loads will result in smaller EHF.
Experience will quickly determine if the critical
load combination for bracing design is:
Maximum wind load, and reduced EHF, or
Reduced wind load and maximum EHF.
.
Additional design cases for bracing systems
In addition to the design cases including the EHF as
described above, the bracing must be checked for
two further design situations which are local to the
oor level:
Horizontal forces to foor diaphragms
Forces due to imperfections at splices
Frame design
to the Eurocode
Figure 1. Frame imperfections.
32 NSC February 2008
Technical
In both these design
situations, the bracing system
is checked locally (the storeys
above and below, as shown in
Figure 2) for the combination of
the force due to external loads
together with the forces due to
either of the above imperfections.
The EHF are not included in either
of these combinations and only
one imperfection needs to be
considered at a time.
The horizontal forces to be considered are the
accumulation of all the forces at the level being
considered, divided amongst the bracing systems.
The importance of restraints to compression
members (in this case columns) is covered in both
BS 5950 and EC3, and this force needs to be taken
to the bracing system. The second check, which
assumes a kink in the column at a splice position,
will be new to UK designers. In many orthodox
structures, the bracing will not need to be increased
as (a) the EHF are not included when carrying out
either check and (b) the additional force in the
bracing members is generally small.
Frame Stability
UK Designers will be of familiar ground here,
except with the presentation of the check for frame
stability. In the UK, we are used to checking frame
stability by calculating
=
h
cr
200
, where h is the
storey height, and is the horizontal storey
deection, under the notional horizontal forces
alone. A small value of
cr
indicates a sensitive
frame, where second-order effects must be allowed
for, usually by a simple amplier.
Under EC3, the equivalent check becomes
=
H
Ed
cr
V
Ed
h

H,Ed
( )( )
, where
H
Ed
is the design value of the horizontal reaction
at the bottom of the storey
V
Ed
is the total design vertical force on the
structure on the bottom of the storey

H,Ed
is the horizontal storey deection
h is the storey height
The signicant change in the Eurocode approach
is that the system of horizontal loads used to check
frame stability is not the EHF alone, but the actual
loads on the bracing in that combination. Typically,
the loads used when checking frame stability will
be the wind loads plus the EHF. (Remember that
there is no direct equivalent load combination to the
BS 5950 gravity loads only, as the wind is always
applied, even if it has a reduced combination factor.)
It would be easy to assume that the EC3
approach would give quite different results to
BS 5950, but this is not the case. For a vertical
bracing system the storey deection is obviously
liked to the level of lateral load, and thus as a
proportion
H
Ed

H,Ed
( )
is remaining about the same.
Looking at the EC3 equation, it could be expressed
as =
EHF
cr
V
Ed
h

H,Ed
( )( )
Wind
V
Ed
+
, which, without
using any of the reduction factors available for tall
or multiple columns, becomes
=
1
cr
200
h

H,Ed
( )( )
Wind
V
Ed
+
As expected, the outcome is close to the BS 5950
value.
How to allow for second order effects
If
cr
> 10, then second order effects can be ignored,
so familiar territory to UK designers. If
cr
< 10,
then a simple amplier can be used as long as
cr
is
larger than 3.
The amplier is
1
11/
cr
( )
, which with some
simple rearrangement, becomes

cr

cr
1 ( )
, again a
familiar friend.
Conclusions
There should be no difculty in understanding
the general principles of frame imperfections,
since these were covered in BS 5950. The logic of
having the EHF appearing in every combination will
perhaps appeal.
As far as second order effects are concerned, the
update to BS 5950 in 2000 has done us all a good
turn, by introducing us to frame stability. Having
become used to checking this to the UK standard,
it will be straightforward to apply our experience
to Eurocode design. The direction in the Eurocode
to use the actual loading conditions to check frame
stability, rather than a separate loading system (the
NHF alone, in BS 5950) is probably an advantage,
reducing the different analyses to be completed.
In truth, the more signicant challenge in both
frame imperfections and frame stability will be to
maintain a good grasp on the loading combinations,
and apply the correct loads (including the
appropriate EHF for that combination) to the
stability systems.
Figure 2. Assumed kink at splice level and
bracing to be checked.