Redundancy analysis - Computational Analysis and Design of Bridge Structures - page 460

Table 15.1 Typ es of redundancy
Type

Description

Load path
redundancy

A member is considered load path redundant if an alternative and sufficient
load path is determined to exist. Load path redundancy is the type of
redundancy that designers consider when they count parallel girders or
load paths. However, merely determining that alternate load paths exist is
not enough. The alternative load paths must have sufficient capacity to
carry the load redistributed to them from an adjacent failed member. If the
additional redistributed load fails, progressive failure of the alternative load
path occurs, and the members could in fact be fracture critical. In
determining the sufficiency of alternative load paths, all elements present
(primary and secondary members) should be considered.

Structural
redundancy

A member is considered structurally redundant if its boundary conditions
or supports are such that the failure of the member merely changes the
boundary or support conditions but does not result in the collapse of
the superstructure. Again, the member with modified support
conditions must be sufficient to carry loads in its new configuration. For
example, the failure of the negative-moment region of a two-span
continuous girder is not critical to the survival of the superstructure if
the positive-moment region is sufficient to carry the load as a simply
supported girder.

Internal
redundancy

A member is considered internally redundant if an alternative and
sufficient load path exists within the member itself such as the multiple
plies of riveted steel member.

Source: FHWA/NSBA/HDR, “Steel Bridge Design Handbook FHWA-IF-12-052—Vol. 9: Redundancy,”
Federal Highway Administration, USDOT, November 2012, http://www.fhwa.dot.gov/bridge/steel/
pubs/if12052/volume09.pdf."

Moreover, Caltrans made a comprehensive flowchart for identifying FCMs
of complex steel bridges in Figure 15.1.
The definition of a narrow plate girder (PG) system varies slightly from
that used in stability discussions when focusing on redundancy. Whereas
the system could contain any number of closely spaced girders in stabil-
ity discussions, twin girder systems alone constitute a narrow system in
the context of redundancy. This is due to the fact that only two primary
elements exist to transfer load. If one of these fails, the second would be
unable to support the entire weight of the structure, resulting in collapse.
Other elements of the bridge, particularly the deck, could be able to carry
additional loads encountered due to a nonredundant member failure and
prevent collapse, which has been seen in the past. This built-in redundancy
is difficult to predict, however, and is not explicitly recognized in the design.
As such, for typical PG bridges, a minimum of three girders are required to
provide alternate load paths and be considered system redundant.
To a lesser degree, studies of concrete bridge redundancy were made assum-
ing cracking concrete, yield reinforcement, or reaching ultimate moment and
shear capacity of the longitudinal or transverse beams (Imhof et al. 2004).