Anda di halaman 1dari 3


AD 113 Collision of vehicles with bridge


The DTp has issued a proposed amendment to BD 37/88 concerning the application of loads
resulting from the collision of a vehicle with the superstructure. Advice on the application has
also been issued and further comment has been given on live loading to be applied to the
damaged bridge. The following comments are given to assist designers in the interpretation
of the requirements and the application of them in the design of the superstructure.

The requirements of the Department of Transport

Our understanding of the Department’s requirements are that to check the adequacy of the
superstructure in response to a vehicle collision, two separate stages are to be considered.

Stage 1. At the moment of impact. A check is to be made at ULS only, using the impact
loads specified in the amendment and partial factors γfL appropriate to load combination 4. No
live load is to be included in this check. In the check, local damage is to be ignored. It is to
be assumed that full transfer of global forces from the point of impact takes place. An
adequate loadpath should therefore exist for these global load effects. If it is found that the
bearings or other support systems are adequate but that elements of the structure are
inadequate in global action (for example, an unbraced bottom flange is not given sufficient
restraint against the lateral force by the bending of a thin web), then it may alternatively be
assumed that the elements concerned are rendered ineffective by the impact. The stage 2
check must then ignore the ineffective elements in global action.

Stage 2. Immediately after the impact. Immediately after the event, the bridge has to be able
to stand up while still carrying traffic which may be crossing. Since the check is one of
survival and the likely traffic is of every-day intensity, it should be carried out at ULS only but
partial load factors normally appropriate to SLS, combination 1 should be used. The partial
factors γm and γf3 should take their usual ULS values. It is understood that HA loading and/or
30 units of HB loading are to be applied for bridges carrying public highways. For this check,
the designer must judge what local damage (including loss of parts of elements) might
reasonably have occurred. This is the area where a designer may feel somewhat hesitant at
first, but it is also the area where the Department considers that the designer is the right
person to judge – arbitrary general rules are not appropriate.

The extent of impact damage

The client authority will be interested in the level of damage which might result from an
impact, on account of considerations of potential disruption due to repair or replacement.
Although local damage is inevitable, damage may only be acceptable up to a prescribed limit,
and the structure will have to be designed accordingly. As perhaps an extreme example, it
might be possible to design the bridge to stand up in stage 2 after one beam has been
grossly damaged and rendered ineffective, but this could be unacceptable because repairs
would be very extensive, costly and disruptive.

From the results of some collision incidents, it is possible to take a view of the likely level of
damage to certain types of bridge, and the following comments offer some guidance on the
consideration of local damage resulting from impact.

Composite girder and slab bridge

For a conventional girder-and-slab bridge with intermediate transverse bracing, impact on a

bottom flange is likely to cause plastic deformation and possibly a small amount of local
tearing of the flange. The flange may be torn locally from the web. If impact occurs at a
‘hard point’ (at transverse cross-bracing) there may be slightly more local deformation than at
softer positions (between bracing).

Design against impact for such a bridge could therefore presume that in stage 1 the specified
impact force is carried as follows.

(a) Horizontal force is spread from the point of impact along the length of the girder, by
bending and shear in the plane of the bottom flange, to points of lateral restraint, i.e.
transverse bracing. From these positions it is transferred through the bracing
members, through the top flanges of the girders and into the plane of the deck slab.
At the supports the force is carried down to the bearings through the support bracing.

(b) Vertical force is applied upward on the line of the web and transferred by global
bending back to the supports.

(c) Inclined forces are simply resolved into components of horizontal and vertical force.
Local effects from an inclined force on the tip of a flange cause only local damage,
and this need not be checked.

For survival in stage 2, the effect of the damage would be as follows. As a tension element
the flange is likely still to be quite effective. We consider that it would be reasonable for a
designer to make only a small allowance for loss of effective section. However, as a
compression element, i.e. in the region close to an intermediate support, the local damage
may be sufficient to initiate large deflection local buckling, particularly if the flange is torn
from the web locally. We would consider it prudent to presume the creation of a pin joint in
the beam which has been struck and carry out a global analysis accordingly. The shear
capacity of the web at that point should be considered carefully, presuming an ineffective
flange and possibly a small reduction of web area, but it is likely that the web would still be
adequate in that condition.

Provided that the design of the bracing and its attachment are adequate for the stage 1
check, there should be no significant damage to those members.

Box girder bridge

For a box girder bridge, local deformation of the web-flange junction is likely, possibly with
minor local tearing. If the impact is at or very close to an internal diaphragm or cross-frame,
some internal damage will also result.

In stage 1 the forces would be transferred by distortional behaviour back to diaphragm or

cross-frame positions, and then by torsion and bending back to supports.

Under stage 2 in midspan regions there should be little reduction in ultimate moment
capacity, as for the beam-and-slab bridge; torsional capacity is also likely to be largely
retained. Adjacent to supports the deformation of the web-flange junction will lead to some
loss of moment capacity, but it is likely that the other lower corner will continue to provide
some bending strength. The designer will have to judge, depending on proportions and plate
thickness, what capacity might remain.

Half-through bridge

The deck of a half-through bridge will provide continuous and direct restraint to the bottom
flange against impact forces. Some tearing of the bottom flange might occur. If the
connection to cross-beams which provide U-frame restraint can be damaged by the impact,
stage 2 should consider the structural action without that restraint at one cross-beam. As for
girder and slab bridges, the effective area of the tension flange should be reduced

Alternative situation where global damage is accepted in stage 1

As mentioned under stage 1, if there is not a complete and adequate load path for the impact
forces on a beam, ‘global damage’ may be presumed and the stage 2 check should allow for
the elimination of the elements concerned. The assessment of such global damage may be
even more difficult than that of local damage. We suggest that it is simply assumed that the
beam concerned has been effectively eliminated from any global action over a suitable length.
The remainder of the structure must then be adequate to carry the live loading. This
alternative scenario would only be appropriate where the client is willing to accept the much
greater extent of damage.