Figure 21: Envelopes of load-displacement curves Figure 18: Compression failure of test unit 1001NT Figure 17: Test setup Figure 15: Cross Section of test units Figure 19: Load displacement response of units with internally bonded tendons Figure 16: Elevation of a typical test unit and tendon layout Figure 20: Load displacement response of units with different ratios of internal to external post-tensioning

The second phase of the seismic research project conducted at the University of California San Diego (UCSD) on precast segmental bridges has been completed. This research project was initiated by the American Segmental Bridge Institute (ASBI) and funded by Caltrans. The objective of the first two phases was to study the seismic performance of segment-to-segment joints in superstructures with different ratios of internal to external post-tensioning. Phase I focused on joints subjected to high bending moments combined with low shears. Phase II focused on joints subjected to high bending moments combined with high shears, representing superstructure joints close to the columns. Seismic performance of superstructure-column systems will be investigated in the third phase of the project. As in Phase I, the test variables in the Phase II test units were: (1) the ratio of internal to external post-tensioning of the superstructure, and (2) presence of continuous mild steel reinforcement in the deck across the segment-to-segment joints. Four large-scale units were tested and the test matrix is given in Table 1 (see page 4). The test units were constructed at a 2/3-scale with respect to a prototype box girder superstructure. Only one half of the prototype box section in the shape of an equivalent I-section was modeled in the experiments to simplify the test setup. Cross section of the test units is shown in Fig. 15. Figure 16 shows an elevation view of a typical test unit. Segments 1 and 2 (see Fig. 16) were epoxy-bonded. In each of the Phase II test units, only one segment-to-segment joint was modeled, which was the epoxy-bonded joint between Segments 1 and 2 (see Fig. 16). The main objective of the experiments was to study the performance of this joint under the combined effect of high bending moments and high shears.

The prototype structure used in design of the test units had a harped-shape tendon with harping points at 1/3 locations of each span, in addition to a horizontal continuity tendon near the bottom surface of the box girder. The harped-shape tendon was modeled in the experiments by the inclined tendon shown in Fig. 16. The horizontal continuity tendon was also modeled in the experiments as shown in Fig. 16. In Test Units 100INT and 100INTCIP, 100 percent of the post-tensioning was achieved by internally bonded tendons. Test Unit 100EXT was post-tensioned with external tendons only (see Fig. 15). In Unit 50INT/50EXT, the horizontal continuity tendon was internally bonded, whereas half of the harped-shape prestressing steel consisted of external tendons and the other half consisted of an internally bonded tendon (see Fig. 15). As in Phase I, Unit 100INTCIP of Phase II had a cast-in-place deck closure at location of the joint between Segments 1 and 2 (Fig. 16), whereas the remaining portions of the precast segments, along the web and bottom slab, were epoxy bonded. Each test unit was subject to shearing force, V, and bending moment, M, at its end (see Fig. 16). The shearing force and bending moment were applied by means of two vertical hydraulic actuators, which were connected to a steel beam (steel nose) as shown in Fig. 17. In the initial testing stage, each test unit was loaded to the reference load level. The reference load is the one required to obtain the correct dead load stresses at the segment-to-segment joint. In the second testing stage, fully reversed cyclic vertical displacements were applied at the tip of the steel nose until failure of the test unit. The forces in the two actuators were related to each other by a prescribed function to obtain the correct simultaneous values of bending moment and shearing force at the segment-to-segment joint throughout the test. To determine the prescribed function that related V to M, it was assumed that the prototype structure was subjected to gravity loads combined with longitudinal seismic forces.

The first crack occurred in all test units at the segment-to-segment joint due to flexure in the downward loading direction. A similar flexural crack occurred in the bottom slab at the joint location under upward loading. Because of the continuity of the deck, the first flexural crack in the deck of Unit 100INTCIP occurred under downward loading at the construction joint between the cast-in-place deck closure joint and the precast segment. The mild steel reinforcement across the segment-to-segment joint controlled the widths of cracks. All test units failed under downward loading by compression in the bottom slab. The test units were subjected to negative bending moments and the failure mode was governed by the compressive force capacity of the bottom slab. Figure 18 shows the onset of compression failure of Unit 100INT under downward loading. Despite the high shearing force transferred at the segment-to-segment joint, no vertical slip was observed between the adjacent precast segments in all test units. The vertical slip was observed only after compression failure of the bottom slab. Figure 19 shows the history of total applied load versus vertical displacement measured at the tip of the steel nose for Test Units 100INT and 100INTCIP; the variable of these two test units was presence of mild steel reinforcement in the deck across the segment-to-segment joint. The sign convention in Fig. 19 is positive for downward loading and displacement. The yield strength of the deck reinforcing bars crossing the joint in Unit 100INTCIP could be developed. Thus the maximum total downward load of Unit 100INTCIP was 141 kips, compared to a maximum total load of 94 kips in Unit 100INT (see Fig. 19). Figure 20 is similar to Fig. 19 but it shows the load versus displacement for Units 100INT, 100EXT and 50INT/50EXT with different ratios of internal to external post-tensioning.

Based on the experimental ultimate moment capacity of Unit 100INT, compression failure in the bottom slab of the prototype structure occurs at 1.92 g seismic vertical acceleration (assuming no longitudinal seismic forces). With cast-in-place deck closure joints (Unit 100INTCIP), failure of the prototype structure occurs at 3.91 g seismic vertical acceleration. With 100 percent external post-tensioning, failure of the prototype structure occurs at 1.98 g vertical acceleration. Compression failure of the superstructure bottom slab near the columns can be avoided by increase of slab thickness, or by provision of closed stirrups to confine the bottom slab and increase their compressive capacity. Figure 21 shows the envelopes of the hysteresis loops of Figs. 19 and 20. Figure 21 confirms findings of Phase I that the ductility and displacement capacity can be significantly increased with 100 percent external post-tensioning. Values of maximum downward displacement reached before failure of Units 100INT, 100INTCIP and 50INT/50EXT were comparable. However, the maximum displacement reached before failure of Unit 50INT/50EXT under upward loading was significantly less than the maximum upward displacements for the other test units (see Fig. 21). Values of the maximum downward load carrying capacity, Vu, maximum downward displacement before failure, Du, and the permanent residual displacement, Dr, are given in Table 1 for all test units. The residual displacement, Dr, is measured at the reference load level during unloading of the test units after reaching a maximum downward displacement of 4.5 in. The Dr values given in Table 1 indicate that residual displacements can be minimized by use of 100 percent external post-tensioning, which agrees with the findings of Phase I.

Table 1. Test Matrix and Limited Test Results
Test Unit	Test Description	Vu′ (kip)	Δu′ (in.)	Δr (in.)
100INT	100% Internal Post-Tensioning (PT)	94.2	6.0	1.53
100INTCIP	100% Internal PT + Cast-in-Place Deck Closure Joint	141.3	6.4	1.99
100EXT	100% External PT	95.6	12.9	0.17
50INT/50EXT	50% Internal PT + 50% External PT	96.0	7.2	0.39

Δr was measured during the 4.5 in. displacement cycle.