considered in design:
r one-way flexural failure of the barrier wall at the barrier
wall–deck slab joint
r punching shear or pull-out failure of the anchor rods
embedded in the deck slab
r fracture of the tensioned rod
r two-way slab failure similar to AASHTO LRFD specifications’
yield-line failure patterns
Accordingly, models M1 to M5 (Fig. 3) were erected.
Model M1 was intended to examine the flexural capacity
of the connection between the precast concrete
barrier wall and the concrete deck slab as well as the
punching shear capacity of the threaded rod–steel plate
connection embedded in the concrete deck slab. Also,
this model would examine the flexural capacity of the
one-way action of the joint between the two inclined
surfaces of the inner face of the barrier wall. Figure 4
presents a cross section of the proposed barrier wall system
and shows the arrangement of the reinforcing steel
in the barrier wall. The barrier was 1200 mm (47 in.) long
to accommodate two threaded rods with
600 mm (24 in.) spacing.
Model M2 was a control specimen representing the
monolithic cast-in-place reinforced concrete barrier wall
used in Ontario bridges for performance level 3 (PL-3).
The amount of wall reinforcement, both vertically and
horizontally, was identical to that used in model M1.
However, the vertical reinforcement in model M2 was
extended into the deck slab to provide concrete anchorage.
Figure 5 presents the details of the reinforcement
in the barrier wall and deck slab for model M2. Both
models M1 and M2 were loaded laterally at 990 mm (39 in.)
above the top surface of the deck slab using a line load
length of 1200 mm (47 in.) (Fig. 3).
Model M3 was intended to examine the overall performance
of the barrier wall segment when connected to
the cantilever part of the concrete deck slab, simulating
the case of slab-on-girder bridges. Figure 6 shows
model M3 with 900 mm (35 in.) projecting deck slab
cantilever length. It also shows the layout of the reinforcing
bars and the threaded rods over the 3000 mm
(118 in.) long barrier segment. This barrier model was
loaded over a 200 mm (8 in.) square area centered at the
midlength of the barrier and 990 mm (39 in.) above the
top surface of the deck slab (Fig. 3).
Model M4 was identical to model M3, but without the
cantilever slab. Model M4 was intended to examine
the overall performance of the barrier wall segment
with respect to yield-line pattern when the barrier wall