Civil Engineering Reference
In-Depth Information
e.
Inertial forces
analysis) are carried out, either due to the level of risk (i.e. in
areas of high seismicity) or due to the judgement being taken
that the static method does not reasonably reflect the behav-
iour of the structure (for example where the structure is highly
irregular and hence the assumption of a principal translational
mode of behaviour is not applicable).
Self-straining load effects are typically modelled as either
'equivalent temperature' effects (e.g. for shrinkage the strains
would be calculated and then converted to an 'equivalent' tem-
perature change) or support movements, and analysed using
an elastic computer programme - it should be noted that these
sorts of load effects are very dependent upon the details of the
analysis and more prone to give conservative results (particu-
larly around boundaries where infinitely stiff translational or
rotational restraints have been modelled).
Blast loads must be modelled accounting for the dynamic
effects. This is typically considered using dynamic 'single
degree of freedom' methods. It is often possible to derive sim-
ple equivalent static UDLs (Cormie, 2009), or it can sometimes
be more efficient to design the structure via an assessment of
the local effects on the remaining structure (i.e. key element
removal). More sophisticated dynamic time history analysis
using nonlinear finite element analysis is sometimes carried
out where simpler methods are not suitable.
Fire loads are typically modelled by varying the combin-
ation factors, rather than by changing the actual magnitude or
distribution of the gravity effects.
Fluid loads are typically modelled as variable magnitude
pressures (UDLs), based upon the height of retained fluid and
its density.
Silo loads are modelled in a similar fashion to fluid loads,
but with generally with more complex variation (to reflect the
'non-fluid' behaviour of a 'grain', for example, as well as the
influence of say 'emptying' a silo.
Soil/earth loads are typically applied as constant UDLs at
the underside of a structure and linearly varying lateral loads
on the sides (but with surcharge loads being applied as uniform
loads).
i.
Dynamic wind loads
ii.
Seismic loads
iii.
Blast loading
Time variable loads can be grouped as:
a.
Load effects expected during typical 'service' - general
live loads say
b.
Rare event loads (extreme events, such as floods, say)
c.
Extremely rare loads, i.e. worst case loads
Conceptually loads can be grouped based upon their time
dependency;
a.
Long-term (permanent) loads - expected to remain static
for expected life of structure
b.
Short-term, fixed magnitude events
c.
Constant or random variability fluctuating forces
d.
Dynamic loads (impact, seismic, blast loading)
For load effects which are variable it is important to consider
the impact of a variable distribution - the simplest example of
this is for a multiple span beam, whether the loading case of
maximum live load on one span with zero live load on adjacent
spans as well as the inverse to capture the peak moments at the
mid-span and supports, as well as potentially a point load adja-
cent to a support and a uniformly distributed load. For more
complex structures (say a 3D canopy structure) the impact of
spatial distribution of time-dependent loading is more difficult
to capture, although analysis methods have been developed
that give less conservative loads than would otherwise need
to be used.
Generally dead and live loads will be applied as uniformly
distributed loads (UDLs) and point loads, depending upon
whether the idealisation of the structure is two- or three-
dimensional the loads will be either area, line or point load
effects.
Wind loads are often applied as UDLs on the surfaces of a
building, or by application of point loads at the centre of pro-
jected area - however, it should be considered that for some
cases (tall buildings for example) the dynamic component of
wind loads (i.e. that portion due to the movement of the struc-
ture under the wind) is significant and those effects are dis-
tributed around the centre of mass, rather than the centre of
projected area.
Seismic loads are simplistically applied as equivalent static
loads, which are typically taken to be a first mode (i.e. half
curve) response which gives an 'inverted triangle' of forces
typically applied at the centre of mass (or more accurately
applied eccentrically around the centre of mass in order to
generate a level of additional torsion to reflect the variability
of mass distribution). More sophisticated dynamic analysis
procedures (e.g. response spectrum analysis or time history
10.3 Combinations of load
The Eurocodes and the various US codes follow roughly simi-
lar approaches of using factored load effects for the ultimate
(or strength in US terminology) limit state (ULS) and unfac-
tored loads at the serviceability limit state (SLS).
All combinations are related to the likelihood of combina-
tions occurring; it should be self-evident that applying simul-
taneously wind, live, temperature and seismic loads effects,
whilst logically sustainable to variable relative levels, is not a
practical approach.
Engineering practice is therefore to define a number of typi-
cal combinations, such as:
Dead + Live
Dead + Live + Wind
