Вы находитесь на странице: 1из 3

Chris Hendy

4/1/07
Transverse local and global moments in beam and slab decks

Local and global moments transversely in deck slabs


forming part of beam and slab bridges
1. Introduction
The results obtained from the computer analysis of a structure are only as good as the
idealisation of the structure allows. Conventional grillages are not particularly good
for modelling transverse effects in beam and slab decks because there is usually
insufficient detail in the grillage mesh representing the slab between beams. This has
led to the traditional approach of calculating global and local effects separately. To
calculate global effects from a grillage model, loads must only be applied to the nodes
and there must not be nodes on the transverse members between the main longitudinal
members. Local effects are calculated from a separate analysis of the slab, usually
using Pucher charts. If this approach is not followed, it is possible to be either
excessively conservative or un-conservative in the design of the slab. Even when this
approach is followed, caution is still required when using autoloaders and INSURF.
The reasons are discussed below.
This note has been prepared in response to a number of queries and misconceptions
encountered recently.

2. Modelling by grillage analysis


2.1 Correct approach to separate determination of local and global effects by
grillage analysis
Figure 1(i) shows the correct modelling and load application for separate calculation
of global effects. If this approach is used, the effects of local loading on the slab
between the main beams must be determined by other means (e.g. Pucher charts) and
added to the global effects. Conservative assumptions are usually made for boundary
conditions in the local analysis i.e. fixed longitudinal edges when calculating hogging
moments and free edges when calculating sagging moments. (It is possible to be less
conservative by always using fixed end conditions to determine the local effects and
then applying the fixed end moments derived from the local analysis to the grillage
transverse members in the opposite sense, as release moments. These additional
moments need to be added to the global moments and local moments derived with
fixed edge conditions).
It should be noted that if an influence surface is generated for the slab transversely in
this model and this is used to determine moments by integration of the area beneath
the surface, the problems in 2.2 below of loading the slab between main beams will
effectively be created. This is the case with INSURF, which should not be used with
this model type to determine transverse moments in slabs, including even global
moments. To determine global moments, loads should be applied manually to the
nodes, using the influence surface as a guide to placement. This problem does not
affect autoloaders which apply point loads at the nodes only, but care should be taken
not to distribute loads to members where the software gives this option. HLOADS
and Autoloader for SuperSTRESS give this choice.

1 of 3

Chris Hendy

4/1/07
Transverse local and global moments in beam and slab decks

2.2 Incorrect approach to separate determination of local and global effects by


grillage analysis
Figure 1(ii) shows inappropriate modelling and load application for separate
calculation of global effects. It is inappropriate because some of the load has been
applied to the slab between the beams, thus causing some additional local effects.
While this may seem desirable to circumvent the need for separate local and global
analyses, the model does not have anywhere near enough detail in the mesh between
beams to make the results accurate. If a single analysis is required, either the number
of slab transverse and longitudinal elements between beams could be increased or a
shell finite element analysis undertaken.
Whilst this model cannot give sensible calculation of the combined effect of local and
global effects, it also cannot be used conservatively to calculate the effects
separately. The analysis in Figure 1(ii) can be seen to produce a hogging moment in
the slab over the internal beams, whilst the analysis in Figure 1(i) produces a sagging
moment. If the total hogging moment in the slab over the beams was obtained by
adding the local effects from a separate analysis to this moment, the net hogging
moment would be considerably overestimated. Similarly, the sagging moment
between the main beams would be underestimated.
Neither INSURF nor other autoloaders can be used with this model to determine slab
transverse moments for the reasons above.
2.3 Considerations when using INSURF
INSURF is really only intended for determining global effects in longitudinal
members and bearing reactions and rotations etc. For transverse members in beam
and slab decks, the problems discussed in 2.1 above apply. In addition, INSURF does
not consider the two HB vehicle case in BD37 clause 6.4.3.1 which needs to be
considered for two way spanning slabs. Accidental wheel loading is also not
considered.

3. Separate consideration of local and global effects at ULS in


design
Clause 4.8.3b) of BS 5400 permits local and global effects in deck slabs to be
considered separately in design at ULS if certain criteria are met. The local and
global effects referred to in this clause are for the longitudinal direction, not for the
transverse direction being discussed above. Local and global effects may be
considered separately in the longitudinal direction, provided the combination is
checked at SLS, because:
the local effect normally affects a small proportion of the total width of the
slab and thus redistribution can occur across the width of the slab;
simple addition of flexural effects ignores membrane action in the slab;
the true local collapse involves a mechanism of yield lines at much greater
load factor than suggested by elastic analysis.
Written by: Chris Hendy
4/1/07

Reviewed by: Bob Bellamy


5/1/07
2 of 3

Chris Hendy

4/1/07
Transverse local and global moments in beam and slab decks

(i) Correct modelling and load application for separate calculation of global
effects

(ii) Inappropriate modelling and load application for separate calculation of


global effects
FIGURE 1

3 of 3

Вам также может понравиться