Examples of Finite Element Models of Steel-Concrete Composite Bridges - Finite Element Analysis and Design of Steel and Steel-Concrete Composite Bridges - page 604

translations. Sharing nodes with two different types of finite elements may
lead to perturbation of element stress and strain at the interface boundary.
Rigid links throughMPCs connect different elements within the structural
model and enforce the kinematics relationships between the degrees of
freedom at each node.
The study [ 7.4 ] outlined that in the case where shell elements were used
to model the bridge deck, the nodes of the girder did not coincide with the
nodes of the shell elements in the deck. The shell elements in the bridge
deck were connected with the prescribed girder models through an MPC.
Typical bending elements, such as the shell element (for the deck and for
girder models G1 and G3) and the beam element (for the girder model G2
and G4), should be avoided for the modeling of the composite girder
bridge since displacement incompatibility occurs at the interface of two
bending elements [ 7.5 ]. This incompatibility was noticeable since the axial
displacements of the deck and the girder. The incompatibility error disap-
peared as the mesh was refined and many methods were proposed to elim-
inate this nonconforming error [ 7.6-7.8 ] . In ABAQUS [1.29], the use of
S8R elements for the concrete deck and B32 elements for the girder pro-
vided full compatibility between the boundaries of two different elements.
The applied loading on a bridge deck consisted of pressure loads applied
through a tire patch. In the finite element modeling, this requirement
imposed the need for a fine mesh in the deck, so that the element is fitted
with the patch size. As a part of the research [ 7.4 ] , the equivalent nodal load
algorithm was employed in order to uncouple the patch load from
the mesh size. Also, bearings were modeled by assigning boundary condi-
tions to the zero-dimensional elements at their real location. For simply
supported beams, rotations in all directions were allowed in order to sim-
ulate the simply supported structure. Minimum restraints were assigned for
longitudinal and transverse movement, while vertical restraint was placed
at the supports. Kinematic constraints were also supplied to nodes between
the girders and the deck. Since the main purpose of the study [ 7.4 ] was
to analyze bridge superstructures, it was assumed that substructures,
such as piers and abutments, did not influence the behavior of the
superstructure.
The use of ductile end cross frames to reduce the transverse seismic
demand in composite steel plate girder bridge superstructures was investi-
gated by Carden et al. [ 7.9 ] . The authors showed that the effectiveness of
these cross frames was strongly influenced by the transverse flexibility of
the superstructure and its capacity for potentially large relative transverse