Global Behavior

The overall responses during Test A are illustrated in Figure 4-18. Up to 60 kips of

loading the bridge specimen showed a linearly elastic response with an initial stiffness of

approximately 90 kips/in. As the load increased from 80 kips to 140 kips, some cracks in

the deck were noticed and at the same time, the stiffness of the specimen was decreasing

visibly. Twisting of the girders was observed as well. At the load of 140 kips, the capacity

of the bridge slightly decreased by 1 kip because one bolt in the bottom connection of the

cross-frame near mid-span was sheared. The intact girder started showing uplift at the

cantilever support. The specimen reached its maximum capacity at 156 kips with 2.5 in. of

displacement.

After reaching the maximum capacity of 156 kips, the concrete deck was crushed

by the loading pad and the load dropped to 133 kips. The test continued and the specimen

was still able to sustain some additional load with many ups and downs until it failed at 5.5

in. of displacement. These ups and downs in this loading period corresponded to several

damages observed, including: 1) the visible cracking and crushing in concrete propagating

toward the ends of specimen, 2) the bottom of the concrete deck between the two girders

both ends of intact girder. As seen in Figure 4-18, during this loading period, the load

fluctuated and reached a local maximum load of 144 kips at 4-in. displacement. It, then,

slowly dropped down to 123 kips at 5.5-in. displacement before the test was halted. This

sequence of failure suggests that the bridge was trying to transfer the load to the intact

girder through an alternative load path after the primary load path had failed. This load-

transfer mechanism is discussed further in the following section.

Figure 4-18 Load vs. deflection curve of the specimen during ultimate load Test A.

A 1 ft by 1 ft grid was marked on the concrete surface to map the damage area more

accurately as illustrated in Figure 4-19. In general, the major damage was observed in the

deck over an area of 5 ft long by 3 ft wide at where the load was applied. Cracking started

along the inside top flange of the damaged girder, then it propagated toward the intact

girder. The cracking pattern was approximated by the black lines in Figure 4-19. This

cracking pattern indicates that the deck failed predominantly in one-way shear failure

mode. The deck damage showed that the damage propagated toward the north support more First Crack

CF Failed

Deck Crushed

Uplift CFs Failed

than the south support. It could be due to the presence of the cantilever portion in the south

end. This cantilever made the south portion of the specimen measured from the loading

point was stiffer than the north portion of the specimen. Eventually, the deck was punched

by the loading pad along the inside top flange of the damaged girder when the specimen

reached its maximum capacity. The spalling of concrete in the bottom of the deck is

illustrated in Figure 4-20.

Figure 4-20 Punching in the bottom of the deck.

The concrete deck was also damaged at both ends of the specimen. The mechanism

for the observed failure can be explained by considering the hypothetical situation in which

the deck is cut longitudinally between the girders leaving two separate and independent

structures. The east structure is then damaged and the load point is displaced downward

the same magnitude as produced in the actual test. Obviously, the magnitude of load

causing this displacement will be much less than what observed in test. Due to the low

flexural capacity at the damage location, the displaced shape will essentially be a

mechanism with hinge rotation about the load point and linear segments to the supports. It

is important to note how the ends of the girder will now project above the supports and that

the undamaged girder has no load applied and is therefore straight and level. Finally,

keeping the load point of the damaged girder at the fixed level of displacement, consider

the forces required to re-join the two separate structures and fuse the deck back together.

At the ends of the girder, the (hypothetical) damaged girder will be above the undamaged

one and need to be pulled downward, which then imparts an equal and opposite upward

Figure 4-21 where the deck above the undamaged girder appears to have been pulled

upwards and seemingly ripped off at the ends.

This same hypothetical situation can be used to examine the failure mechanism near

the load point. In this region, the (hypothetical) damaged girder will be below the

undamaged. Since the displacement level is being held by the loading beam, the

undamaged girder will be pulled downward when the two separate structures are

hypothetically re-joined. The effect of the resulting shear force can be seen in Figure 4-20.

Figure 4-21 Damage at North end (left) and South end (right).

Another important behavior observed was the uplift at supports of the intact girder.

Displacements near the supports were monitored at location 1 (14 in. from the bearing line

of the South support) and location 2 (17 in. from the north end of steel box girders) as

depicted in Figure 4-22. Figure 4-23 and Figure 4-24 plot the vertical displacements at

these two locations. These plots show that the intact girder (or the West Girder) began to

uplift at the south support shortly at 100 kips before the specimen reached its maximum

capacity of 156 kips while the north support was uplifted just right after the maximum Shear stress pulling up Shear stress pulling down Shear stress pulling up

capacity was reached. The final uplifts were measured as 0.48 in. near the south support

and 0.11 in. near the north support.

Figure 4-22 Location of potentiometers at supports.

Figure 4-24 Displacement of the both girders near the north support.

(a) North support (b) South support