Full-Depth Precast Concrete Bridge Deck Panel Systems (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

17 This chapter presents the results and findings of Tasks 3 through 6 listed in Chapter 1. Additional information is pro- vided in Appendices B through F. The following issues are discussed in this chapter: • Recommendations for full-depth precast concrete bridge deck panel systems. Two systems were developed, a trans- versely pretensioned system and a transversely convention- ally reinforced system. Neither system uses longitudinal posttensioning or an overlay. New panel-to-panel and panel-to-girder connection details were developed and used in these systems, as follows: – Panel to panel connection. Four connection details were initially developed and tested in direct tension. Based on the structural performance of these details, two details were selected and used in the recommended systems. – Panel to concrete girder connection. A new connection detail that uses clusters of three 11⁄4 in. (31.8 mm) diameter double-headed steel studs was developed. The clusters are spaced at 48 in. (1220 mm). – Panel to steel girder connection. A new connection detail that uses clusters of eight 11⁄4 in. (31.8 mm) diameter steel studs was developed. The clusters are spaced at 48 in. (1220 mm). • Analytical and experimental investigation of selected details: – Panel to panel connection  Direct tension test using pullout specimens  Full-scale bridge specimen – Precast panel to concrete girder connection: full-scale direct tension test – Precast panel to steel girder connection  Push-off (direct shear) specimens tested for ultimate capacity  Push-off (direct shear) specimens exposed to fatigue loading and then tested for ultimate capacity  Full-scale beam test (two composite beams) • Guidelines for design, detailing, fabrication, and installation of full-depth precast concrete bridge deck panel systems. • Proposed revisions to the AASHTO LRFD Bridge Design Specifications. Recommended Full-Depth Precast Concrete Bridge Deck Panel Systems Design Criteria The following criteria were set in advance to pave the way for the development of the deck systems. The criteria were decided after a careful study of the bridges covered in the lit- erature review and national survey. The criteria were also dis- cussed with a panel of experts on this type of construction. • Type of superstructure—The slab/I-girder bridge type was used. This decision was made based on the fact that 50% to 60% of the bridges in United States are of this type (33). • Construction material—The deck slab is made from con- ventionally or prestressed reinforced concrete. The support- ing I-girder can be made of prestressed concrete or steel. • Composite versus noncomposite superstructure—It was evident from the literature review that the superstructure of the majority of bridges built with precast deck panels is made composite with the deck. Composite systems typi- cally have many advantages over noncomposite systems, such as shallower depth of the superstructure, longer spans, smaller deflection and less vibration caused by mov- ing traffic, and larger clearance. • New construction projects versus deck replacement projects—The details of the precast deck systems presented in this chapter were developed to fit new construction proj- ects, as well as deck replacement projects. This decision was made because of the almost 50/50 split between new con- struction and deck replacement projects nationwide. C H A P T E R 3 Research Results

• No longitudinal posttensioning was used—This criterion was set by the research problem statement. • No overlay was used—This criterion was set by the re- search problem statement. Two systems were developed. The general features of these systems are listed in Table 1. Although these systems were developed without utilizing longitudinal posttensioning or an overlay, the systems can be easily modified to accept those features. The following model bridge was considered to develop the recommended systems: Total width: 44 ft (13.41 m) Superstructure: Four steel girders spaced at 12 ft (3.66 m) with a top flange width of 12 to 14 in. (300 to 356 mm), or four BT-72 or NU1800 pre- stressed, precast concrete gird- ers spaced at 12 ft (3.66 m). The 12 ft (3.66 m) girder spacing was chosen to provide extreme straining actions in the deck and, consequently, the highest amount of reinforcement. Concrete deck panels: Total thickness = 81⁄4 in. (210 mm) Structural slab thickness = 8 in. (200 mm) Normal weight concrete, unit weight = 150 lb/ft3 = 23.6 kN/m3 Compressive strength at 28 days = 6.0 ksi (41.37 MPa) Grout material: Compressive strength at time of opening the bridge for traffic = 6.0 ksi (41.37 MPa) Live load: HL-93 (AASHTO LRFD specifica- tions) Side barriers: Jersey barrier, 600 plf (2.19 kN/m) per side The design was carried out in accordance with the AASHTO LRFD specifications (7). Recommended System CD-1 Figures 20 to 29 show the details of recommended system CD-1. The panel is 8 ft (2.44 m) long and covers the full width of the bridge (44 ft, or 13.41 m). Although the panel has a structural thickness of 8 in., it is made 81⁄4 in. (210 mm) thick because no overlay is used. The top 1⁄4 in. (6 mm) of the panel is used as a sacrificial layer that allows for texturing the top surface of the slab. After the panels are installed and grouted, the texture is applied by machine grinding. The tex- turing process helps ensure a uniform elevation of the fin- ished deck slab and provides a high-quality riding surface. The panel is transversely reinforced with eight 1⁄2 in. (12.7 mm) diameter pretensioned strands and 12 No. 5 (16) bars dis- tributed on two levels. A 2 in. (50 mm) top and bottom clear concrete cover is provided for the two layers of reinforcement. This amount of reinforcement is sufficient to cover the required flexural capacity at positive and negative moment area. Step-by- step design calculations of the system are given in Appendix B. The longitudinal reinforcement of the panel is made of No. 6 (19) bars at 13.3 in. (338 mm). In order to splice these 18 System Designation CD-1 CD-2 Reinforcement Type: Transverse Longitudinal Pretensioned Conventional Conventional Conventional Supporting girder and construction type: New construction projects Deck replacement projects Alteration to existing shear connectors Steel or concrete girders Steel girders High Steel or concrete girders Steel or concrete girders Minimum Made composite with the girder Yes Yes Longitudinal posttensioning No No Use of overlay No No Panel can be crowned to match the bridge profile No Yes Notes Two panel-to-panel connection details were developed for this system (CD-1A and CD-1B). A full-scale bridge mockup using this system was constructed and tested in this project. This system was not tested in this project. Table 1. General features of the conceptual designs of CD-1 and CD-2.

bars across the transverse panel-to-panel joints, two connec- tion details were developed—CD-1A and CD-1B. • CD-1A (Figures 20 to 23). On one side of the panel, the No. 6 bar is embedded 6 in. (152 mm) in a galvanized bulged hollow structural steel (HSS) tube (HSS 4 × 12 × 3⁄8 in., or 102 × 305 × 10 mm). On the other side of the panel, the No. 6 bar extends 71⁄2 in. (190 mm) outside the panel. The HSS tube is a 4 in. (102 mm) cut and is installed on its side. It is bulged in the middle to a total height of 5 in. (127 mm). To keep the HSS tube empty during casting of the panel’s concrete, its sides are covered with thin cardboard sheets. A 1 in. (25 mm) diameter plastic pipe is attached to the top surface of the HSS tube and is used to fill the tube with flowable grout. The HSS tube has an oversize 13⁄4 in. (45 mm) diameter hole on the free side of the panel to help in installing the new panel without interference with the shear connectors. The panel is installed so that the HSS tubes are ready to receive the No. 6 bars of the next panel, as shown in Figure 23. As the next panel is being installed, it will be tilted to avoid interference with the shear con- nectors of the superstructure. • CD-1B (Figures 24 to 26). On both sides of the panel, the No. 6 bar is embedded 12 in. (305 mm) in an HSS tube (4 × 12 × 3⁄8 in.). The dimensions of the HSS tube are exactly the same as those for the HSS tube in CD-1A. In this case, however, the HSS tube is not bulged, and it is provided with a 1.5 in. (38 mm) wide top slot. The slot extends all the way to the top surface of the panel. The new panel is in- stalled vertically, and then a 241⁄2 in. (622 mm) long splice bar is dropped from the top surface of the panel through the slot. The goal of using the HSS tube is to confine the grout sur- rounding the No. 6 bar, which enables the bar to develop its yield strength in a shorter distance than required for uncon- fined bars. According to the LRFD specifications, an uncon- fined No. 6 bar requires at least 18 in. (457 mm) to develop its full yield strength (7). The No. 6 bar has only about 6 in. (152 mm) of embedment length and 12 in. (305 mm) of lap splice length in CD-1A and CD-1B, respectively. These new details were tested for direct tension due to static load and for flexure due to repeated loading; the test results found full development of the No. 6 bar yield strength. A similar technique that uses a 19 Cross Section of the Bridge Plan View of the Precast Panel showing Reinforcement 8" + 1 /4 " * * 1'-0" Two 2-1/2" strands 270 ksi, LL 1'-0"1'-0" 10#6 bars@13.3" 1'-0"1'-0" 1'-0"1'-0"3 #6 @13.3" 9.4" 1'-0" 3#6 @13.3" 9.4"1'-0" Two 2-1/2" strands 270 ksi, LL 7' -1 1" C C A A See Figure 3 2% Detail H 1"-φ leveling screw 12'-0" 12'-0" 12'-0"4'-0" 4'-0" 6x2#5 10#6 bars@13.3" 10#6 bars@13.3" B B #6 @ 13.3 in. EE Short pieces of 2#4 2#4 Figure 20. Cross section and plan view of CD-1A (**1⁄4 in. is used as a sacrificial layer for texturing).

high-strength spiral wire was successfully used with the NUDECK precast system (2, 3). Forms needed for grouting the transverse joints can be built by attaching strips of plywood to the top surface of the panel, as discussed in Chapter 2. This method is recommended to ensure the shear key is completely filled with grout. The panel is made composite with the supporting girder through hidden shear pockets. The shear pockets are 12 in. (305 mm) wide, 14 in. (356 mm) long, and 5 in. (127 mm) high, and they are spaced at 48 in. (1220 mm). The dimen- sions of the pockets are optimized to minimize the volume of grout needed to fill a pocket, which will make the system more economical. The 48 in. (1220 mm) spacing of the pock- ets was chosen to simplify the fabrication process of the pan- els by minimizing the number of shear pockets to be formed, which will reduce the fabrication cost. An experimental vali- dation was conducted using push-off specimens and full- scale beam testing because the 48 in. (1220 mm) spacing was in violation of the LRFD specifications that limit the spacing to 24 in. (610 mm) (7). After the panels are installed and their elevation is adjusted using the leveling screws, the shear pockets and transverse joints between panels, including the HSS tubes, are filled with flowable grout. Two panel-to-girder connection details were developed. The first detail is used for steel girders, where 11⁄4 in. (31.8 mm) diameter steel studs are used. The reason for using the 11⁄4 in. studs, rather than the 7⁄8 in. (22 mm) diameter studs commonly used, is to minimize the dimensions of the shear pockets; two 7⁄8 in. studs are replaced with one 11⁄4 in. stud (34). The studs are set in groups at 48 in. (1220 mm), and each group has eight studs, as shown in Figure 27. A 3 in. (76 mm) spacing between studs in the longitudinal direction is proposed, which violates the LRFD specifications that stip- ulate a minimum spacing of four times the stud diameter to be used (4 × 1.25 in. = 6 in. [152 mm]). The intent of the LRFD limit has been to guarantee that the compressive stresses in concrete or grout in front of the stud will not ex- ceed the allowable bearing strength due to overlapping stress 20 Section B-B Section A-A #6 bar @13.33 in. 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"5" 5" 5" 5" 7 1/2" Section C-C 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"5" 5" 5" 5" 3/4" φvent 2" φgrouting pipe Shear pocketShear pocket 1'-2" 2'-10" 1'-2" 1'-4 1/2" HSS 14x10x1/4", 6" high piece 7" 7" 7" 7" Two 1/2" strands 270 ksi, LL 2 x 2#5 2 x 2#5 2 x 2#5 2#4 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 7" 7"7" 7" 8" +1 /4 * 3/4" 1'-2" 9 1/2" 3 3/ 4" 4 1/ 4" +1 /4 "* 6 3/4" 1"φgrouting pipe HSS 4x12x3/8", 4" long 9 1/2"7" 7" 2" +1 /4 * 2" 1'-4 1/2" 1'-2" 2#4 5" 3" + 1/ 4" * Figure 21. CD-1A, Sections A-A, B-B, and C-C (*1⁄4 in. is used as a sacrificial layer for texturing).

distributions of adjacent studs. The shear pocket is confined with an HSS tube in order to increase the compressive strength of the grout. Another alternative is to confine the shear pocket using three No. 6 closed ties. Both alternatives were experimentally investigated. The number of studs per pocket was determined based on a parametric study by Tadros and Baishya on a wide range of steel bridges with spans between 60 and 130 ft (18 and 40 m) and girder spacing between 6 and 12 ft (1.82 and 3.66 m) (2). The study concluded that one 11⁄4 in. (31.8 mm) stud at 6 in. (152 mm) would be sufficient to maintain full composite ac- tion between the deck and the steel girder. It is recommended that the designer run the analysis for the horizontal shear and determine the required number of studs for the bridge under consideration. Figure 28 shows the recommended detail for concrete girders, where clusters of three 11⁄4 in. (31.8 mm) diameter double-headed studs are set at 48 in. (1220 mm). The studs are embedded in the top flange of the prestressed concrete girder. The reason for using double-headed studs as hori- zontal shear reinforcement, rather than the commonly used vertical web shear reinforcement of the girder, is to mini- mize the shear pocket dimensions. If vertical web shear rein- forcement were used for this model bridge, 12 legs of No. 5 bars would be needed per cluster. In addition, the stud’s head provides full anchorage without consuming a large amount of space compared with the No. 5 bars that need to be bent to an L-shape or an inverted U-shape. The number of studs per cluster was determined based on information collected from the design examples given in Chapter 9 of the PCI Bridge Design Manual (35) and from bridge designers in the DOTs. However, it is recommended that the designer run the analysis for the horizontal shear and determine the required number of studs for the bridge under consideration because the amount of horizontal shear reinforcement required depends on many variables, such as the width and surface condition of the interface, span length, girder spac- ing, and girder depth. 21 Shear key details 3/4" 3/4" 1 1/2" 3 1/2" 1 1/2" 8" + 1 /4 " * * 3/4" 3/4" 1" 1'-0" Ø 1" 4"4" 5" 4" Ø 1 3/4" 4" Galvanized bulged HSS 4x12x3/8" 6" #6 Section G-G Galvanized bulged HSS 4x12x3/8 F F #6 G G Section F-F 3/4"1'-0" 3 3/ 4" 4 1/ 4" +1 /4 "* 4" 1"φgrouting pipe #6 1" 3/4"1'-0" 7 1/2" Figure 22. CD-1A, panel-to-panel connection detail, Detail D (*1⁄4 in. is used as a sacrificial layer for texturing).

Figure 29 shows the recommended detail for the panel-to- barrier connection. A closed pin bar is cast in the panel and extended outside the top surface of the panel. The bar’s size and spacing depend on the barrier’s design. The top surface of the panel at the interface between the barrier and the panel is intentionally roughened to enhance the bond capacity. The transverse edges of the panel are provided with a ver- tical shear key. The dimensions of the shear key are designed to guarantee full transfer of wheel loads from one side of the joint to the other. The modified shear friction theory (36) is used to determine the vertical shear strength of the shear key joint. The theory depends on depicting possible modes of failure of the joint. The failure modes are shown in Figure 30 and are described below. Bearing failure at side bc of the shear key (1) where φ = strength reduction factor for bearing = 0.7 (Section 5.5.4.2.1, AASHTO LRFD Bridge Design Specifications [7]), P f Lu c bc≤ ( )( )φ 0 85 12. *' kip/ft 22 old panel new panel 15 .0 0 ° 15 .0 0 ° 10 .0 0 ° 5. 00 ° Figure 23. CD-1A, installation of a new panel.

= specified concrete strength of the precast panel or the grout material, whichever is smaller, = 6.0 ksi. Lbc = length of the side bc of the shear key = 1.06 in., and Pu = factored wheel load with dynamic allowance, calcu- lated in kip per linear foot in the transverse direction. Pu ≤ (0.7 × 0.85 × 6.0 × 12 × 1.06) = 45.4 kip/ft Shear failure along line be inside the grout filling the shear key (2) where φ = strength reduction factor for shear = 0.9 (Section 5.5.4.2.1, AASHTO LRFD Bridge Design Specifica- tions), c = cohesion strength of the grout material = 0.15 ksi for concrete cast monolithically (Section 5.8.4.2, AASHTO LRFD Bridge Design Specifications), Lbe = length of the distance from b to e = 5.0 in., μ = friction coefficient of the grout material = 1.4 for con- crete cast monolithically (Section 5.8.4.2, AASHTO LRFD Bridge Design Specifications), Av = longitudinal reinforcement crossing the shear inter- face per foot = 0.44 × 12 / 13.3 = 0.397 in2/ft, P c L A fu be v y≤ +( )φ μ* *12 kip/ft fc' fy = yield strength of the longitudinal reinforcement = 60 ksi, and Pu = factored wheel load with dynamic allowance, calcu- lated in kip per linear foot in the transverse direction. Pu ≤ 0.9 (0.15 × 12 × 5 + 1.4 × 0.397 × 60) = 38.1 kip/ft Therefore, Pu = 38.1 kip/ft (556 kN/m). According to Section C3.6.1.2.5 of the LRFD specifications (7), which provides guidelines for determining the tire con- tact area of the design truck of the HL-93 live load, the width of the contact area in inches = P/0.8, where P = design wheel load in kip = 16 kip. Therefore, the width of the contact area = (16/0.8) = 20 in. The applied factored wheel load = P (load factor for live loads) (dynamic load allowance, IM) = 16 × 1.75 × 1.33 = 37.24 kip/20 in. = 22.3 kip/ft (325 kN/m) < 38.1 kip/ft (556 kN/m) Recommended System CD-2 The empirical design method given in Section 9.7.2.4 of the LRFD specifications (7) was used to design the required reinforcement. The LRFD specifications limit the use of the 23 Plan View of the Precast Panel showing Reinforcement Cross Section of the Bridge 4'-0" 8" + 1 /4 " * * 1'-0" 10#6 bars @13.3" 1'-0"1'-0" 1'-0"1'-0" 3 #6 @13.3" 9.4"1'-0" 7' -1 1" C2 C2 A2 A2 2% 12'-0"12'-0" 6x2#5 10#6 bars @13.3" B2 B2 #6 @ 13.3 in. EE Short pieces of 2#4 Short pieces of 2#4 See Figure 7 1'-0" Two 2-1/2" strands 270 ksi, LL 1'-0"3 #6 @13.3" 9.4" 1'-0" Two 2-1/2" strands 270 ksi, LL Detail H 1"-φ leveling screw 12'-0"4'-0" 10#6 bars @13.3" Figure 24. Cross section and plan view of CD-1B (**1⁄4 in. is used a sacrificial layer for texturing).

empirical method to CIP slabs because all the validation tests for this design method were conducted on CIP slabs. The empirical method depends on the arching effect, where the bottom layer of transverse reinforcement acts as a tension tie to the concrete arch that is developed between adjacent girderlines. The research team believes that this method can be equally applied to precast panel deck systems if the following condi- tions are satisfied: first, the arching effect is successfully developed by anchoring the bottom layer of transverse rein- forcement at the girderlines to make it able to fully develop its yield strength, and second, the transverse panel-to-panel joints are constructed to simulate monolithic CIP slabs by splicing the longitudinal reinforcement of the precast panels. The empirical design method requires any section of the slab between the exterior girders to have a top and bottom mesh. Each mesh is made of two layers of reinforcement. The amount of reinforcement required for each layer of the top mesh = 0.18 in2/ft (381 mm2/m), and the amount of rein- forcement required for each layer of the bottom mesh = 0.27 in2/ft (572 mm2/m). Maximum spacing of reinforcement in any layer = 18 in. (457 mm), and the minimum thickness of the slab = 7.0 in. (178 mm). The recommended system CD-2 is made of an 81⁄4 in. (210 mm) thick solid panel. The top 1⁄4 in. (6 mm) of the panel thickness is used as a sacrificial layer for texturing the top sur- face of slab. Texturing is executed by machine grinding after the panels are installed and grouted. The texturing process helps to maintain a uniform elevation of the finished deck slab and provides a high-quality riding surface. Figures 31 to 35 provide details of the recommended system. The precast panel has a partial depth continuous channel at the girderlines. The channel is covered with a 3 in. (76 mm) thick slab that houses the transverse top layer of reinforce- ment. After the precast panels are installed and their elevation is adjusted using leveling screws, the continuous channels are 24 #6 bar @13.33 in. 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 3/4" f vent 2" fgrouting pipe Shear pocketShear pocket 1'-2" 2'-10" 1'-2" 1'-4 1/2" HSS 14x10x1/4", 6" high piece 7" 7" HSS 4x12x3/8", 4" long with top slot 2 x 2#5 2 x 2#5 2 x 2#5 2#4 Section B2-B2 Section A2-A2 8" +1 /4 * 2" +1 /4 * 2" 7" 7" 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5"7" 7" Two 1/2" strands 270 ksi, LL 3/4" 1'-2" 3 3/ 4" 4 1/ 4" + 1/ 4" * Section C2-C2 7" 7" 1'-4 1/2" 1'-2" 2#4 5" 3" + 1/ 4" * Figure 25. CD-1B, Sections A2-A2, B2-B2, and C2-C2 (*1⁄4 in. is used as a sacrificial layer for texturing).

filled with nonshrink grout through grouting pipes provided in the 3 in. slab. The panel is reinforced with three layers of reinforce- ment—transverse top and bottom reinforcement layers and one longitudinal reinforcement layer provided near the mid height of the panel. The longitudinal layer combines the two longitudinal layers of reinforcement required by the empiri- cal design method. The amount of reinforcement for these layers satisfies the reinforcement requirements of the empir- ical design method, as follows: Top transverse layer between the exterior girders = 1 No. 6 @ 18 in. = 0.293 in2/ft > 0.18 in2/ft Bottom transverse layer between the exterior girders = 1 No. 6 @ 18 in. = 0.293 in2/ft > 0.27 in2/ft To fit the 18 in. spacing between the transverse bars, the panel is made 9 ft long. Longitudinal layer of reinforcement = 1 No. 8 Grade 60 steel with 4 in. long threaded ends @ 15 in. = 0.601 × 12/15 = 0.481 in2/ft > (0.18 + 0.27) = 0.45 in2/ft The longitudinal No. 8 (25) bars are spliced using HSS 8 × 4 × 3⁄16 in. (203 × 102 × 5 mm), 31⁄2 in. (89 mm) long cut, Grade 36 tubes, and heavy-duty nuts, as shown in Figure 33. The HSS tube is installed in 10 × 6 in. (254 × 152 mm) pre- fabricated pockets located on one transverse edge of the panel. The No. 8 bars extend about 4 in. (102 mm) inside the pocket and about 4 in. outside the other transverse edge of the panel. The panel to be installed is vertically lowered and then it is moved horizontally until the No. 8 bars are inserted 25 #6, 24.5" long splice bar 2'-0" Section G2-G2 Galvanized bulged HSS 4x12x3/8, 4" long F2 F2 G2 G2 Section F2-F2 Detail D2 3 3/ 4" 4 1/ 4" + 1/ 4" * 11 1/2" 1" 11 1/2" 1.5" wide vertical slot X2 X2 Section X2-X2 1 1/2" 4" 3/4"1'-0" 11 1/2" 3 3/ 4" 4 1/ 4" + 1/ 4" * #64" 1 1/ 2" #6 #6 #6 Shear key details Galvanized bulged HSS 4x12x3/8, 4" long 1'-0" 4" 10 3/4"1 1/2" 4" 4" 1 1/ 2" 1 3/4" 1" 1 1/4" 1 1/2" 1" 1 1/2" 2" 1 1/4" 3/4" 8" + 1 /4 "* 11" 11" 1'-0" 3/4" 3/4" 1 1/2" 3 1/2" 1 1/2" 3/4" 3/4" 1" Figure 26. CD-1B, panel-to-panel connection detail, Detail D2 (*1⁄4 in. is used as a sacrificial layer for texturing).

3 1/2" 1 1/4" studs 1 " 2" φgrouting pipe Light weight angles used as grout barrier and to adjust for the panel elevation Rectangular bar 5 " 3 " + 1 / 4 " * Section C-C 1'-2" 2'-10" 1'-2" 1'-4 1/2" 2 1/2" 8- 1 1/4" studs 5 " 3 " + 1 / 4 " * 1 " Top surface of the steel girder flange 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 7" 7"7" 7" 3/4" φvent 2" φgrouting pipe HSS 14x10x1/4", 6" high piece 5"3 1/2" Section E-E 9" 1'-0" 2 1/2" 1'-4 1/2" 5 " 3 " + 1 / 4 " * Figure 27. Sections C-C and E-E for steel girders (*1⁄4 in. is used as a sacrificial layer for texturing).

in the HSS tubes. The thickness of the HSS tube is designed to provide 125% or more of the yield capacity of the No. 8 bar, as follows: Yield capacity of the No. 8, Grade 60, with threaded ends = 0.601 × 60 = 36.1 kip (161 kN) Yield capacity of the 3⁄16 in. thick HSS = × 36 = 47.5 kip > 125%(36.1) = 45.1 kip (201 kN) > 36.1 kip (161 kN) If the bridge owner requires corrosion protection measures to be used for the deck reinforcement, the top and bottom transverse reinforcement layers can be made of epoxy-coated 2 3 16 3 1 2 × × 27 Top surface of the concrete girder flange 3- 1 1/4" double headed studs 5" 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 3" + 1/ 4" *3/4" f vent 2" f grouting pipe 1'-4 1/2" 1'-2" 2'-10" 1'-2" 1'-4 1/2" 7" 7"7" 7" HSS 14x10x1/4", 6" high piece 1" 5" 3" +1 /4 "* 3" 4" 4" 3" Section E-E 5" 3" + 1/4" ** 1" 1'-0" 2" f grouting pipe 1 1/4" double headed stud 1" f backer rod Section C-C 5 7/8" 3'-2 3/8" 5 1/2" 5 1/4" 4'-0 1/4" 5 1/ 2" 8 1/ 2" 2 7 /8 " 1 1/4 in. stud 2" Figure 28. Sections C-C and E-E for concrete girders (*1⁄4-in. is used as a sacrificial layer for texturing).

reinforcement, while the longitudinal reinforcement layer with the coupling accessories can be made of galvanized steel. The transverse edges of the precast panels are provided with a female shear key. The dimensions of the shear key are identical to those with CD-1. Using the modified shear fric- tion theory (36) shown in CD-1, it can be seen that the shear key detail with No. 8 bars at 15 in. (381 mm) has enough capacity to transfer the weight of the HL-93 design load. The empirical design method does not apply to the over- hanging part of the slab. It is thus necessary to design the overhang for collision effects. The design calculations of the overhang are provided in Appendix B of this report. It is also important to check the stresses in the bottom layer of reinforcement—six No. 6 (19) bars at the girderline loca- tions—during shipping and handling. If the panel is lifted at the girderline locations (the continuous blockout channels), this area will be in negative moment. The compression force of this moment will be carried by the bottom six No. 6 bars, and the tension force will be carried by the top six No. 6 and top 12 No. 8 bars. The bottom six No. 6 bars need to be checked against buckling, as follows: Panel weight = (8/12)(0.150)(9) = 0.9 kip/ft Negative moment = (0.9)(122)/(10) = 12.96 kip/ft Tension force = (12.96 × 12)/(4.25) = 36.6 kip (distributed on six No. 6 bars) = 36.6/6 = 6.1 kip/bar = 6.1/0.44 = 13.9 ksi Allowable stress of No. 6 bar, Fa = where Es = modulus of elasticity of the bar = 29,000 ksi Fy = yield strength of the bar = 60 ksi Kl = effective buckling length of the bar = 1.0 × 12 in. = 12 in. r = radius of gyration of the bar = 0.25 × the bar diameter (Kl/r) = (12)/(0.25×0.75) = 64.00 Fa = 25.1 ksi > 13.9 ksi (safe) Panel to Panel Connection Development of the panel-to-panel connection details, which were used in the recommended systems CD-1A and CD-1B, was achieved using the following approach: Kl r Cc / . . ⎛⎝⎜ ⎞⎠⎟ = <0 655 1 0 Cc = = 2 29 000 60 97 67 2π ( , ) . F Kl r C Kl r C y c c 1 1 2 5 3 3 8 2 − ⎛⎝⎜ ⎞⎠⎟ ⎡ ⎣⎢ ⎤ ⎦⎥ + ⎛⎝⎜ ⎞⎠⎟ − / / 1 8 1 0 2 3 2 Kl r C Kl r C C E F c c c s / , / . ,⎛⎝⎜ ⎞⎠⎟ ⎡ ⎣⎢ ⎤ ⎦⎥ ≤ = π y 28 Intentionally roughened surface Width of the base of the barrier 8" + 1/ 4" * * #5 closed loop c a b d e f P R Bearing failure mode 8" #6@13.3" c a b d e f P Shear failure mode Figure 29. Detail H, panel-to-barrier connection detail (**1⁄4-in. is used as a sacrificial layer for texturing). Figure 30. Design parameters of the shear key.

• Investigation and Review of Grout Materials Available in the Market—This investigation was conducted to decide on the type of grout material to be used in the experimen- tal investigation. • Group 1 Specimens—Initially, four connection details were developed and tested in direct tension using 16 pull- out specimens. In these details, an HSS tube is used to con- fine the grout surrounding the spliced bars. Confinement typically increases the grout strength, resulting in a signif- icant reduction in the length required to fully develop the yield strength of the bar. Based on the experimental results of these specimens, the top three successful connection details were considered in the next step. • Group 2 Specimens—Nine pullout specimens, representing the top three successful details of Group 1, were fabricated and tested in direct tension to confirm the experimental results obtained in Group 1. Based on the experimental results, two connection details were chosen as the final con- nection details used in the development of the recom- mended systems CD-1A and CD-1B. • Full-Scale Bridge Specimen—After the recommended sys- tems CD-1A and CD-1B were developed, a full-scale bridge specimen made of three precast panels and utilizing the candidate connection details was fabricated and tested for 2,000,000 cycles of fatigue load. Investigation of Various Grout Materials To determine what type of grout material should be used in the experimental investigation, the research team reviewed the specifications of many grout materials commercially avail- able in the market. The products chosen in the review process were selected based on the results of the literature review and national survey. Table 2 compares some of the products that were considered in the review process. To choose the grout material that will be used in the entire experimental program, the research team set the following criteria: (a) nonshrink grout, (b) 6 ksi (41.4MPa) compressive strength at 1 day, (c) not reactive with steel, (d) high flowability, and (e) can be mixed with pea gravel to increase the yield volume. 29 Cross Section of the Bridge 2% Detail H Plan View of the Precast Panel showing Reinforcement Detail E2#6@18" #8 bar with 5" long threaded ends C C D D 6" 6" 6" 8 spacings x 15" 6" 6" 1'-0" 6" 6" 8 spacings x 15" 2 sp ac in gs @ 15 "1'-0" 1'-0" KK 4- 1" φgrouting holes per girderline 4'-0"12'-0"12'-0"12'-0"4'-0" 2% 2% 8" + 1 /4 " * 8" + 1 /4 " * 8 spacings x 15" 2 sp ac in gs @ 15 " 1'-0" 6" 6" 6" 4#8@18" Two 1" φ leveling screws 8' -1 1" A A B B Figure 31. Cross section and plan view of CD-2 (*1⁄4 in. is used as a sacrificial layer for texturing).

The research team decided to use SS Mortar for the exper- imental investigation of the connection details. This decision was based on the fact that SS Mortar is exclusively designed for splice connections where steel tubes are used to confine reinforcing bars, which is the case with the new connection details developed for CD-1A and CD-1B, where HSS tubes are used. In addition, SS Mortar has a relatively high flowa- bility, which helps in filling tight connection details. The research team monitored the compressive strength gain of 2 × 2 × 2 in. (51 × 51 × 51 mm) SS Mortar cubes over a period of 28 days. It was found that the early-age compressive strength measured by the research team was higher than that given by the manufacturer. The SS Mortar was able to reach a compressive strength of 6.0 ksi in less than 1 day, which makes this type of grout suitable for use in weekend construction projects. However, the 28-day compressive strength measured by the research team was less than that specified by the manu- facturer, as shown in Figure 36, but it was higher than the tar- get compressive strength specified by the research team for use with the recommended systems (6.0 ksi [41 MPa]). No shrink- age cracks were observed in either the SS Mortar filling the tubes of the pullout specimens or the 2 × 2 × 2 in. cubes. Based on a discussion with the SS Mortar manufacturer on how to increase the yield volume of the mortar, a trial mix of SS Mortar and 1⁄4 in. diameter pea gravel was made. The ratio of pea gravel to SS Mortar was 1 to 2 by weight. The trial mix was placed in 4 × 8 in. cylinders to monitor the compressive strength gain with time, as shown in Figure 36. SS Mortar with 50% pea gravel showed slightly slower gain of compres- sive strength than the SS Mortar without pea gravel. How- ever, both mixes reached almost the same compressive strength at 28 days. Pure SS Mortar, with no pea gravel, was used for the pullout specimens. Group 1: Direct Tensile Test of Four Connection Details Sixteen pullout specimens were fabricated and tested. The following variables were considered in making the 16 specimens: 30 Section D-D #8 bar with 5" long thread ends 4" 4" 4" 5" Section B-B Section A-A 2#6 bar @ 18" Section C-C 3/4" 1'-6" 1'-6" 1'-6" 1'-6" 5" 3" 5" 2#6 bar @ 18"2#8 bar @ 18" 1 1/2" 3/4" 3 1/2" 3/4" 1 1/2" #8 bar with 5" long thread ends3/ 4" 3/ 4" 4" 4" 4" 5" 8'-11" 9'-0" 1/2" 1/2" 3/4" 3/4" 2" 2" 1" 10" 1" 2" 1" 2" 1" 10" 1'-6"1'-1" 6" 4" 6" 4" 2" 6" + 1/ 4" * 2" 6" + 1 /4 "* 8" + 1 /4 "* 8" + 1/ 4" * Figure 32. CD-2, Sections A-A and B-B (*1⁄4 in. is used as a sacrificial layer for texturing).

• Size of the HSS Tube. Two sizes were used—HSS 3 × 12 × 1⁄4 in. (76 × 305 × 6 mm) and HSS 4 × 12 × 3⁄8 in. (102 × 305 × 10 mm). For both sizes, a 4 in. (102 mm) long strip was used. These sizes were chosen because they fit the 8 in. (203 mm) thickness of the panel, while satisfying the minimum top and bottom concrete cover for reinforcement in deck slabs as spec- ified by the AASHTO LRFD specifications (7), and because they are commercially available from many producers. • Size of Spliced Bar. Two bar sizes were considered—No. 6 (Metric No. 19) and No. 7 (22), Grade 60 (414 MPa) uncoated. • Connection Details. – Detail A—A bulged HSS tube with two side holes. The spliced bars were embedded for 6 in. (152 mm) inside the tube, but they were not overlapped. The purpose of bulging the tube was to increase the volume of grout that is confined within the tube and optimize the required development length. – Detail B—A straight HSS tube with a 12 in. (305 mm) long slot located on the top surface of the tube. The developed bars were embedded 11 in. (279 mm) inside the tube, and they overlapped each other. – Detail C—A straight HSS tube with a side slot. The devel- oped bars were embedded for 6 in. (152 mm) inside the tube, but set head to head. This detail was similar to Detail A except that the tube had no bulge. – Detail D—A bulged HSS tube with a 6 in. (152 mm) long slot on the top surface of the tube. The spliced bars were embedded for 6 in. (152 mm) inside the tubes and set head to head. Table 3 shows the design criteria of the 16 pullout speci- mens, and Figure 37 shows the details of the test specimens. Each HSS tube was embedded in an 8 × 12 × 24 in. (203 × 305 × 610 mm) concrete prism, and the concrete prism was reinforced with two No. 4 (13) top bars and two No. 5 (16) bottom bars. This amount of reinforcement was chosen to simulate the reinforcement required by the empirical design method given by the AASHTO LRFD specifications (7). The tube was set flush with one face of the concrete prism, and one of the two developed bars was embedded in the prism and extended inside the HSS tube to represent the longitudi- nal reinforcement of the panel. This bar was extended outside the concrete prism from the other side so it could be hooked 31 Shear key details 3/4" 1" 3/4" 1 1/2" 3 1/2" 1 1/2" 8" + 1 /4 " * 3/4" 3/4" Detail H Intentionally roughened surface Width of the base of the barrier 8" + 1/ 4 " * HSS 8x4x3/16", A36 Section G-G 1"φall thread bar, 150 ksi Section F-F, Detail E 6" + 1 /4 " * 2 1/4" 3 1/2" 1/4" HSS 8x4x3/16" GG 1"φHex nut with 1/4" thick washer F 1"1/ 2" 1/ 2" F 4" 4" 2" 6"5" 1/ 2" 1/ 2" 9" 8 5/8" 1/4" 2" 1/4" 2 5/8" 2" 1/4" 1/4" 4" 3 5/ 8" φ = 2 " 3" 2" 3/ 4" 8" φ = 1 1 /4 " Figure 33. CD-2, Details E and H (*1⁄4 in. is used as a sacrificial layer for texturing).

with the grip of the testing machine. To keep the HSS tubes from filling during concrete casting of the prisms, the sides of the tubes were covered with thin pieces of cardboard. Figure 38 shows the specimens during fabrication. A normal weight concrete mix with a specified 28-day concrete strength of 6.0 ksi (41 MPa) was used. The specimens were moist cured for 7 days, and 4 × 8 in. (102 × 203 mm) concrete cylinders were made to monitor the compressive strength gain with age. Figure 36 shows the compressive strength gain with age. After the compressive strength of concrete mix reached 6.0 ksi (41 MPa), the second bar was embedded in the tube and the tube was then filled with SS Mortar grout. No pea gravel was added to the grout mix. The specimens were tested when the grout was 3 days old using the Tinius Olsen Machine, as shown in Figure 39a. The specimens were loaded at a fixed rate of 300 lb/sec (1334 N/sec) until failure. Two modes of failure were observed. The first was bar slippage (as shown in Figure 39b), where the failure load was measured at the moment when the bar started to slip away from the concrete prism. This moment was identified when a sudden drop of the applied load was observed on the load gauge of the test- ing machine. The second failure mode was prism failure (as shown in Figure 39c), where the concrete around the HSS tube failed in axial tension. Based on the test results that are given in Table 3, the following conclusions were reached: • Specimens A-1, B-1, and D-3 had shown higher developed strength than the rest of the specimens. These specimens exceeded 1.25 times the specified minimum yield strength (60 ksi) of the spliced Grade 60 steel bars. • In all connection details, both sizes of the HSS tubes showed almost identical behavior and developed almost the same amount of strength for the same spliced bar sizes. • Connection details made with No. 7 (22) bars have shown lower developed strength than those made with No. 6 (19) bars. This was expected because the same amount of con- finement and development length was used for the No. 6 and No. 7 bars for every connection detail. Based on the test results of the pullout specimen, the re- search team decided to consider the connection details A and B with No. 6 bar and HSS 4 × 12 × 3⁄8 in. (102 × 305 × 10 mm) 32 Top surface of the flange of the steel girder Section B-B 5" 3" + 1 /4 " *Steel studs 1" 2"φgrouting pipe 3" 5" Light weight angles used as grout barrier Rectangular bar 9'-0" 1'-0" Section K-K K K 5 1/ 2" Figure 34. CD-2, Sections B-B and K-K for steel girders (*1⁄4 in. is used as a sacrificial layer for texturing).

tube in the next step of investigation. Although the small-size HSS 3 × 12 × 1⁄4 in. (76 × 305 × 6 mm) tube showed almost the same structural behavior as, and developed a bar strength similar to that developed with, the HSS 4 × 12 × 3⁄8 in. tube, the research team decided to use the larger tube as it provides higher construction tolerance. Group 2: Direct Tensile Test of Selected Connection Details In this group, nine pullout specimens were tested, repre- senting three connection details and three specimens per detail. In all specimens, No. 6 (19) bars and HSS 4 × 12 × 3⁄8 in. (102 × 305 × 10 mm) tubes were used. The connection details considered in this group were as follows (see Figure 37): • Detail A—Same as Detail A of Group 1. • Detail BB—Same as Detail B of Group 1 except that the slot on the top surface of the HSS tube was open all the way to the top surface of the concrete prism. This change was made to simulate the connection detail that would be used later on the recommended system CD-1B. • Detail AA—Same as Detail A of Group 1 except that the HSS tube was not bulged. This detail was added to check the effect of bulging the HSS tube on the developed bar strength. Table 4 shows the design criteria of the nine pullout spec- imens, and Figure 40 shows details of the specimens during fabrication. The concrete mix and grout material used for the specimens of Group 1 were used for the specimens of Group 2 (see Figure 36 for the compressive strength gain with age). The specimens were tested in direct tension when the con- crete strength of the grout was about 6.5 ksi (44.82 MPa). At that time, the concrete strength of the prism was about 6.1 ksi (42.06 MPa). Bar slippage failure occurred in all of the specimens, where the failure load is measured at the moment when the bar 33 Section B-B K 3" 5" 5" 1" 1'-0" Shear connector 3" + 1/ 4" * 2"φgrouting pipe K Top surface of the flange of the girder 9'-0" Section K-K Figure 35. CD-2, Sections B-B and K-K for precast concrete girders (*1⁄4 in. is used as a sacrificial layer for texturing).

Set 45 Set 45 HW Construction Grout SS Mortar Masterflow 928 747 Rapid Setting Grout S Grout Sonogrout 10K Description Magnesium phosphate patching and repair mortar. It sets in 15 minutes. Hot weather, magnesium phosphate patching and repair mortar Noncatalyzed multi-purpose, mineral aggregate grout. High-precision, high-strength, cement-based, metallic aggregate mortar. It is used for NMB splice sleeve splicing system. High-precision, hydraulic-cement- based, mineral aggregate grout. Ideal for grouting machines or plates with precision load bearing support. Nonmetallic cement-based grout. It is used wherever a rapid setting material is needed. Shrinkage compensated, nonmetallic, cement-based grout. It is used for applications requiring strength and durability. Shrinkage compensated, portland- cement-based, high-strength grout. 1 hour 2.0 @ 72 o F – – – – – 3 hours 5.0 @ 72 o F 3.0 @ 95 o F – – – – 6 hours 5.0 @ 72 o F 1.2 @ 36 o F 5.0 @ 95 o F – – – – – – – – – – 1 day 6.0 @ 72 o F 5.0 @ 36 o F 6.0 @ 95 o F 1.5 4.0 @ 70 o F 7.5 @ 77 o F 4.0 3.5 @ 77 o F 1.6 @ 70 o F 3 days 7.0 @ 72 o F 7.0 @ 36 o F 7.0 @ 95 o F 5.0 5.4 @ 70 o F 8.2 @ 77 o F 5.0 5.0 @ 77 o F 3.8 @ 70 o F 7 days – – 6.0 7.0 @ 70 o F 10.5 @ 77 o F – 6.0 @ 77 o F 5.1 @ 70 o F Com- posite strength (ksi) 28 days 8.5 @ 72 o F 8.5 @ 36 o F 8.5 @ 95 o F 7.0 11.0 @ 70 o F 12.6 @ 77 o F 8.0 8.0 @ 77 o F 6.2 @ 70 o F Features - High early strength at 1 hour - Superior bonding - Very low drying shrinkage - Resistant to freeze/thaw, sulfate and deicing chemicals - Superior bonding - Very low drying shrinkage - Resistant to freeze/thaw, sulfate and deicing chemicals - Can be extended with pea gravel - Designed for the 50 oF to 90 oF range - Non rusting - Nonshrink grout - High flowability, suitable for pumping in tight spaces - Can be used over wide range of temperature - Design for use with splice sleeve system - Nonshrink grout - Resistant to freeze/thaw & sulfates - High flowability, suitable for pumping in tight spaces - Designed for the 40 oF to 90 oF range - Nonshrink grout - High early strength at 1 day - Chloride free - Nonrusting - Not recommended for placing below 35 oF - Recommended for shear key grouting - Can be extended by adding pea gravel - High flowability - Cannot be extended by adding gravel - Shrinkage compensated grout Yield (ft3/bag) 0.39 w/o gravel 0.58 w 60% gravel 0.39 w/o gravel 0.58 w 60% gravel 0.45 w/o gravel 0.42 w/o pea gravel 0.50 w/o pea gravel – 0.5 w/o gravel 0.6 w 27% gravel 0.69 w 55% gravel 0.40 Source: manufacturers’ literature. Table 2. Comparison between various types of commercial grout material.

started to slip away from the concrete prism, as shown in Figure 41 Specimen A. This moment was identified when a sudden drop of the applied load is observed on the load gauge of the testing machine. For the specimens made with Detail BB (i.e., specimens with a top slot), cracks between the grout filling the slot and the specimen were observed very close to the failure load, as shown in Figure 41 Specimen BB. The fail- ure load and the equivalent developed bar strength are given in Table 4. The following conclusions were drawn from the experimental program of the pullout specimens. • All of the connection details tested with No. 6 (19) spliced bars were able to develop a bar strength equal to or greater than 125% of the 60 ksi (413.7 MPa) yield design strength, which is consistent with the require- ment of Section 5.11.5.2.2 of the AASHTO LRFD speci- fications (7). • Connection Detail AA made with straight tubes and no slots showed about a 5% increase in developed strength compared with connection details made with slotted tubes (Detail BB). • Connection details made with bulged tubes and no slots (Detail A) showed about a 10% increase in developed strength compared with connection details made with straight tubes (Detail AA). Based on the test results of this group, connection Details A and BB were considered in the development of the recom- mended systems CD-1A and CD-1B. Full-Scale Bridge Specimen After connection Details A and BB were used to develop the recommended systems CD-1A and CD-1B, respectively, the structural behavior of these connections as a result of fatigue flex- ural loading was investigated. The experimental investigation was conducted by building a full-scale bridge specimen. The bridge was made of a concrete deck measuring 20 ft (6.1 m) wide, 24 ft (7.31 m) long, and 8 in. (203 mm) thick and supported by two W18 × 119 steel beams. The steel beams were set 12 ft (3.66 m) on center. The concrete deck was made of three precast con- crete panels, each 20 ft (6.10 m) wide by 8 ft (2.44 m) long. The design and details of these panels were according to the recom- mended system CD-1. The panel-to-panel connection details, Details A and BB, were used on these panels as follows: • Panel P1—Detail A was used on the north and south trans- verse joints. • Panel P2—Detail A was used on the north transverse joint, and Detail BB was used on the south transverse joint. • Panel P3—Detail BB was used on the north and south transverse joints. 35 0 2 4 6 8 10 12 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Age (days) Co m p. S tre ng th (k si) SS Mortar Tested 2x2" Cubes SS Mortar Specifications Concrete Mix of the pullout specimens SS Mortar with Pea Gravel Tested 4x8" Cylinders Figure 36. Compressive strength versus time for SS Mortar and concrete mix.

Connection Detail Size of HSS Tube (in.) Type of Slot NS = No Slot TS = Top Slot SS = Side Slot Straight (S) or Bulged (B) Tube Size of Bar Embedment Length (in.) Mode of Failure Failure Load (kip) Developed Bar Strength, fd (ksi) 60 %df 1 HSS 4x12x3/8 NS B No. 6 6” head-to-head Prism 37.7 85.3 142% 2 HSS 4x12x3/8 NS B No. 7 6” head-to-head Bar Slip 43.4 72.2 120% 3 HSS 3x12x¼ NS B No. 6 6” head-to-head Bar Slip 32.8 74.2 124% A 4 HSS 3x12x¼ NS B No. 7 6” head-to-head Bar Slip 45.2 75.2 125% 1 HSS 4x12x3/8 12”-TS S No. 6 12” overlapped Prism 36.0 81.5 136% 2 HSS 4x12x3/8 12”-TS S No. 7 12” overlapped Prism 40.9 68.0 113% 3 HSS 3x12x¼ 12”-TS S No. 6 12” overlapped Prism 37.0 83.8 140% B 4 HSS 3x12x¼ 12”-TS S No. 7 12” overlapped Prism 40.3 67.0 112% 1 HSS 4x12x3/8 SS S No. 6 6” head-to-head Bar slip 23.3 52.7 88% 2 HSS 4x12x3/8 SS S No. 7 6” head-to-head Bar slip 34.6 57.5 87% 3 HSS 3x12x¼ SS S No. 6 6” head-to-head Bar slip 34.4 77.8 130% C 4 HSS 3x12x¼ SS S No. 7 6” head-to-head Bar slip 30.0 49.9 83% 1 HSS 4x12x3/8 6”-TS B No. 6 6” head-to-head Bar slip 35.5 80.4 134% 2 HSS 4x12x3/8 6”-TS B No. 7 6” head-to-head Prism 24.1 40.0 67% 3 HSS 3x12x¼ 6”-TS B No. 6 6” head-to-head Bar slip 28.5 64.5 108% D 4 HSS 3x12x¼ 6”-TS B No. 7 6” head-to-head Bar slip 30.5 50.7 85% Table 3. Design criteria and test results of the pullout specimens (Group 1).

37 #6 or #7 2'-0" 1'-6 1/2" 2" 1" 4"8" #5 bottom #4 topHSS 4x12x3/8" Connection Detail AA, Group #2 (A modified version of Detail A) 6" 6" 1'- 0" 8" 4" #6 or #7 2'-0" 1'-6 1/2" 2" 1" 4"8" #5 bottom #4 topHSS 4x12x3/8" Connection Detail C, Group #1 6" 6" 1'- 0" 8" 4" #6 or #78" Connection Detail D, Group #1 1" 2'-0" 1'-6 1/2" 2" 5" 4" #5 bottom #4 topBulged HSS 4x12x3/8" 6" 6" 1'- 0" 8" 4" #6 or #78" Connection Detail A, Group #1 & #2 1" 2'-0" 1'-6 1/2" 2" 5" 4" #5 bottom #4 topBulged HSS 4x12x3/8" 6" 6" 1' -0 " 8" 4" #6 or #7 2'-0" 1'-6 1/2" 2" 1" 4"8" #5 bottom #4 topHSS 4x12x3/8" Connection Detail BB, Group #2 (A modified version of Detail B) 11" 1'- 0" 8" 4" Connection Detail B, Group #1 #6 or #7 2'-0" 1'-6 1/2" 2" 1" 4"8" #5 bottom #4 topHSS 4x12x3/8" 11" 1' -0 " 8" 4" Group #1 Tubes Figure 37. Details of the pullout specimens of Groups 1 and 2.

38 (a) Test Setup (b) Bar slippage failure (c) Tension failure Figure 38. Specimens used in Group 1 during fabrication. Figure 39. Test setup and failure modes of Group 1 specimens. Connection Detail Size of HSS Tube (in.) Type of Slot Straight (S) or Bulged (B) Tube Size of Bar Embedment Length (in.) Mode of Failure Failure Load (kip) Developed Bar Strength, fd (ksi) 60 %df 1 Bar Slip 37.6 85.5 2 Bar Slip 34.2 77.8 3 HSS 4x12x3/8 No slot B No. 6 6 in. head-to-head Bar Slip 38.7 88.0 A Average 83.8 139.7% 1 Bar Slip 32.9 74.8 2 Bar Slip 32.9 74.8 3 HSS 4x12x3/8 12 in. Top slot S No. 6 12 in. overlapped Bar Slip 33.8 76.8 BB Average 75.5 125.8% 1 Bar Slip 34.7 78.9 2 Bar Slip 32.9 74.8 3 HSS 4x12x3/8 No slot S No. 6 6 in. head-to-head Bar Slip 35.4 80.5 AA Average 78.0 130.0% Table 4. Design criteria and test results of the pullout specimens (Group 2).

Figures 42 and 43 show details of the bridge specimen. Fig- ures 44 and 45 show the precast panels during fabrication and after 7 days of moist curing. Top and bottom layers of the strands were initially tensioned to 205 ksi (0.76 fpu), and con- crete was cast on the next day of tensioning the strands. A normal weight concrete mix of 6.0 ksi specified compressive strength was used. After the concrete was cast and consolidated, the panels were continuously moist cured for 7 days using wet burlap. Three days after the concrete was cast, the strands were released using a mechanical hydraulic system that allows gradual release of the tension force of the strands at one end of the bed. This technique of prestress release was used to protect the panels against cracking that could result from sudden release of the prestress force. The research team also used this technique during fabrication of similar precast pan- els (2, 3, 5, 6). No cracks were observed at the panel edges during or after the prestress release. No shrinkage cracking was observed on the top surface of the panel. Figure 46 shows the strength gain of the concrete mix with age. The curing process continued for 4 days after the strands were released, and the panels were kept exposed to the laboratory environ- ment afterward, where the average temperature was about 80 °F in the morning and 70 °F at night and the average rela- tive humidity was about 40% to 50%. Regular checking of the top and bottom surfaces of the panels at different ages did not reveal any shrinkage cracks. Figure 47 shows the test setup, where a self-equilibrium frame was built at the transverse joint. The self-equilibrium frame consisted of a top and bottom beam connected to- gether with four 2.0 in. diameter high-strength threaded rods. A 110 kip (489 kN) hydraulic actuator and a load spreader beam were used to apply the fatigue load. The spreader beam was supported by the precast panel at two points spaced at 6 ft (1.82 m) using two Neoprene pads meas- uring 9 × 22 in. (229 × 559 mm) each. The dimensions of the Neoprene pads were determined according to the LRFD specifications (7). The supports were positioned on one side of the transverse joint. This load arrangement simulated the center axle of an HS20 truck. The applied load fluctuated between 4.00 kip (17.8 kN) and 46.56 kip (207.1 kN). The 4 kip (17.8 kN) load was determined in order to maintain sta- bility of the test setup, while the 42.56 kip (189.3 kN) differ- ence between high and low loads was determined based on the weight of the center axle of the HS20 truck plus dynamic allowance, 32 kip × 1.33 = 42.56 kip (189.3 kN). The fatigue load was applied for 2,000,000 cycles at 2 cycles per second, as recommended by ASTM D6275 (37). 39 Figure 40. Specimens used in Group 2 during fabrication. Specimen A Specimen BB Figure 41. Failure modes of Group 2 specimens.

The research team used the chance of building a full-scale bridge specimen to address several questions that were raised about the construction feasibility of the recommended sys- tem CD-1, as follows: • Would panels made with connection Detail A be installed without interfering with the shear stud cluster? This issue was addressed by welding steel pipes, 21⁄2 in. (63.5 mm) diameter and 51⁄2 in. (140 mm) high, on the top surface of the steel beams to simulate the footprint of eight 11⁄4 in. (31.8 mm) diameter studs. The pipes were set in clusters at 48 in., four pipes per cluster. The four pipes in each cluster were welded at the corners of the perimeter of the stud cluster, as shown in Figure 48. • Would the 1 in. gap in the shear key be wide enough to allow efficient filling and consolidation of the grout? This issue was addressed by attaching 6 in. (152 mm) wide strips of plywood to the bottom surface of the precast pan- els at the transverse joints. The plywood strips were hung from the top surface using short pieces of threaded rods. 40 Cross Section of the Bridge 8" 20'-0" 4'-0" 12'-0" 4'-0" 11" 1'-0 7/8" W18x119 11" 1'-0 7/8" W18x119 A B C A B C 8' -0 " 8' -0 " 8' -0 " P1 P2 P3 3' -8 " 1'-0" 3' -8 " 1/2" strand #5 bar #6 bar #4 bar Detail A Detail BB Plan view showing the reinforcement details Figure 42. Cross section and plan view of the full-scale bridge specimen.

41 5" 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 3" 3/4" φ vent 2" φ grouting pipe Section C-C of Panels P1, P2 & P3 5" 3" Shear pocketShear pocket HSS 14x10x1/4" 6" high piece 1'-4 1/2" 1'-2" 2'-10" 1'-2" 1'-4 1/2" Section B-B of Panels P1 #6 bar @ 13.33 in. 3/4" 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 3 3/ 4" 4 1/ 4" 7 1/2" 6 3/4" 1" φ grouting pipe HSS 4x12x3/8", 4" long Section A-A of Panels P1, P2 & P3 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" Section B-B of Panel P2 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 3 3/ 4" 4 1/ 4" 6 3/4" 3/4" 1'-0" 1 1/4" 1 1/4" 3/4"1'-0" HSS 4x12x3/8", 4" long HSS 4x12x3/8", 4" long with top slot Section B-B of Panel P3 1'-2" 2'-0" 1'-7" 2'-0" 1'-2" 9 1/2" 9" 9 1/2" 9 1/2"9 1/2" 5" 5" 5" 5" 3 3/ 4" 4 1/ 4" 3/4"1 1/4"3/4" 1'-0" HSS 4x12x3/8", 4" long with top slot HSS 4x12x3/8", 4" long with top slot 8" 2" 2" Two 1/2" strands 270 ksi, LL 2#42#4 2 x 2#5 2 x 2#5 2 x 2#5 1" φ grouting pipe Figure 43. Sections A-A, B-B, and C-C of Panels P1, P2, and P3.

42 HSS 12x4x3/8 in. Precast panels ready to receive concrete Panel P1: Detail A, North side Panel P1: Detail A, South side Panel P3: Detail BB, North and South sides Figure 44. Panels P1, P2, and P3 during fabrication.

• Would the grout be able to travel the 48 in. distance between the shear pockets to completely fill the haunch between the precast panels and the steel beam? This issue could not be addressed in this part of the experimental investigation because no 11⁄4 in. (31.8 mm) studs were welded on the steel beams in order to save the beams for the full-scale beam specimens. This issue was, however, addressed during the construction of the full-scale beams. The following steps were taken to build the full-scale bridge specimen: • Backer rods measuring 1 in. in diameter were glued to the top surface of the steel beams to form the haunch between the panels and the steel beams. • Panel P2 was installed vertically and set on the steel beams using 1.0 in. high steel shims. • Panel P1 was lifted from the prestressing bed. The panel was tilted about 15 degrees by shortening the length of the chains on one side of the panel. The No. 6 bars were inserted into the oversize holes provided on the transverse side of Panel P2; the panel was then lowered and moved horizontally. The installation process took about 120 sec- onds and went smoothly, with no need to change the tilt- ing angle of the panel during installation. • Panel P3, which had connection Detail BB, was installed vertically. • Plywood strips were used for wood forms on the bottom side of the panel-to-panel joints. The plywood strips were hung from the top surface of the panels using threaded rods. 43 Panel P1 Panel P2 Panel P3 Figure 45. Panel P1, P2, and P3 after 7 days of moist curing.

• The shear pockets were filled with SS Mortar through the 2 in. (50 mm) diameter tubes until the grout came out from the 1 in. (25 mm) diameter vent tubes on the far side of the pockets (7). The transverse shear key joints were filled with the SS Mortar grout. The grout had sufficient flowability to set without the need for any external vibrators. Figures 49 and 50 show some of the construction steps. When the grouting material reached the minimum required strength of 6.0 ksi (41.37 MPa), the test setup was built around the north transverse joint P1–P2, between Panel P1 and P2. The load was positioned in the transverse direction between the steel beams to produce the highest flexural effects, as shown in Figures 47, 51, and 52, where each of the two Neo- prene pads that support the load spreader beam were set 3 ft (0.914 m) from the centerline of the supporting steel beam. This arrangement provided a 6 ft (1.828 m) spacing between the Neoprene pads to simulate the LRFD HS20 truck. To investigate the effect of the fatigue load on the struc- tural behavior of the joint, the following actions were taken: • A series of strain gauges and displacement devices were installed around the joint on the top and bottom surface of the precast panels, as shown in Figure 52. First, the full fatigue load, 42.56 kip (189 kN), was applied as a static load, and the strain and displacement measurements were recorded with a data acquisition system. • The fatigue load, varying from 4.00 to 42.56 kip (17.8 to 189.3 kN), was applied for 2,000,000 cycles at 2 cycles per second. • A 3⁄4 in. (19 mm) deep water pool was built around the joint covering the full width of the bridge, as shown in Figure 51. The pool was kept full of water before and while the fatigue load was applied. Every 12 hours the bottom surface of the join was checked for water leakage. • The full fatigue load, 42.56 kip (189 kN), was then applied as a static load, and the strain and displacement measure- ments were collected. • These steps were repeated at south transverse joint P2–P3, between Panel P2 and P3. Test Results The clustered stud shear connectors did not obstruct the in- stallation of Panel P1 that was made with connection Detail A. The idea of building the grout forms of the transverse joints on the bottom surface of the panels worked very well. No leakage was observed as the joints were filled with grout, and no air voids were noticed on the top or bottom surface of the joints. The size and number of the grouting and venting ports of the shear pockets was sufficient to provide complete filling of the shear pockets and the haunch. No water leakage was detected before, while, or after the 2,000,000-cycle fatigue load was applied. No tension cracks were observed on the bottom surface of the transverse joints or the panels after the 2,000,000-cycle fa- tigue load was applied. This observation showed that no slip- page occurred to the spliced No. 6 (19) bar of Detail A and Detail BB. 44 0 1 2 3 4 5 6 7 8 9 10 0 7 14 21 28 35 42 49 56 Time (days) Co m pr es siv e str en gt h (ks i) SS Mortar Grout Concrete Mix Figure 46. Concrete strength gain versus time for the concrete mix and SS Mortar grout.

45 Load spreader beam W24x104, L= 8 ft Bottom reaction beam W18x56, L= 14 ft W18x119, L= 32ft Neoprene pad 22"x9"x2" 1'- 0" 4' -0 " 2' -4 " 8" 6'- 0" 2 in. diameter threaded bars 110-kips Hydraulic Actuator Top reaction beam 1'-0" 12x12x2" bearing plates 2" diameter heavy duty nut 8-in. thick precast panel 8' -0 " 8'- 0" 8'- 0" P1 P2 P3 3' -8 " 3' -8 " 9" 1'-10" 9'-8 1/2"9'-3 1/2" 1" 2'-2 1/2" 1'-4 1/2" 1" 2'-2 1/2" 1'-4 1/2" 1'-0" 9" 1'-10" 3 1/2"8 1/2" Detail A Detail BB 3-in. diameter holes 3'-0" 6'-0" 3'-0" 4'-0" 12'-0" 4'-0" 20'-0" Figure 47. Test setup.

46 1'-3" 11 1/2" 10 " 7 1/ 2" (b) 4- 2.5 in. diameter pipes Section A-A 1'-3" 11 1/2" 10 " 7 1/ 2" Perimeter of the cluster footprint (a) 8- 1.25 in. studs 8- 1.25 in. studs 9" 3" 15" 3" 2 1/2" A A 4- 2.5 in. pipes 9" 3" 15" 3" 2 1/2" B B 4 1/ 2" 1" Section B-B 4 1/ 2" 1" 5" Perimeter of the cluster footprint 8- 1.25 in. studs 4-2.5 in. pipes 5" (a) Panel P2 being lifted from the prestressing bed (b) Vertical installation of Panel P2 (c) Installation of Panel P1 (d) Installation of Panel P1 Figure 48. Arrangement of the four 21/2-in.-diameter pipes. Figure 49. Installation of the precast panels.

No signs of concrete crushing were observed at the top sur- face of the joint or the panels after the 2,000,000-cycle fatigue load was applied. No separation between the grout and the vertical surface of the shear key was observed. The strain measurements at the P1–P2 and P2–P3 trans- verse joints are summarized in Figures 52, 53, and 54. Table 5 summarizes the displacement measurement at both joints. Studying the strain and displacement measurements revealed the following: • The strain gauges oriented in the transverse direction (1uE, 8uE, 3uE, 6uE, 9uE, and 12uE) showed high stresses compared with the strain gauges oriented in the longitudi- nal direction (2uE, 7uE, 4uE, 5uE, 10uE, and 11uE). This observation confirms the logic that is used by the equiva- lent strip method of the LRFD specifications (7), where the deck slab is assumed to act as a one-way slab in the trans- verse direction. • Comparable gauges on the sides of each joint showed almost the same amount of transverse strains (compare 47 Figure 50. Grouting of the shear pockets and shear keys. Figure 51. Test setup at the north transverse joint (left) and the water pool around the joint (right).

1uE with 8uE, 3uE with 6uE, and 9uE with 12uE). This in- dicates that both joints (Detail A and BB) were able to transfer the full applied load. • The strain measurements on the north and south sides of the P1–P2 and P2–P3 joints were almost identical. This indicates that the structural behavior of the deck system was not affected by type of panel-to-panel connection, as long as the connection is capable of transferring the full load. • The stress and displacement measurements of both joints before and after the 2,000,000 cycles of fatigue load were almost the same, which indicates that no stiffness deterio- ration occurred as a result of the fatigue load. • Comparing the strain and displacement measurements of this test with those calculated using the equivalent strip method of the LRFD specifications (7) showed that the LRFD equation used to calculate the width of the equiv- alent strip leads to a conservative design, as it distributes the wheel load on a smaller distance than it should be, which results in higher flexural stresses. This observation may be due to the fact that the panels used in this test were transversely pretensioned. Transverse pretension- ing increases the panel stiffness, which causes the wheel load to be distributed on a wider strip. The effect of trans- verse pretensioning is not recognized by the LRFD spec- ifications, as the same equation is used to calculate the equivalent width of the strip for CIP and precast concrete slabs. Demolition of the Precast Panels To demolish the bridge, the transverse joints between Pan- els P1 and P2 and Panels P2 and P3 were saw cut. The small diameter of the blade did not allow for cutting through the full 8 in. thickness of the panel; therefore, only the top 6 in. (152 mm) of the joint was cut. This was sufficient to cut the No. 6 (19) spliced bars inside the joints. The center panel, P2, was then lifted by an overhead crane, which caused the bot- tom 2 in. (51 mm) of the joint to break. Investigation of the cut joints showed: (a) complete filling of the joint with grout, with no air voids, (b) no grout crushing, (c) no bond failure between the grout and the shear key of the panels, and (d) no bond failure between the grout and the No. 6 (19) spliced bars, as shown in Figure 55. 48 2u E P1 P2 1uE 3uE 4u E 11 uE 12uE 10 uE 9uE 1uE to 8uE: Top surface strain gauges 9uE to 12uE: Bottom surface strain gauges D1: Bottom surface displacement device (same arrangement was used at the second joint) D1 1' -0 " 2' -4 " 4'-0" 12'-0" 4'-0" 6uE 8uE 7u E 7" 7" 3" 3'-0" 6'-0" 3'-0" 9" 1'-10" 5u E 3" 3" 2'-1" 1'-10" 2'-1" 2'-1" 1'- 10" 2'-1" Figure 52. Locations of the measuring devices (uE  strain gauges, D  vertical displacement device).

Analytical Investigation of the Development Length of Confined Reinforcing Bars A concrete member’s strength can be significantly in- creased with the use of lateral confinement. Many researchers have investigated this technique over the past two decades. Saatcioglu et al. and Sun et al. provide a summary of various research activities conducted in this area (38, 39). Lateral con- finement can be provided by spiral reinforcement, as in the case of circular columns; circular steel tubes, as in the case of concrete-filled tube structures; or other shapes of structural steel, such as the HSS tubes that were used in this project. Lateral reinforcement produces lateral confining pressure on 49 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 0 15 30 45 60 75 90 120 135105 150 165 180 195 210 225 240 Time (seconds) M ic ro S tra in 1 uE 2 uE 3 uE 4 uE 5 uE 6 uE 7 uE 8 uE 9 uE 10 uE 11 uE 12 uE 3uE 6uE 8uE 1uE 12uE 7uE 9uE (dead) 2uE (a) Before applying the 2,000,000-cycle Fatigue Load -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 Time (seconds) M ic ro S tra in 1 uE 2 uE 3 uE 4 uE 5 uE 6 uE 7 uE 8 uE 9 uE 10 uE 11 uE 12 uE 3uE 6uE 8uE 1uE 12uE 7uE 9uE (dead) 2uE (B) After applying the 2,000,000-cycle Fatigue Load Figure 53. P1-P2 joint, connection Detail A.

50 -80.00 -70.00 -60.00 -50.00 -40.00 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 0 15 30 45 60 75 90 120 150 180 195165135105 210 240225 Time (seconds) M ic ro S tra in 1 uE 2 uE 3 uE 4 uE 5 uE 6 uE 7 uE 8 uE 9 uE 10 uE 11 uE 12 uE 3uE 8uE 12uE 9uE 6uE (dead) 1uE (dead) (a) Before applying the 2,000,000-cycle Fatigue Load -80.00 -70.00 -60.00 -50.00 -40.00 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 Time (seconds) M ic ro S tra in 1 uE 2 uE 3 uE 4 uE 5 uE 6 uE 7 uE 8 uE 9 uE 10 uE 11 uE 12 uE 12uE 9uE 6uE 3uE 8uE 1uE 11uE 5uE 2uE 7uE (B) After applying the 2,000,000-cycle Fatigue Load Figure 54. P2-P3 joint, connection Detail BB. Displacement (in.) Joint Before Applying the Fatigue Load After 2,000,000 Cycles of Fatigue Load P1–P2 0.0401 0.0390 P2–P3 0.0379 0.0388 Table 5. Displacement measurements at P1–P2 and P2–P3 joints.

51 (a) The Haunch fully filled with Grout with no Air Voids (b) Panels Stacked after Demolition P2 P3 P1 Detail A (P1-P2 connect ion) Detail A (P1-P2 connect ion) Detail BB (free side) Detail BB (P2-P3 conn ection) Detail BB (P2-P3 conn ection) Detail A (free side) (c) Condition of the Transverse Edges after Saw Cutting P1, Detail A (P1-P2 connection) P1, Detail A (free side) P3, Detail BB (P2-P3 connection) P2, Detail BB (P2-P3 connection) P2, Detail A (P1-P2 connection) Figure 55. Precast panels after demolition.

the concrete core, which significantly reduces the core ten- dency for internal cracking and increases the concrete com- pressive strength and ductility. Many mathematical models that describe the stress- strain relationship of confined concrete have been devel- oped (38). Among the latest models is the one presented by Sun et al. (39) that can be used for noncircular lateral confinement: fc0 = f0 + 4.1kfl (3) where fc0 = confined concrete strength, f0 = unconfined concrete strength (i.e., for concrete cylinders), k = a factor that relates the average lateral pressure fl to the equivalent uniform pressure (k can be taken = 1.0 for the case of using HSS tubes), fl = effective lateral confining pressure, (4) As = area of lateral confinement steel, fyh = confinement steel strength, s = pitch of lateral confinement, and bc = core dimension, center-to-center of perimeter of lat- eral confinement. Providing lateral confinement to the concrete core sur- rounding a reinforcing bar can significantly reduce its devel- opment length. As the reinforcing bar, which is in tension, tries to slip away from the concrete surrounding it, high longitudi- nal compressive stresses are created in the concrete. Because concrete is a semielastic material, the longitudinal compressive stresses force the concrete surrounding the bar to expand lat- erally, which may cause the concrete to split along the bar and the bar to then slip away from the concrete. The lateral con- finement resists the lateral expansion of concrete and protects it against splitting. Two approaches can be used to calculate the development length of steel reinforcement bars embedded in laterally con- fined concrete. The first approach is to develop a mathematical model through an experimental program. This method can be used for a specific type of lateral confinement, where a large num- ber of pullout specimens are tested for various variables that may affect the development length, such as bar size and con- crete strength. The mathematical model uses the unconfined concrete, , as a base for calculating the development length. This method provides an accurate estimate of the develop- ment length and a flexible model that can be easily adjusted for a wide range of variables. However, a large number of specimens need to be tested to get a reliable model. fc' = ∑ A f sb s yh c fc' The second approach is to use the development equation that is given by a code or specification for bars confined by regular stirrups, but replacing the unconfined concrete strength with the confined concrete strength, fc0, given in Equation 3. Then the reduced development length can be ver- ified through a limited number of pullout specimens. This method gives conservative estimates of the development length. The second approach was used in this project because the development length of reinforcing bars confined with HSS tubes had not been experimentally investigated before. The development length of the No. 6 (19) bars confined by an HSS 4 × 12 × 3⁄8 in. tube was estimated as follows: Step 1 Determine the confined concrete strength, fc0 fc0 = f0 + 4.1kfl = 6,000 + 4.1 × 1.0 × 6,750 = 33,675 psi (232.2 MPa) Step 2 Determine the development length of the No. 6 (19) bar using: Article 5.11.2.1.1 of the LRFD specifications (7) (5) where Ab = cross-sectional area of the bar = 0.44 in2, fy = bar yield strength = 60 ksi, and Equation 12-1 of the ACI318-05 (40) (6) where fy = bar yield strength = 60,000 psi, ψt = reinforcement location factor = 1.0, and ψe = reinforcement coating factor = 1.0. Check ψtψe = 1.0 < 1.7 ψs = reinforcement size factor = 0.8, λ = light weight concrete factor = 1.0, and (this is the upper limit specified by AC1318-05). c k d tr b +⎡ ⎣⎢ ⎤ ⎦⎥ = 2 5. l f f c k d dd y t e s c tr b b= +⎡ ⎣⎢ ⎤ ⎦⎥ 3 40 0 ψ ψ ψ λ ld = × × = 1 25 0 44 60 33 675 . . . 5.68 in. (144 mm) l A f f d b y c = 1 25 0 . f A f sb l s yh c = = × ×( )( ) ×( ) = ∑ 2 12 38 36 000 12 4 6 750 , , psi fc' 52

The ACI318-05 upper limit was recommended for this case because calculations of this term yielded a much higher value than 2.5 db = bar diameter = 0.75 in. Therefore, 6.0 in. (152 mm) development length of the No. 6 (19) bar in Detail A was used for the recommended sys- tem CD-1A. Step 3 Determine the lap splice length of the No. 6 (19) using: Article 5.11.5.3.1, LRFD specifications, (7) Lap splice length = 1.7ld = 1.7 × 5.68 = 9.66 in. (245 mm) Section 12.15, ACI318-05 (40) Lap splice length = 1.3ld = 1.3 × 5.88 = 7.65 in. (194 mm) Therefore, an 11.0 in. (279 mm) lap splice length of the No. 6 (19) bar was used in Detail BB, which was used in rec- ommended system CD-1B. Panel to Concrete Girder Connection Description of the Connection Detail Typically, the girder web reinforcement is extended outside the top surface of the girder and is embedded in the concrete slab to create full composite action. The maximum size of the girder web reinforcement is No. 5 (16) bar, and these bars are made of an L- or inverted U-shape to develop their yield strength at the interface. As a result of extending the maximum spacing between shear connectors to 48 in. (1220 mm), a large number of No. 5 (16) bars need to cluster and be made to fit into the shear pocket dimensions, which cannot be practically done. Also, the minimum bending diameter of the No. 5 (16) bar will require (a) increasing the girder web thickness if the in- verted U-bars are set in the transverse direction, (b) increasing the length of the shear pocket if the inverted U-bars are set in the longitudinal direction, or (c) significantly increasing the width of the shear pocket if the L-shaped bar is used (Figure 56). A new connection detail for creating full composite action for slab/concrete girders systems, such as system CD-1, was developed. The new connection detail minimizes the inter- ference between the horizontal shear reinforcement of the girder and the shear pockets of the panel. The new detail uses clusters of 11⁄4 in. (31.8 mm) diameter double-headed studs spaced at 4 ft (1220 mm). The studs in each cluster are spaced at 3 in. and embedded 81⁄2 in. (216 mm) in the concrete girder. The studs are made from SAE 1018 steel that is used ld = × × × × × × × 3 60 000 1 0 1 0 1 0 0 8 40 33 675 2 5 0 7 , . . . . , . . 5 5 88= . in. (149 mm) to make the 11⁄4 in. (31.8 mm) studs for steel girders (34). The top surface of the concrete girder is intentionally roughened to 1⁄4 in. (6 mm) amplitude. (The details of the connection are shown in Figure 28.) A precast NU I-girder is used in devel- oping this detail, for the following reasons: • The NU girder represents the most critical conditions en- countered with thin top flanges and webs. Web thickness of the NU girder is 5.9 in. (150 mm), and the top flange thickness is 2 7⁄8 in. (73 mm). • Most of the new series of I-girders developed in the United States, such as the Washington State Super Girder, the New England Bulb Tee, and the Iowa Bulb Tee, have almost the same features as the NU girders. To determine the amount of horizontal shear reinforcement, the design examples given by the PCI Bridge Design Manual were considered (35). Four design examples of slab/I-girder bridge systems are given in the design manual; the bridge struc- tures range from a simply supported span to three continuous span structures, with a span length up to 120 ft and girder spac- ing from 9 to 12 ft. Studying these examples revealed that the maximum horizontal factored shear force at the interface between the deck slab and the precast concrete girders is about 3.7 kip/in. (0.65 kN/mm) of the longitudinal direction of the girder. Therefore, the required horizontal nominal shear strength for a precast panel measuring 8 ft (2.44 m) long is Vn = (3.71 kip/in.) (8 × 12 in.)/(φ = 0.9) = 396 kip/panel (1761 kN/panel) For example, try three 11⁄4 in. (31.8 mm) diameter double- headed studs per pocket, with a cluster spacing of 48 in. (1220 mm) and with one stud per row. The studs are made from SAE 1018, 54 ksi (372 MPa) yield strength, 64 ksi (441 MPa) ultimate tensile strength steel. The pocket dimensions are 14 in. wide and 14 in. (356 mm) long. The shear friction theory (7) was used to design for the required reinforcement. The nominal shear resistance of the interface plane according to Equation 5.8.4.1-1 of the LRFD specifications (7) is Vn = c Acv + μ Avf fy (7) where c = cohesion factor, 0.1 ksi for concrete placed against clean, hardened concrete with surface intentionally roughened (LRFD specifications, Article 5.8.4.2), μ = friction factor, 1.0 for concrete placed against clean, hardened concrete with surface intentionally rough- ened (LRFD specifications, Article. 5.8.4.2), Acv = area of concrete engaged in shear transfer = (12 in. × 14 in.)(2 pockets) = 336 in2, 53

Avf = area of shear reinforcement crossing the shear plane (3 studs per pocket)(2 pockets) = 7.38 in2/panel, and fy = yield strength of the horizontal shear reinforcement, 54.0 ksi for SAE 1018 steel. Vn = (0.1 ksi)(336 in2) + 1.0(7.38 in2)(54 ksi) = 432.1 kip/panel (1922 kN/panel) > 396 kip/panel (1761 kN/panel) = × ⎛⎝⎜ ⎞⎠⎟3 14 1 25 4 2 . . Limits on Vn given by Equations 5.8.4.1-2 and 5.8.4.1.3 of the LRFD specifications (7) are not used here, as the shear pockets are confined with HSS tubes or closed ties that protect the grout surrounding the studs from crushing at the limits given by these equations. Experimental Investigation The shear friction theory depends on the assumption that the shear connectors will be able to develop their tensile yield strength. The axial tension force will be provided in the studs 54 (a) No. 5 Inverted U-shape Bar set Transversely (b) No. 5 Inverted U-shape Bar set Longitudinally (c) No. 5 L-shape Bar set Transversely 1'-3" 5 x 2 1/2"1 1/4" 1 1/4" K3 K3 1'-11 3/4" 5" 7 1/2" 7 1/2"4 1/2" 4 1/2" K2 K2 1'-3" 5 x 2 1/2"1 1/4" 1 1/4" K1 K1 No. 5 Inverted U-shape bar Section K1-K1 1" 1" 5 7/8" 8 3/8" Minimum required thickness 5" 1" 3" 1'-0" 5" NU-Girder Section K2-K2 1'-0" 5 7/8" 1" 1" 5" 1" 3" No. 5 Inverted U-shape bar NU-Girder Section K3-K3 5 7/8" 1'-10" 5" 1" 3" 1" 1" No. 5 L-shape bar5" 2 1/2" 1 5/8" NU-Girder Figure 56. Various options for setting the No. 5 bar.

once the deck slab starts to slide horizontally on the concrete girder. Due to the roughness of the top surface of the girder, the horizontal sliding of the deck slab will be accompanied by vertical separation at the interface. The head of the stud will resist the vertical separation, causing axial tension force in the studs and compression force in the concrete around the stud. The double-headed stud that is used in the new connection detail should be fully developed on both sides of the interface. On the girder side of the interface, the studs are embedded in thin elements with light reinforcement (i.e., the top flange and the web of the girder), which may not provide enough confinement to fully develop them. It was thus important to test this connection detail on the girder side to make sure that enough confinement is provided and that the studs can de- velop their tensile strength. Anchorage of the headed stud on the slab side was checked with the panel to steel girder con- nection discussed later in this chapter. Figures 57 and 58 provide details of the test specimens. The specimens were full-size top parts of an NU I-girder, and each specimen was made with one cluster of three 11⁄4 in. (31.8 mm) studs. Two groups of specimens were designed, with three spec- imens in each group. The first group of specimens was made with the exact amount of web reinforcement that is usually used with NU girders, which is No. 4 @ 4 in. (13 @ 102 mm) on each side of the web, as shown in Figure 57, while the second group of specimens was made with a higher amount of reinforcement in the web, as shown in Figure 58. The amount of web rein- forcement provided in the second group was determined based on matching the yield strength of the studs, as follows: Ultimate tensile force of three studs = (3 studs) (54 ksi) = 200.0 kip Yield strength of No. 4 @ 4 in., which is the typical web reinforcement of the NU girder, = (0.20 in2 per leg)(2 legs)(4 rows of reinforcement)(60 ksi) = 96.0 kip Use an additional two No. 4 (13) inverted U-shaped bars per stud = (0.20 in2 per leg)(2 legs)(6 rows of reinforcement)(60 ksi) = 144.0 kip Total yield strength of the web reinforcement in the vicinity of the three studs = 96.0 + 144.0 = 240.0 kip (1067 kN) > 200.0 kip (890 kN) Figure 59 shows the test setup, where one cluster of three studs was embedded in a full-size top part of an NU girder. The studs were tested in direct tension by anchoring them with a top reaction beam that was supported by two hy- draulic jacks. The specimen was tied to the strong floor using high-strength threaded rods. Two steel side forms, which are used in fabricating the NU I-girders, were borrowed from a precast concrete producer and used to fabricate the six specimens, as shown in Figure 60. π 4 1 252 2× ⎛⎝⎜ ⎞⎠⎟. in 55 4'-0 1/4" 1'-4 1/8" 5 7/8" 2 1/2" 1 1/8" 5 3/8" 2'- 0" 8 1/ 2" #4 @ 6 in. 8#5 6" 6" 6" 6" 6" 6" 35 " 6 1/ 2" 46" 8" 1 1/4 in. stud #4 @4 in. 4" 4" 4" 4" 4" 4" 4" #4 @ 4 in. #4 @ 4 in. #3 #3 1'- 10 " 1" 3" 3" Figure 57. Group 1 of the slab/concrete girder specimens.

A steel beam was used to temporarily support the 11⁄4 in. (31.8 mm) long studs and keep them perfectly vertical until the concrete gained sufficient strength to support the studs. A high-performance concrete mix with 8.0 ksi (55 MPa) speci- fied concrete strength at 28 days was used for all the speci- mens, as shown in Part b of Figure 60. The specimens were moist cured using wet burlap for 7 days. Concrete cylinders were made and cured by the specimens to monitor the strength gain with age. Figure 60 also shows the compressive strength gain with time. The specimens were anchored to the strong floor using 2 in. (51 mm) diameter high-strength threaded bars, as shown in Part d of Figure 60. Two synchronized hydraulic jacks, 300 kip (1334 kN) capacity each, and a stiff reaction beam were used to apply load on the studs. The load was applied at a rate of 300 kip/sec (1334 kN/second) until failure occurred. Test Results and Discussion Group 1 Specimens with Regular Web Reinforcement. At a relatively low load of about 90 kip, two horizontal hair cracks developed on the side surfaces of the specimen. These cracks were at the junction between the top flange and the vertical web of the specimen and were very close to the level of the head of the studs embedded in the flange. During this stage, no signs of failure were observed, and the top reaction beam was perfectly horizontal. Also, the recorded load from the two hydraulic jacks was almost identical. These signs gave a clear indication that the applied load was uniformly distributed between the three studs, and no stud slippage occurred. When the total applied load approached about 90 to 98 kip (400 to 436 kN), the side cracks started to widen and could be easily observed from a distance, as shown in Figure 61. Also, the recorded load from the two hydraulic jacks started to show a small difference, and the reaction beam started to lose its perfect horizontal alignment. These signs showed that the applied load was not perfectly distributed between the three studs. Note that the 90 kip (400 kN) applied load is about the maximum tensile capacity of web reinforcement provided in the specimen. When the applied load reached about 105 kip (467 kN), some cracks started to form on the top surface of the speci- men close to the exterior studs. The three studs with the con- crete surrounding them started to pull out of the concrete specimen. The top surface cracks continued to widen until failure occurred, as shown in Figure 61. The recorded failure loads of the three specimens were 116.4, 131.4, and 116.2 kip (517, 584 and 517 kN), with an average value of 121.0 kip (538 kN), which was about 61% of the yield capacity of the stud group. Failure occurred when the studs pulled out of the concrete specimen and a sudden drop in the recorded load was reported. The amount of stud slippage at failure ranged from 1.5 to 2.0 in. (38 to 51 mm). 56 4'-0 1/4" 1'-4 1/8" 5 7/8" 2 1/2" 1 1/8" 5 3/8" 2' -0 " 8 1/ 2" #4 @ 6 in. 8#5 6" 6" 6" 6" 6" 6" 35 " 6 1/ 2" 46" 8" 1 1/4 in. stud 4" #4 @4 in. 4" 4" 4" 4" 4" 4" #3 #3 1'- 10 " 1" 3" 3" #4 @ 4 in.#4 on each side of the stud 3" 1' -6 1 /8 " Figure 58. Group 2 of the slab/concrete girder specimens.

It was clear that failure started when the tensile stresses generated at the junction between the top flange and the ver- tical web exceeded the tensile strength of the provided web reinforcement of 90 kip (400 kN). However, the heads of the studs protected them from pulling out of the concrete spec- imen by applying compressive stresses on the concrete sur- rounding the stud stems. The compressive stresses confined the stud stems and made the studs and the concrete around them act as a unit that took the shape of an inverted pyra- mid. When the applied load reached 105 kip (467 kN), the web reinforcement of the girder yielded and started to show plastic deformation. This behavior was evident by the sud- den widening of the horizontal cracks on the sides of the specimen. Failure finally occurred when the concrete at the junction between the top flange and the web could not resist the tensile stresses generated. In one of the three tested specimens, the studs were com- pletely pulled out from the specimen by jackhammering the concrete around them, as shown in Figure 61. Inspection of the concrete around the studs found no air pockets were observed in the concrete specimen in the area around the studs, which indicated that although this area was congested 57 HSS 10x10x1/2 3'-0" 5'-0" 6'-0" 1 3/4" threaded rod placed in a 2" ID plastic tube & tied to the strong floor 2 1/4" 5 1/2" 2 1/4" 2'-1" 11 " 18 1/2" 44 0 ki p H J 44 0 ki p H J 43 1 /2 " 1 1/2"9" 4"4" 1 1/4 in. stud 2" 2 1/4" 1 1/4" Load cell 18.5" deep reaction beam A A 4'-0 1/4" 1'-4 1/8" 5 7/8" 2 1/2" 1 1/8" 5 3/8" 2'- 0" 8 1/ 2" 11 " 18 1 /2 " 5 1/ 2" 43 1 /2 " 2" 18.5" deep reaction beam 1 1/4 in. stud Section A-A Figure 59. Test setup of the slab/concrete girder specimens.

with heavy reinforcement, standard consolidation practices were able to remove air voids from the concrete. The inspec- tion also detected no crushing of the concrete in the vicinity of the studs and no permanent deformation on the stud head that was buried in the concrete. Group 2 Specimens with Additional Web Reinforce- ment. In general, the structural behavior of the specimens with additional web reinforcement was superior to that of specimens without additional reinforcement. The number and size of cracks was smaller, and the failure capacity was almost doubled. The first sign of cracking started to appear when the ap- plied load was about 150 kip (667 kN), where one hair crack was formed on each side of the specimen. These cracks were at the junction between the top flange and vertical web of the specimen, and very close to the level of the stud heads em- bedded in the flange. At this stage, no cracks were observed on the top surface of the specimen, and the reaction beam was in perfect horizontal alignment. Also, the recorded loads from the two hydraulic jacks were almost identical. These signs gave an indication that no slippage of any of the three studs occurred and that the applied load was uniformly dis- tributed between the studs. When the applied was about 190 kip, some minor hair cracks started to form on the top surface, and the hair cracks on the side surface started to open and became visible. It was clear that the inverted pyramid, made of the studs with the concrete surrounding them, was trying to pull out of the con- crete specimen. The top surface cracks continued to widen until failure occurred, as shown in Figure 62. The recorded failure loads for the three specimens were 215.4, 213.9, and 58 (a) The test specimens ready to receive concrete (b) Casting of concrete 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 0 5 10 15 20 25 30 Time (days) Co m pr es si ve s tre ng th (p si) (c) Compressive Strength Gain versus Time of the Concrete Mix (d) Test Setup Figure 60. Fabrication and test setup of the slab/concrete girder specimen.

203.6 kip (958, 952 and 907 kN), with an average value of 211.0 kip (938 kN), which is about 107% of the yield capacity of the stud group. Failure occurred when the studs pulled out of the concrete specimen and a sudden drop in the recorded load was reported. The amount of stud slippage at failure was about 1.0 in. (25 mm). The failure load was higher than the yield capac- ity of the studs and smaller than the ultimate strength capacity. It was clear that the additional web reinforcement signifi- cantly helped in anchoring the studs to the concrete specimen and fully developed their yield strength. Also, the number and size of side cracks was significantly reduced, compared with specimens of Group 1. Based on the test results, it is recom- mended to use stud yield strength of 54 ksi (372 MPa) to determine the number of shear connectors and to provide 59 (a) Location of the horizontal side surface cracks that started at about 105 kip (b) Top surface cracks at failure (c) Side surface cracks at failure (d) Studs after being pulled away from the specimen Figure 61. Structural behavior of Group 1 specimen of the slab/concrete girder.

additional web reinforcement in the vicinity of the stud clus- ters to achieve full composite action. As a result of the test setup, the top flange of the tested specimens in Group 1 and Group 2 was under longitudinal tensile flexural stresses that expedited failure. In real bridges, the top flange at the strength limit state is typically under compressive flexural stresses, which will help to confine the concrete around the studs and increase the failure load. Therefore, the results obtained from this test would be con- sidered conservative if compared with real bridge behavior. Analytical Investigation Finite element analysis was used to investigate the behav- ior of the slab/concrete girder pullout specimens. A commer- cial program, Nastran, was used in the analysis. The concrete specimen was modeled using the eight-node, cubic, three- dimensional element. Each node has three displacement de- grees of freedom, in the x, y, and z directions. The x direction is transverse to the girder longitudinal axis, the y direction is parallel to the girder longitudinal axis, and the z direction is parallel to the girder height. The following mechanical prop- erties were assigned to the concrete specimen: compressive concrete strength = 8 ksi (55 MPa), unit weight = 150 lb/ft3 (23.6 kN/m3), and Poisson ratio = 0.15. The 11⁄4 in. (31.8 mm) studs and web reinforcement bars were also modeled using the 20-node cubic element. The cir- cular cross-sectional area of the stud’s stem and head and the web reinforcement bars were replaced with the equivalent square cross-sectional area, as shown in Table 6. This simpli- fication helped to refine the mesh in the vicinity of the 11⁄4 in. (31.8 mm) studs. The following mechanical properties were assigned to the stud: tensile strength = 64 ksi (441 MPa), yield strength = 54 ksi 60 (a) Side surface cracks at failure (b) Top surface cracks at failure Figure 62. Structural behavior of Group 2 specimen of the slab/concrete girder.

(372 MPa), unit weight = 490 lb/ft3 (76.9 kN/m3), and Poisson ratio = 0.30. The following mechanical properties were assigned to the vertical web reinforcement: yield strength = 60 ksi (414 MPa), unit weight = 490 lb/ft3 (76.9 kN/m3), and Poisson ratio = 0.3. Details of the finite element model are given in Appendix F. Each stud was loaded with a tensile axial force equivalent to the stud yield capacity—66.4 kip (295 kN). This load was applied as a surface load uniformly distributed on the stud cross-sectional area—54 ksi (372 MPa). To study the internal stress concentration around the studs, three sections were chosen, as shown in Figure 63. Sec- tion 1-1 is at the free side of the external stud, Section 2-2 is at the mid distance between two adjacent studs, and Section 3-3 is at the centerline of the center stud. Appendix F gives the z direction and principal stress distributions for these three sections, as well as the principal stress distribution on the top and side surfaces of the specimens. Studying these fig- ures reveals the following: • For Group 2 specimens, the additional web reinforcement helped in widening the base area of the inverted pyramid, which resulted in a lower stress concentration at the junc- tion between the top flange and the web. This observation is consistent with the experimental program results, where the size of the side crack at failure was wider for the Group 1 specimens than for Group 2 specimens. • The additional web reinforcement helped to distribute the tension force provided by the studs on a wider and deeper volume, resulting in reduced stress concentrations around the studs. This can be seen from the following observations: – Stress concentration at the flange-to-web junction in the Group 1 specimens is higher than that of the Group 2 specimens. – The concrete stress in the vicinity of the stud’s stem in the Group 1 specimens is higher and extends deeper than that of the Group 2 specimens • The stress distribution at Section 3-3 (in the z direction or principal stress) shows that the proposed 18 in. (457 mm) embedment of the additional web reinforcement is quite enough to develop its yield strength. The high tensile stresses generated in concrete between adjacent rows of additional web reinforcement do not extend to the bottom surface of the concrete specimen. This finding is consistent with the experimental test results, where no signs of slip- page or vertical side-surface cracking parallel to the addi- tional web reinforcement were observed. • The principal stresses at all sections are higher than the z- direction stresses due to the specimen setup that puts the top flange of the specimen in tension. • The compressive stress at the flange-web junction is about 2.0 ksi (14 MPa), which is less than the concrete bearing strength, 0.85 × 8 ksi = 6.8 ksi (47 MPa). This observation is consistent with the test result as no concrete crushing in this location was observed at failure. It is believed that the web reinforcement helped to confine the concrete and consequently protected it from premature cracking. Panel to Steel Girder Connection Steel studs welded to the top surface of steel girders and embedded in the concrete slab have been the typical tech- nique used to create full composite action for slab/steel girder construction (2, 34). The 3⁄4 in. (19 mm) and 7⁄8 in. (22 mm) diameter studs have been the common sizes used in bridges. Recently, a 11⁄4 in. (31.8 mm) diameter stud was developed by a group of researchers at the University of Nebraska (34). The stem of the 11⁄4 in. (31.8 mm) diameter stud has double the cross-sectional area of a 7⁄8 in. (22 mm) diameter stud. There- fore, one 11⁄4 in. (31.8 mm) stud replaces two 7⁄8 in. (22 mm) studs. There are many advantages to using the 11⁄4 in. (31.8 mm) stud, including (a) higher speed of construction, as a smaller 61 Actual Diameter (in.) Cross Sectional Area (in.2) Equivalent Square Area (in. x in.) Stud Stem 1.25 1.227 1.108 x 1.108 Stud Head 2.5 4.909 2.216 x 2.216 No. 4 Bar 0.5 0.200 0.447 x 0.477 Table 6. Dimensions of the equivalent square area used for finite element analysis. Group #1 Specimen Group #2 Specimen 1 2 3 1 2 3 Figure 63. Location of Sections 1, 2, and 3.

number of studs are welded; (b) less congestion of the girder top flange, especially in areas of high horizontal shear stresses; (c) easier deck removal; and (d) less damage to the girder top flange during deck removal. The 11⁄4 in. (31.8 mm) stud has been successfully used in bridges in some of the Midwest states (36, 41, 42). The use of the 11⁄4 in. (31.8 mm) studs with precast concrete panels adds another advantage, as the shear pocket dimensions are reduced by about 40%, resulting in a smaller volume of grout to be used and a more economical system. As discussed in Chapter 2 of this report, the maximum spac- ing between shear connectors is 24 in. (610 mm) in the LRFD specifications (7). Investigation of the background of this limit revealed that a very limited amount of testing was conducted with stud spacing greater than 24 in. (610 mm). In addition, the majority of these tests were made for CIP slabs, where the studs are uniformly spaced across the specimen and not clus- tered in groups, as is the case with precast panel construction. Recently, two attempts have been made to address the issues of clustering the studs in groups for precast panels and extend- ing the 24 in. (610 mm) maximum spacing to 48 in. (1220 mm) (31, 32). (A brief summary of those attempts is given in Chap- ter 2 of this report.) The first attempt (31) focused only on the effect of the number of studs and their orientation per cluster on the ultimate capacity, while the second attempt focused only on the effect of extending the maximum spacing limit to 48 in. (1220 mm) on the fatigue capacity (32). Study of these attempts revealed the following: • Extending the maximum spacing to 48 in. (1220 mm) has no negative effect on the fatigue capacity of clustered 7⁄8 in. (22 mm) studs. • Clustered studs may not be able to produce their ultimate capacity as a result of premature crushing failure of the grout surrounding the studs or premature failure of the concrete slab surrounding the shear pocket. • None of these attempts was able to simultaneously investi- gate fatigue and ultimate capacity of clustered studs. • Both attempts used 3⁄4 in. (19 mm) and 7⁄8 in. (22 mm) di- ameter studs. A review of the literature found that the fatigue and ulti- mate capacities of shear studs were studied individually, which means that the effect of the fatigue load on the ultimate stud capacity was not investigated. In a real bridge, there is a fair chance that the studs will be exposed to a large number of live load cycles before the bridge is overloaded and the studs are loaded up to their maximum strength. Description of the Connection Detail As discussed earlier in this chapter, recommended system CD-1 uses clusters of eight 11⁄4 in. (31.8 mm) studs spaced at 48 in. (1220 mm). The number of studs per cluster was determined based on the parametric study conducted by Tadros and Baishya, in which a large number of slab/steel girder bridges, with spans ranging from 60 to 130 ft (18.2 to 39.6 m) and girder spacing ranging from 6 to 12 ft (1.82 to 3.66 m), were analyzed (2). The study revealed that the maximum horizontal shear stress at the interface required one 11⁄4 in. (31.8 mm) stud set at 6.0 in. (152 mm) spacing. To prevent the grout surrounding the studs from prema- ture cracking due to the high compressive stresses generated by the stud group, the shear pocket was confined with an HSS tube, as shown in Figures 20 to 30. Another alternative for confining the grout that was considered in the experi- mental investigation was using three individual No. 6 (19) closed ties. A 2 in. (50 mm) clear concrete cover was main- tained on the lower tie, and a 1 in. (25 mm) clear spacing was maintained between the ties. This arrangement resulted in setting the tie group as close as possible to the bottom surface of the panel, where the bearing stresses of the studs on the grout reach their highest value close to the base of the stud. This finding was revealed by the finite element analy- sis of the push-off specimens that will be discussed later in this chapter, and it was also confirmed by other researchers (43, 44). Two options for manufacturing the 11⁄4 in. (31.8 mm) stud were investigated: (a) produce a headed stud where the head is made integral with the stud stem, and (b) produce a headless stud with a heavy-duty nut and washer to form the stud head. The two options were investigated with three stud manufacturers located in different states; it was found that the first option would reduce the cost of making the stud and save time and effort required to install the heavy nut. But producing the headed stud requires a special forging ma- chine that may not be available at every stud manufacturer. The headed stud was used for the push-off specimens, while the headless stud with a heavy-duty nut and washer was used for the full-scale beams. Figure 64 shows the dimensions of the headed and headless 11⁄4 in. (31.8 mm) diameter studs. The weight of the 11⁄4 in. (31.8 mm) headed stud was 2.37 lb, compared with 1.10 lb for a 7⁄8 in. (22 mm) stud. SAE 1018 steel was used to make both the headed and the headless studs. To validate the proposed concept of extending the maxi- mum stud spacing to 48 in. (1220 mm) and to study the effect of fatigue load on ultimate capacity, the following activities were conducted: • Push-off specimens: Group 1 was tested directly for ulti- mate capacity and Group 2 was exposed to 2,000,000 cycles of fatigue load and then tested for ultimate capacity. • Full-scale beam testing: Two full-scale beams were tested. The first beam was made with clusters of four 11⁄4 in. studs 62

spaced at 24 in., and the second beam was made with clus- ters of eight 11⁄4 in. studs spaced at 48 in. Push-Off Specimens Description of the Push-Off Specimens Two groups of push-off specimens were fabricated and tested. Group 1 consisted of eight specimens tested for ulti- mate capacity. Group 2 consisted of eight specimens exposed to 2,000,000 cycles of fatigue load and then tested for ulti- mate capacity. Table 7 gives the design criteria for these spec- imens. Figures 64 to 73 show the details of the specimens. Figure 74 shows the specimens during fabrication. The specimen details of both groups are identical, with the following exceptions: • The specimens of Group 1 were made with a 11⁄4 in. (31.8 mm) thick haunch between the concrete specimen and the steel plate, while the specimens of Group 2 were made with- out a haunch. The haunch was eliminated in the second group of specimens in order to compare the test results with the ultimate capacity as given by the LRFD specifications (7) and other sources, such as Ollgaard et al. (45), Oehlers and Bradford (43), and Viest (46), where the equations were de- veloped using a symmetric specimen with no haunch pro- vided in the push-off specimens. Symmetric specimens are typically made with a steel beam with studs welded on both flanges and with a concrete prism on each side of the steel beam. The symmetric specimen could not be used in this re- search because a very high load would be required to break a specimen with sixteen 11⁄4 in. (31.8 mm) studs, which was beyond the capability of the testing facility. • External confinement was added to the specimens of Group 2 by two side plates attached to the specimens. The plates were anchored by 1⁄2 in. (12.7 mm) diameter threaded bars and nuts. The threaded bars were embedded in the specimens and extended 3 in. (76 mm) outside the specimen on each side, as shown in Figure 73. The external confinement was added to simulate real bridge deck sys- tems where the slab has extended length on both sides of the girderline to help confine the shear pockets. This tech- nique was successfully used during the development of the 11⁄4 in. (31.8 mm) diameter studs (42, 47). The standard welding gun used to weld the 7⁄8 in. (22 mm) studs was also used to weld the headed 11⁄4 in. (31.8 mm) studs. A special chuck that can fit the headed 11⁄4 in. (31.8 mm) stud was fabricated and used, as shown in Figure 75. The studs of first and second specimens of Group 1 were welded using a tri-legged support to adjust the verticality of the studs, as shown in Figure 75. Once the technician gained enough confidence in the welding process, however, he shot 63 1/2" 1 1/4" 3/16 in. diamter flux ball 1 1/ 8" 2 1/8" 3" Heavy duty nut 3/16 in. thick washer 2 1/2" 1/2" 4 1/4" 1/2" 5 1/4" 1/2" 1 1/4" 3/16 in. diamter flux ball Headed stud Headless stud Figure 64. Dimensions of the 11/4-in.-diameter stud. Push-Off Specimen Number of Specimens Number of Studs per Specimen Type of Grout Confinement Test Type Group 1 P-4-CT-U 2 4 3 No. 6 closed ties (CT) P-4-ST-U 2 4 Steel tubes (ST) P-8-CT-U 2 8 3 No. 6 closed ties (CT) P-8-ST-U 2 8 Steel tubes (ST) Ultimate (U) Group 2 P-4-CT-F/U 2 4 3 No. 6 closed ties (CT) P-4-ST-F/U 2 4 Steel tubes (ST) P-8-CT-F/U 2 8 3 No. 6 closed ties (CT) P-8-ST-F/U 2 8 Steel tubes (ST) Fatigue/Ultimate (F/U) Table 7. Design criteria of the push-off specimens.

the rest of the studs without using the tri-legged support. The studs were welded using a direct current power supply of 2,600 A. The welding was successful at an average rate of 1.8 sec/stud. The quality of the stud welding was checked using the following three measures: • Visual inspection. The weld was visually inspected to make sure that the melted material formed a complete and uni- form flash (i.e., dam or weld collar) at the base of the stud with no flaws, as shown in Figure 75. Also, the bottom sur- face of the 1 in. (25 mm) thick steel plate was inspected to make sure that the generated heat did not melt the full thickness of the plate. • Bending the stud to 45 degrees. Most of the state agency specifications require that a stud be bent 45 degrees with- out failure. Figure 75 shows the 11⁄4 in. (31.8 mm) stud was successfully bent to 45 degrees with no sign of failure at the base. • Using a portable hydraulic jacking device. A portable hy- draulic jacking device that could be used in the field or in the shop was developed (34, 42). The device consists of two collars placed around two adjacent studs, a small hydraulic jack, and a top tie, as shown in Figure 75. The collar con- sists of two steel blocks tied together with four screws. By tightening the four screws, the collar is placed in full con- tact with the 11⁄4 in. (31.8 mm) stud. The base of the collar is recessed to accommodate the weld at the stud base. A compact 100 kip (445 kN) hydraulic jack is placed between the collars to provide lateral shearing force at the stud base. The top tie, which consists of two plates and two threaded rods, is used to protect the studs from bending and to pro- tect the technicians during the test. The quality control test is conducted by applying a horizontal force that would cause an axial tension failure in the stud. This force can be calculated by analyzing the studs with the top tie as a closed frame action, where the studs are fixed at their base and hinged at the top. The device was successfully used to test studs used in the experimental program. An 85 kip (378 kN) force was applied on two adjacent studs, and no signs of failure were observed at the stud base. 64 Section B-B Plan View C-C 1'-3 5/8" 1 9/16" 1'-0 1/2" 1 9/16" 2" 2' -0 " Base plate 1-in. thick 5 sp ac in gs @ 4" = 2 0" 2" 8" 1.25" 1" 1'-0 1/2" A A 10 " 10 " 1' -8 " 3'-8" B 2'-0" B C C 4" 1'-4" A A Section A-A 2' -0 " 7" 7" 6 1/2" 11" 6 1/2" 7" 6 1/ 2" 11 " 6 1/ 2" 6 1/2" 11" 6 1/2" 9" 4" 1'-4" 3" 3" 6 1/2" 11" 6 1/2" 5" 3-#4 closed ties Figure 65. Concrete dimensions of P-4-CT-U.

A normal weight concrete mix with 6 ksi (41 MPa) speci- fied concrete strength was used to make the specimens. The shear pockets of the specimens were filled with an SS Mortar mix containing 50% pea gravel (1⁄4 in., 6 mm, diameter). Fig- ure 76 shows the compressive strength gain with age of the concrete and grout mixes. The specimens were tested using a horizontal self-equilib- rium frame, as shown in Figure 77. Because a nonsymmetric specimen was used, it was expected that the specimen would move upward at the bearing end of the specimen where the load was applied, which would lead to a premature and unrealistic failure. Therefore, a steel frame was built around the bearing area of the specimen, as shown in Figure 77. The steel frame was provided with roller supports to allow for horizontal sliding of the specimen. All specimens were tested when the grout was 28 days or older. The specimens of Group 1 were tested by applying the load at mid height of the 8 in. (203 mm) thick slab at a rate of about 5 kip (22 kN) per second. The relative horizontal movement between the steel plate and the concrete specimen was recorded with a linear variable displacement transducer (LVDT). The specimens of Group 2 were tested using the following steps: (a) the specimen was loaded with a static load equal to the fatigue capacity of the stud group as determined by the LRFD specifications (7), and the relative horizontal move- ment between the steel plate and the concrete specimen was recorded; (b) the specimen was exposed to 2,000,000 cycles of fatigue load, and the upper limit of the fatigue load was the fatigue capacity of the stud group as determined by the LRFD specifications (7), and the lower limit was 5 kip (22 kN) to maintain equilibrium of the specimen; and (c) the upper limit of the fatigue load was applied as a static load, and relative horizontal movement was recorded. At all steps, the load was applied at mid height of the 8 in. (203 mm) thick slab. Fatigue and Ultimate Capacities of Steel Studs Fatigue Capacity. The fatigue capacity was estimated in accordance with Equation 6.10.10.2-1 of the LRFD specifica- tions (7). No other model of the of the fatigue capacity was considered in this study because the literature review revealed 65 10 " 10 " 1' -8 " 4" 1'-4" 3'-8" 8" 1.25" 1" 2'-0" 1'-0 1/2" B B A 5 3/4" 1'-0 1/2" 5 3/4" 5 3/4" 1'-0 1/2" 5 3/4" A Section B-B Plan View C-C 1'-3 5/8" 1 9/16" 1'-0 1/2" 1 9/16" 2" 2' -0 " Base plate 1-in. thick 5 sp ac in gs @ 4 " = 2 0" 2" Section A-A C C 2' -0 " 7 1/2" 9" 7 1/2" 4" 1'-4" 5 3/ 4" 1' -0 1 /2 " 5 3/ 4" HSS 12.50X0.188 9" 3" 3" 5" Figure 66. Concrete dimensions of P-4-ST-U.

that this equation gives a fair estimate for all sizes of studs used on bridges, including the 11⁄4 in. (31.8 MPa) stud. (8) (English Units) α (ksi) = 34.5 – 4.28 log (N) (9) (English Units) where Zr = fatigue resistance force of shear connector (kip), d = stud diameter (in.) = 1.23 in., and N = number of cycles. For 2,000,000 cycles and 11⁄4 in. stud: α = 7.53 ksi > ksi, and Zr = 7.53 × 1.252 = 11.77 kip/stud Four-stud cluster: Zr = 4 × 11.77 = 47.08 kip (209 kN) Eight-stud cluster: Zr = 8 × 11.77 = 94.16 kip (418 kN) Ultimate Capacity. The following sources were used to estimate the ultimate capacity of the stud group. 5 5 2 . Z d dr = ≥α 2 2 5 5 2 . • The design equation developed by Viest (46) for studs with diameter greater than 1.0 in. (25 mm): (10) (English Units) where Qcr = critical load (lb) ds = stud diameter (in) = 1.25 in. = compressive strength of the grout mix surround- ing the stud (ksi) = 9.6 ksi Four-stud cluster: Qcr = 4 × 48.4 = 193.6 kip (861.1 kN) Eight-stud cluster: Qcr = 8 × 48.4 = 387.2 kip (1722.2 kN) This equation was considered in this research because it was the only equation that was found in the literature that was developed for studs with a diameter greater than 1.0 in. (25 mm). It was also reported by Issa et al. (31) that this equa- tion correlates well with test results when it was used to de- termine the ultimate capacity of studs clustered in groups. Qcr = × × =5 1 25 9 6 4 0 9 6 48 42. . . . . kip. fc' Qcr s c cd f f= 5 4 02 ' '. / 66 Section B-B Plan View C-C 1'-3 5/8" 1 9/16" 1'-0 1/2" 1 9/16" 2" 2' -0 " Base plate 1-in. thick 5 sp ac in gs @ 4 " = 2 0" 2" 8" 1" 1'-0 1/2" A A 10 " 10 " 1' -8 " 3'-11" B 2'-0" B C C 4" 1'-4" A A Section A-A 2' -0 " 1'-1" 7" 1'-1" 1'-5" 6 1/2" 3" 6 1/2" 11" 6 1/2" 3 1/2" 6 1/ 2" 11 " 6 1/ 2" 6 1/2" 1'-5" 3 1/2" 4" 1'-4" 2'-0" 9" 3-#6 closed ties 3" 3" 3" Figure 67. Concrete dimensions of P-8-CT-U.

• The design equation developed by Ollgaard et al. (45): This equation was developed using statistical analysis of push- off specimens, where the slab had not failed prematurely through splitting. (11) (English Units) where Dmax = critical load (kip) Ash = cross-sectional area of 11⁄4 in. stud (in2) = 1.23 in2 = compressive strength of the grout mix surround- ing the stud (ksi) = 9.6 ksi Ec = modulus of elasticity of the grout mix surround- ing the stud (ksi) (12) (English Units) where wc = unit weight of grout mix surrounding the studs (kcf) = 0.145 kcf Dmax = 1.1 × 1.23(9.6)0.3(4,463)0.44 = 119.3 kip/stud = ( ) =33 000 0 145 9 6 5 6461 5, . . ,. ksi = 33 000 1 5, . 'w fc c fc' D A f Esh c cmax = ( ) ( )1 1 0 3 0 44. ' . . Four-stud specimens: Dmax = 4 × 119.3 = 477.2 kip (2122.6 kN) Eight-stud specimens: Dmax = 8 × 119.3 = 954.4 kip (4245.2 kN) In the four- and eight-stud specimens, the estimated shear capacity is greater than the ultimate tensile capacity. • The design equation developed by Oehlers and Johnson (48): Using an approach similar to that used by Ollgaard et al., Oehlers and Johnson developed the following equation for the maximum shear capacity of steel studs: (13) (English Units) where Dmax = critical load for push-off specimens per stud (kip) Ash = cross-sectional area of 11⁄4 in. studs per group (in2) = 4 × 1.23 or 8 × 1.23 in2 fu = ultimate tensile strength of the stud material (ksi) = 64 ksi D n A f f f E E sh u c u c s max = − ⎛⎝⎜ ⎞⎠⎟ ⎛ ⎝⎜ ⎞ ⎠⎟5 3 1 3 0 35 . . ' . ⎛⎝⎜ ⎞⎠⎟ 0 40. 67 Section B-B Plan View C-C 1'-3 5/8" 1 9/16" 1'-0 1/2" 1 9/16" 2" 2' -0 " Base plate 1-in. thick 5 sp ac in gs @ 4 " = 2 0" 2" 8" 1.25" 1" 1'-3 5/8" A A 10 " 10 " 1' -8 " 3'-11" B 2'-0" B C C 4" 1'-4" A A Section A-A 2' -0 " 1'-1" 1'-1" 1'-4" 7" 4" 2'-0" 7" 9" 6" 1'-0" 6" 6" 1' -0 " 6" HSS 16x12x5/16 3" 3" 3" 3" 7" 1'-4" 4" 4" 1'-4" Figure 68. Concrete dimensions of P-8-ST-U.

= compressive strength of the concrete surround- ing the stud (ksi) = 9.6 ksi Ec = modulus of elasticity of the concrete (ksi) = 5,645 ksi Es = modulus of elasticity of the stud material (ksi) = 29,000 ksi n = number of studs per group = 4 or 8. Four-stud specimens = 391.7 kip (1742.3 kN) Eight-stud specimens = 815.4 kip (3626.9 kN) In both the four- and eight-stud specimens, the estimated shear capacity is greater than the ultimate tensile capacity. • Equation 6.10.10.4.3-1 of the AASHTO LRFD specifications (7): This equation was derived from the equation developed by Ollgaard et al. (45) after changing the exponents of , Ecfc' Dmax = − ⎛⎝⎜ ⎞⎠⎟ ×( )( )⎛⎝⎜ ⎞⎠⎟5 3 1 3 8 8 1 23 64 9 6 64 0 . . . . . . , , 35 0 40 5 645 29 000 ⎛⎝⎜ ⎞⎠⎟ Dmax = − ⎛⎝⎜ ⎞⎠⎟ ×( )( )⎛⎝⎜ ⎞⎠⎟5 3 1 3 4 4 1 23 64 9 6 64 0 . . . . . . , , 35 0 40 5 645 29 000 ⎛⎝⎜ ⎞⎠⎟ fc' to make the equation dimensionally correct and limiting the shear capacity by the ultimate tensile capacity of the stud. (14) (English Units) where Qn = nominal capacity (kip) Asc = cross-sectional area of 11⁄4 in. stud (in2) = 1.23 in2 = compressive strength of the concrete surrounding the stud (ksi) = 9.6 ksi Ec = modulus of elasticity of the concrete surrounding the stud (ksi) = 5,645 ksi Fu = ultimate tensile strength of the stud material (ksi) = 64 ksi. Four-stud cluster: Qn = 4 × 78.7 = 314.8 kip (1400.2 kN) Eight-stud cluster: Qn = 8 × 78.7 = 629.6 kip (2800.4 kN) • Equation 5.8.4.1-1 of the LRFD specifications (7): This equation is derived from the shear friction theory and is commonly used for the design of horizontal shear rein- forcement for slab/concrete girder composite beams. Qn = × × = × = least kips0 5 1 23 9 6 5 645 143 2 1 23 64 . . . , . . 78 7 78 7 . . kips kip/stud= fc' Qn sc c c sc uA f E A F= ≤0 5. ' 68 Section B-B Plan View C-C 8" 1'-0 1/2" A A 10 " 10 " 1' -8 " 3'-8" B B C C 1'-4" A Section A-A 2' -0 " 7" 7" 10 1/2"11"6 1/2" 7" 6 1/ 2" 11 " 6 1/ 2" 10 1/2"11"6 1/2" 6 1/2" 6 1/2" 1'-4" 11" 7" 3" 3" 1'-0" 4 3/4" 3 1/4" 1" 1'-4 1/4" 1' -0 " 2" 4" 4" 2" 6 3/4" 3" 6 3/4" 4 1/ 2" 3" 4 1/ 2" 1'-4 1/2" 2" 1'-0 1/2" 2" 1-in. thick base plate A 3-#4 closed ties 4 3/ 4" Figure 69. Concrete dimensions of P-4-CT-F/U.

However, the LRFD specifications (7) give values for c and μ if steel beams are used. Using Equation 7 gives Vn = (0.025 ksi)(113 in2) + 0.7(4.92 in2)(54 ksi) = 188.8 kip (839.8 kN) (four-stud specimens) Vn = (0.025 ksi)(192 in2) + 0.7(9.84 in2)(54 ksi) = 376.8 kip (1676.0 kN) (eight-stud specimens) Limits on Vn given by Equations 5.8.4.1-2 and 5.8.4.1-3 of the LRFD specifications (7) are not used here as the shear pockets are confined with HSS tubes or closed ties, which protect the grout surrounding the studs from crushing at the limits given by these equations. Table 8 summarizes the ultimate capacity using various sources. Test Results and Discussion The test results of Groups 1 and 2 are summarized in Table 9 and Figures 78 to 81. Table 9 gives the failure load and the mode of failure for all the specimens. This table also gives the failure load, Ff, as a per- centage of estimated ultimate capacity according to Viest (46), Ollgaard et al. (45), Oehlers and Johnson (48), and Equations 6.10.10.4.3-1 and 5.8.4.1-1 of the LRFD specifications (7). Figure 78 shows the failure modes of Group 1 specimens. Figure 79 shows the load-displacement relationship of Group 1 specimens when they were tested for ultimate capacity. Figure 80 gives the load-displacement relationship of Group 2 specimens due to fatigue load before and after ap- plying the 2,000,000 cycles of fatigue load. Figure 81 shows the failure mode of Group 2 specimens. Fatigue Capacity of Clustered Studs. No signs of con- crete/grout crushing, weld failure, or local distress around or inside the shear pockets were observed when the push-off specimens, with four and eight 11⁄4 in. studs, were exposed to 2,000,000 cycles of fatigue load. Also, as shown in Figure 80, there was almost no change in the load-displacement rela- tionship of the push-off specimens after applying the 69 10 " 10 " 1' -8 " 4" 1'-4" 3'-8" 3" 8" 1'-0" 1'-0 1/2" B B A A Section B-B Plan View C-C Section A-A C C 2' -0 " 1'-4" HSS 9.0X0.188" 9" 1" 3 1/4" 4 3/4" 7 1/ 2" 9" 7 1/ 2" 1'-4"1'-0" 3" 7 1/2" 9" 7 1/2" 7 1/2"9"7 1/2" 1' -0 " 2" 4" 4" 2" 6 3/4" 3" 6 3/4" 4 1/ 2" 3" 4 1/ 2" 1'-4 1/2" 2" 1'-0 1/2" 2" 1-in. thick base plate Figure 70. Concrete dimensions of P-4-ST-F/U.

2,000,000 cycles. This observation is consistent with the fa- tigue test that was conducted on full-scale beams tested later in this research, and the fatigue test results of a half-scale beam tested by Markowski et al. (32). The research team strongly believes that this equation can be satisfactorily used for the de- sign of composite beams made with clusters of eight 11⁄4 in. (31.8 mm) studs spaced as much as 48 in. (1220 mm) apart. Ultimate Capacity of Clustered Studs. Comparing the test results of the four-stud and eight-stud specimens shows that, regardless of the type of confinement used around the stud group, the ultimate capacity did not proportionally in- crease when the number of studs was doubled. Regardless of the number of studs, the ultimate capacity of a stud group confined with the steel tube is about 5% to 15% higher than the ultimate capacity of the same stud group confined with individual closed ties. The difference is more pronounced with the four-stud group than with the eight-stud group. Regardless of the number of studs per group and the type of stud confinement, Equation 6.10.10.4.3-1 of the AASHTO LRFD specifications (7) overestimated the ultimate capacity by as much as 50%. The same observation applies to the equations developed by Ollgaard et al. (45) and Oehlers and Johnson (48), where the ultimate capacity is overestimated by as much as 60%. For push-off specimens tested directly for ultimate capac- ity, Equation 6.10.10.4.3-1 of the AASHTO LRFD specifica- tions (7) and the equation developed by Viest (46) correlate very well with the test results. This observation is consistent with the findings of Issa et al. (31) that were obtained from testing of quarter-scale symmetric specimens made with two, three, and four 7⁄8 in. (22 mm) stud groups. Comparison between the test results of Group 1 and Group 2 in Table 9 shows that the 2,000,000 cycles of fatigue load reduced the ultimate capacity by about 5% to 18%. The reduction is more pronounced with (a) stud groups confined with closed ties than those confined with steel tubes, (b) spec- imens made with eight studs than those made with four studs, and (c) Equation 6.10.10.4.3-1 of the AASHTO LRFD specifications (7) and the equation developed by Viest (46) than the other three equations. A bond failure between the lower tie and the concrete slab was observed in most of the specimens made with closed in- 70 1-in. thick base plate Section B-B Plan View C-C 8" A A 10 " 10 " 1' -8 " 3'-11" B B C 1'-4" A Section A-A 2' -0 " 1'-1" 7" 1'-5"6 1/2" 6 1/2" 11" 6 1/2" 6 1/ 2" 11 " 6 1/ 2" 2'-0" 1" 3" 1'-4 1/2" 7" 6 1/2" 1'-5" 7 1/2" 1'-4" 1'-1" 1'-8" 9" 4 3/4" 3 1/4" 7 1/2" C A 1' -8 " 2" 4" 4" 4" 4" 2" 1'-4 1/2" 2" 1'-0 1/2" 2" 5 1/ 2" 9" 5 1/ 2" 6 3/4" 3" 6 3/4" 3-#6 closed ties Figure 71. Concrete dimensions of P-8-CT-F/U.

dividual ties and subjected to the 2,000,000 cycles of fatigue load. It is believed that this failure occurred because of the large size of the bar used in this detail, which led to high stress concentration in this area. Using No. 4 or No. 5 closed ties might help avoid this failure. Shape of the Push-Off Specimens. In future investiga- tions, it is recommended to use symmetric push-off speci- mens instead of the L-shaped specimen that was used in this research. However, due to the expected high load that is re- quired to break the symmetric specimen, half- or quarter- scale specimens should be used. Comparing the failure modes of Group 1 and Group 2 shows that the side external confinement of the specimen is very important to overcome the limited-width problem of the push-off specimens. All the specimens of Group 1 had slab failure, while almost all of the specimens of Group 2 had stud failure. Unfortunately, no mathematical models are available to quantify the amount of the side confinement needed to simulate a real bridge. Finite Element Investigation of the Push-Off Specimens The finite element method was used to investigate the be- havior of the push-off specimens. A commercial finite element package (Nastran) was used in the analysis. The push-off con- crete specimen and the grout filling the shear pocket were modeled using a eight-node cube element. Each node has three translational degrees of freedom (x, y, and z direction). The confining tube and the individual closed ties were modeled using the thin shell element. The circular cross section of the studs was replaced with a square cross section with equivalent area. The studs were modeled using the 20-node cube element. The following mechanical properties were assigned to the concrete mix of the specimen: compressive strength = 6.2 ksi (42.7 MPa), unit weight = 150 lb/ft3 (23.6 kN/m3), and Pois- son ratio = 0.15. The following mechanical properties were assigned to the grout mix: compressive strength = 9.6 ksi (66.2 MPa), unit weight = 145 lb/ft3 (22.8 kN/m3), and Poisson ratio = 0.15. 71 1-in. thick base plate Section B-B Plan View C-C 8" A A 10 " 10 " 1' -8 " 3'-11" B B C C 1'-4" A Section A-A 2' -0 " 2'-0" 9" HSS 15 x 9 x 5/16 1'-4"7 1/2" 9" 7 1/2" 1" 3" 1'-4 1/2" 7 1/2" 1'-3" 8 1/2" 1'-8" 3" 9" 3" 4 3/4" 3 1/4" 7 1/2" 1'-3" 8 1/2" A 7 1/ 2" 9" 7 1/ 2" 1' -8 " 2" 4" 4" 4" 4" 2" 1'-4 1/2" 2" 1'-0 1/2" 2" 5 1/ 2" 9" 5 1/ 2" 6 3/4" 3" 6 3/4" Figure 72. Concrete dimensions of P-8-ST-F/U.

CB A 1'-4" 2 ' - 0 " Section C-C C Section B-B B A C B 2 ' - 0 " A 7 1 / 2 " 9 " 7 1 / 2 " 7 1/2" 1'-3" 8 1/2" 1'-4" Section A-A 1/2 in. diameter, 28 in. long, threaded rod 1/2 in. thick plate with 4- 3/4" f holes 1'-7 1/2" 2 1/4" 4 1/2" 1'-7 1/2" 8 " 2 1/4" 2 1/4" 1 1/4" 2'-0" 1'-7 1/2" B A C 4" Section B-B Section C-C 6 1/2"11"6 1/2" 6 1 / 2 " 1 1 " 6 1 / 2 " Section A-A Figure 73. Typical reinforcement of the push-off specimens. Group 1 specimens on left, Group 2 specimens on right.

73 Specimen P-4-CT-U Specimen P-4-ST-U Specimen P-8-CT-U Specimen P-8-ST-U Figure 74. Fabrication of the push-off specimens.

The following mechanical properties were assigned to the stud: tensile strength = 64 ksi (441 MPa), yield strength = 54 ksi (372 MPa), unit weight = 490 lb/ft3 (76.9 kN/m3), and Poisson ratio = 0.30. Details of the finite element model are given in Appendix F. To check the validity of Equation 6.10.10.4.3-1 of the AASHTO LRFD specifications (7) for studs clustered in groups, each spec- imen was loaded with a horizontal load equal to the ultimate horizontal shear capacity determined by this equation. The load was surface loaded on a 10 × 10 in. (254 × 254 mm) area on the bearing block of the specimen to simulate the test setup. The re- sult of the surface load was at mid height of the 8 in. (203 mm) thick slab. Appendix F gives the results of the finite element 74 Figure 75. Welding of the 11⁄4-in. studs and the quality control tests. (continued text on page 86)

75 0 1 2 3 4 5 6 7 8 9 10 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Age (days) Co m p. S tre ng th , k si Concrete Mix SS Mortar Grout with pea gravel Test specimenL-beam Self equilibrium frame 4" Cyclic load jack Figure 76. Compressive strength versus age of the concrete mix and grout. Figure 77. Test setup. Ultimate Capacity, kip (kN) Source Four Studs Eight Studs Viest (46) 193.6 (861.1) 387.2 (1722.3) Ollgaard et al. (45) 477.2 (2122.6)* 954.4 (4245.2)* Oehlers and Johnson (48) 391.7 (1742.3)* 815.4 (3626.9)* LRFD Specifications (7), Equation 6.10.10.4.3-1 314.8 (1400.2)** 629.6 (2800.5)** LRFD Specifications (7), Equation 5.8.4.1-1 188.8 (839.8) 376.8 * Shear capacity is greater than the tensile ultimate capacity. ** Shear capacity is controlled by the tensile ultimate capacity. Table 8. Ultimate capacity of the stud cluster using various models.

Viest (46) Ollgaard et al. (45) Oehlers and Johnson (48) LRFD Specs. (7) Equation 6.10.10.4.3-1 LRFD Specs. (7) Equation 5.8.4.1-1 Test Failure Load, Ff kip Failure Mode f crQ F max f D F max f D F f nQ F f nV F See Figure Group 1: Ultimate test, 1.25-in. haunch, no external confinement on the specimen, four studs per specimen. 237 The load was applied by mistake at the interface level Slab failure: 1. Inclined crack on the side of the concrete specimens. 2. No grout crushing 3. All studs bent to about 5 degrees 122% 50% 61% 75% 126% 78-a P - 4 - S T - U 313 Slab failure: 1. Vertical crack on the side of the concrete specimens. 2. No grout crushing 3. All studs bent to about 10 degrees 162% 66% 80% 99% 166% 78-b Average for P-4-ST-U (four studs and steel tubes) 142% 58% 71% 87% 146% 241 Slab failure: 1. Horizontal crack on the side of the concrete specimen 2. The slab lifted away from the steel plate and the specimen could not take any more load 124% 51% 62% 77% 128% 78-c P - 4 - C T - U 259 Slab failure: 1. Horizontal crack on the side of the concrete specimen 2. The slab lifted up from the steel plate and the specimen could not take any more load 134% 54% 66% 82% 137% 78-d Average for P-4-CT-U (four studs and closed ties) 129% 53% 64% 80% 133% Average for P-4-ST-U and P-4-CT-U (all four-stud ultimate specimens) 136% 56% 68% 84% 140% Table 9. Test results of the push-off panel-to-steel specimens.

Viest (46) Ollgaard et al. (45) Oehlers and Johnson (48) LRFD Specs. (7) Equation 6.10.10.4.3-1 LRFD Specs. (7) Equation 5.8.4.1-1 Test Failure Load, Ff kip Failure Mode f crQ F max f D F max f D F f nQ F f nV F See Figure Group 1: Ultimate test, 1.25-in. haunch, no external confinement on the specimen, eight studs per specimen. 400 Slab failure: 1. Concrete bearing failure at the bearing block of the specimen 2. No grout crushing 3. Studs remained almost vertical 103% 42% 49% 64% 106% 78-e P - 8 - S T - U 346 Slab failure: 1. Concrete bearing failure at the bearing block of the specimen 2. No grout crushing 89% 36% 42% 55% 92% 78-f Average for P-8-ST-U (eight studs and steel tubes) 96% 39% 46% 60% 99% 376 Slab failure: 1. Horizontal crack on the side of the concrete specimen 97% 39% 46% 60% 100% 78-g P - 8 - C T - U 318 Slab failure: 1. Horizontal crack on the side of the concrete specimen 82% 33% 39% 51% 85% 78-h Average for P-8-CT-U (eight studs and closed ties) 90% 36% 43% 56% 92% Average P-8-ST-U and P-8-CT-U (all eight-stud ultimate specimens) 93% 38% 45% 58% 96% Average of all specimens in Group 1 (ultimate testing) 115% 47% 57% 71% 118% Table 9. (Continued).

Viest (46) Ollgaard et al. (45) Oehlers and Johnson (48) LRFD Specs. (7) Equation 6.10.10.4.3-1 LRFD Specs. (7) Equation 5.8.4.1-1 Test Failure Load, Ff kip Failure Mode f crQ F max f D F max f D F f nQ F f nV F See Figure Group 2: Fatigue/ultimate test, no haunch, with side external confinement, four studs per specimen. – Could not be tested because the specimen rotated in the horizontal plan due to improper setup Test failed. – – – – – 81-a P - 4 - S T - F / U 231 Stud failure: 1. All studs failed at the welding area. 2. The grout around the studs did not crush. 3. Some of the concrete outside the confinement tube failed. 4. At failure load, some cracks around the confinement tube were observed on top of the specimen. 119% 48% 59% 73% 122% 81-b Average for P-4-ST-F/U (four studs and steel tubes) 119% 48% 59% 73% 122% 308 Stud/grout failure: 1. Two of the studs failed at the welding area. The other two bent about 30 degrees. 2. No cracks were observed on top of the specimen. 3. Grout inside the confinement area was crushed. 4. Concrete outside the confined grout area was crushed. 5. Bond failure between the bottom tie and the surrounding concrete. 159% 65% 79% 98% 163% 81-c P - 4 - C T - F / U 220 Stud/concrete failure: 1. One stud failed at the welding area. The remaining studs bent about 25 degrees. 2. A cone-shape failure was observed in the grout around the studs. 3. Bond failure between the bottom tie and the surrounding concrete 4. A pronounced crack was observed on top of the specimen. 114% 46% 56% 70% 117% 81-d Average for P-4-CT-F/U (four studs and closed ties) 137% 56% 68% 84% 140% Average for P-4-ST-F/U and P-4-CT-F/U (all four-stud fatigue/ultimate specimens) 128% 52% 64% 79% 131% Table 9. (Continued).

Viest (46) Ollgaard et al. (45) Oehlers and Johnson (48) LRFD Specs. (7) Equation 6.10.10.4.3-1 LRFD Specs. (7) Equation 5.8.4.1-1 Test Failure Load, Ff kip Failure Mode f crQ F max f D F max f D F f nQ F f nV F See Figure Group 2: Fatigue/ultimate test, no haunch, with side external confinement, eight studs per specimen. 379 The slab lifted off from the steel plate at the far edge and the specimen could not take any more load. 1. Plate did not come off the specimen. 2. No grout failure was detected 3. No cracks were observed on top or around the specimen. 98% 40% 46% 60% 101% 81-e P - 8 - S T - F / U 300 Stud failure: 1. All studs failed. Two studs failed at the base material, four studs failed at the weld location, and the remaining two sheared off. 2. Grout crashed around the studs. 3. Concrete outside the steel tube confinement did not crack. 4. Slippage occurred between the steel tube and the grout inside 5. At failure load, there was a two-crack V shape at the side of the specimen. 6. At failure load, there was a very fine crack around the steel tube on top of the specimen. 77% 31% 37% 48% 80% 81-f Average for P-8-ST-F/U (eight studs and steel tubes) 88% 36% 42% 54% 91% 245 Stud failure: 1. Two of the studs sheared off, the following two failed at the base material, and the remaining four bent about 20 degrees. 2. A cone-shaped failure was observed in the grout around the studs. 3. Bond failure between the bottom tie and the surrounding concrete 63% 26% 30% 39% 65% 81-g P - 8 - C T - F / U 245 Bond failure of the lower closed tie: 1. The first four studs were bent about 15 degrees. The remaining four studs were slightly bent. 2. The failure was cone-shaped and formed around the group of studs. 3. Bond failure between the bottom tie and the surrounding concrete. 63% 26% 30% 39% 65% 81-h Average for P-8-CT-F/U (eight studs and closed ties) 63% 26% 30% 39% 65% Average for P-4-ST-F/U and P-4-CT-F/U (all eight-stud fatigue/ultimate specimens) 76% 31% 36% 47% 78% Average of all specimens in Group 2 (fatigue/ultimate testing) 102% 42% 50% 63% 105% Table 9. (Continued).

80 (a) P-4-ST-U, Sp. A (b) P-4-ST-U, Sp. B (c) P-4-CT-U, Sp. A (d) P-4-CT-U, Sp. B (e) P-8-ST-U, Sp. A (f) P-8-ST-U, Sp. B (g) P-8-CT-U, Sp. A (h) P-8-CT-U, Sp. B Figure 78. Failure modes of Group 1 push-off specimens.

350,000 400,000 Sp. B -0.800 -0.700 -0.600 -0.500 -0.400 -0.300 -0.200 -0.100 0.000 0 50,000 100,000 150,000 200,000 250,000 300,000 Load (lb) D i s p l a c e m e n t ( i n . ) Sp. A P-4-ST-U Load (lb) Sp. A Sp. B 350,000 400,0000 50,000 100,000 150,000 200,000 250,000 300,000 -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 D i s p l a c e m e n t ( i n . ) P-4-CT-U Load (lb) 350,000 400,0000 50,000 100,000 150,000 200,000 250,000 300,000 -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 D i s p l a c e m e n t ( i n . ) Sp. A Sp. B P-8-CT-U 350,000 400,0000 50,000 100,000 150,000 200,000 250,000 300,000 P-8-ST-U -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 Load (lb) D i s p l a c e m e n t ( i n . ) Sp. A Sp. B Figure 79. Load-displacement relationship of Group 1 push-off specimens.

No results are available. The specimen could not be tested because the steelplate rotates in its horizontal plan. Load (lbs) After fatigue load P-4-ST-F/U Sp. A P-4-ST-F/U Sp. B Before fatigue load After fatigue load Before fatigue load -0.1 -0.08 -0.06 -0.04 -0.02 0 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 Load (lbs) P-4-CT-F/U Sp. A Load (lbs) D i s p l a c e m e n t ( i n . ) -0.1 -0.08 -0.06 -0.04 -0.02 0 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 D i s p l a c e m e n t ( i n . ) -0.10 -0.08 -0.06 -0.04 -0.02 0.00 D i s p l a c e m e n t ( i n . ) 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 P-4-CT-F/U Sp. B Load (lbs) -0.10 -0.08 -0.06 -0.04 -0.02 0.00 D i s p l a c e m e n t ( i n . ) Before fatigue ld. After fatigue ld. The LVDT is dead Figure 80. Load-displacement relationship of Group 2 push-off specimens due to fatigue load before and after the 2E+6 cycles.

-0.10 -0.08 -0.06 -0.04 -0.02 0.00 D i s p l a c e m e n t ( i n . ) Load (lbs) No data was recorded for this specimen P-8-ST-F/U Sp. B 0 20,000 40,000 60,000 80,000 100,000 120,000 Before fatigue load After fatigue load -0.10 -0.08 -0.06 -0.04 -0.02 0.00 D i s p l a c e m e n t ( i n . ) Load (lbs) P-8-CT-F/U Sp. B 0 20,000 40,000 60,000 80,000 100,000 120,000 P-8-CT-F/U Sp. A -0.10 -0.08 -0.06 -0.04 -0.02 0.00 0 20,000 40,000 60,000 80,000 100,000 120,000 Load (lbs) D i s p l a c e m e n t ( i n . ) Before fatigue ld. After fatigue ld. Before fatigue load After fatigue load P-8-ST-F/U Sp. A -0.12 -0.10 -0.08 -0.06 -0.04 -0.02 0.00 0 20,000 40,000 60,000 80,000 100,000 120,000 Load (lbs) D i s p l a c e m e n t ( i n . ) The LVDT was dead Figure 80. (Continued).

84 (a) P-4-ST-F/U Sp. A (b) P-4-ST-F/U Sp. B (c) P-4-CT-F/U Sp. A (d) P-4-CT-F/U Sp. B Figure 81. Failure modes of Group 2 push-off specimens.

85 (e) P-8-ST-F/U Sp. A (f) P-8-ST-F/U Sp. B (g) P-8-CT-F/U Sp. A (h) P-8-CT-F/U Sp. B Figure 81. (Continued).

analysis of various specimens, and Table 10 gives a summary of the maximum stresses in the stud, grout, and confinement tool. The four- and eight-stud specimens are not able to deliver the horizontal ultimate shear capacity as given by Equation 6.10.10.4.3-1 of the AASHTO LRFD specifications (7). This can be seen from the average axial tensile stress at the stud base, which is higher than the ultimate tensile strength of the SAE 1018 stud material (64 ksi, or 441.3 MPa). The upper limit of Equation 6.10.10.4.3-1 of the AASHTO LRFD specifications (7), Asc Fu, does not recognize the fact that the stud close to failure is subjected to a combination of axial tensile and normal flexural stresses. This can be seen by check- ing the average principal tensile stress at the stud base, which is about 155% of the axial tensile stress, as shown in Table 10. This means that the upper limit of Equation 6.10.10.4.3-1 overestimates the stud’s shear capacity. This finding was con- firmed by the push-off test, where almost none of the speci- mens were able to reach the capacity determined by Equation 6.10.10.4.3-1 of the AASHTO LRFD specifications (7). Maximum bearing stress in the grout is located in front of each stud close to the stud base. It extends vertically for a dis- tance approximately equal to the stud diameter. The maxi- mum bearing stress is about 30 ksi (206.9 MPa), which is about 310% of the compressive strength of unconfined grout mix (9.6 ksi, or 66.2 MPa). However, if confinement is pro- vided around the shear pocket, the compressive strength of the grout can be significantly increased, as follows: Effective lateral confining pressure, fl, (for steel tube confinement) (for closed ties confinement) = × ×( )( ) ( 2 0 44 3 60 1 75 legs in per leg bars ksi in. 2. . )( ) =15 6 034in. . = × × ⎛⎝⎜ ⎞⎠⎟ ( ) ( ) 2 1 5 16 36 1 12 sides in. in. ksi in. in.( ) = 1 875. ksi = ∑ A f sb s yh c Confined grout strength, fc0 = f0 + 4.1kfl = 9.6 + 4.1 × 1 × 1.875 = 17.3 ksi (119.3 MPa) (for steel tube confinement) = 9.6 + 4.1 × 1 × 6.034 = 34.3 ksi (236.8 MPa) (for closed ties confinement) The confinement around the stud group helps to distrib- ute the bearing stresses of the grout volume on the concrete slab in front of the grout volume. The highest bearing stress is about 2.30 ksi (15.9 MPa), and the average bearing stress over the slab height is about 2.0 ksi (13.8 MPa). The confinement provided by the steel tube helps to dis- tribute the bearing stresses on a wider part of the slab, result- ing in a reduction in the compression in the slab compared with when closed ties are used. The truncated shape of the shear pocket and grout volume helps in distributing the bearing stresses more uniformly across the slab height. Full-Scale Beam Test The objective of the full-scale beam testing was to investi- gate the feasibility of extending the AASHTO maximum stud spacing from 2 to 4 ft (610 to 1220 mm) by checking differ- ences in structural performance of two composite beams due to fatigue and ultimate loads. Two full-scale composite beams, each 32 ft (9.75 m) long, were fabricated. The beams were identical except that the spacing between the stud clusters was 2 ft (610 mm) for the first beam and 4 ft (1220 mm) for the second beam. Each composite beam was made of an 8 in. (203 mm) thick precast slab supported by a W18 × 119 steel beam. The slab and the steel beam were made composite using sixty-four 11⁄4 in. (31.8 mm) studs over the full-span length. The studs on Beam 1 were clustered in 16 groups spaced at 24 in. (610 mm), with four studs per group. The 24 in. (610 mm) spac- ing is the current limit according to the AASHTO LRFD specifications (7). 86 Four-Stud Specimens Eight-Stud Specimens P-4-ST-U (Steel Tube) P-4-CT-U (Closed Ties) Average P-8-ST-U (Steel Tube) P-8-CT-U (Closed Ties) Average Applied horizontal load (kip)* 314.8 629.6 Maximum axial tensile stress at base of the stud (ksi) 58.4 99.1 78.8 99.9 74.9 87.4 Maximum tensile principal stress at base of the stud (ksi) 92.5 157.0 124.8 162.0 117.0 139.5 Maximum longitudinal movement of the stud head (in.) 0.0075 0.0103 0.0089 0.0109 0.00954 0.01022 Maximum axial tensile stress in confinement material in the transverse direction of the specimen (ksi) 21.0 3.7 NA 30.7 5.3 NA Maximum bearing stress in grout in front of the stud (ksi) 29.1 31.8 30.5 27.1 31.6 29.4 Maximum bearing stress in the concrete in front of the grout volume (ksi) 2.31 2.31 2.31 2.30 2.30 2.30 * Determined using Equation 6.10.10.4.3-1 of the AASHTO LRFD Specifications (7 ). Table 10. Summary of the finite element analysis results for the push-off specimens.

The studs on Beam 2 were clustered in eight groups, spaced at 48 in. (1220 mm), with eight studs per group, as shown in Figures 82 to 84. The spacing between the studs in each group was 3 in. (76 mm) in the longitudinal direction. Two studs per row spaced at 5 in. (127 mm) in the transverse direction were used. In each beam, the stud clusters on the south half of the beam were confined with HSS 9 × 7 × 0.188 in. (229 × 178 × 5 mm) and 13 × 9 × 5⁄16 in. (330 × 229 × 8 mm) tubes, and the stud clusters on the north half were con- fined with individual No. 4 (13) and No. 6 (19) closed ties for the 2 ft (610 mm) and 4 ft (1220 mm) clusters, respectively. The concrete slab of each beam was made of one precast panel, which was reinforced with two welded wire reinforce- ment (WWR) meshes. The top mesh was made of 6 × 6 – D10 × D10 (152 × 152 – MD65 × MD65), and the bottom mesh was made of 6 × 6 – D14 × D14 (152 × 152 – MD90 × MD90). This amount of reinforcement was provided in accordance with the minimum reinforcement requirements of the empirical design method given in Article 9.7.2 of the AASHTO LRFD specifications (7). Figures 82 to 84 show the details of the full-scale beams. Wood forming and a normal weight, 7 in. (178 mm) slump, 6 ksi (41.4 MPa) concrete mix were used in making the panels, as shown in Figure 85. The panels were moist cured for 7 days and then stored in the laboratory until they were in- stalled on the steel beams. No shrinkage cracks were observed on the panels. On both beams, 11⁄4 in. (31.8 mm) headless studs with heavy-duty nuts and washers were used. Due to the lack of a high-voltage source at the testing facility, the studs were manually welded, as shown in Figure 86. The welding quality was checked by visual inspection, bending the stud to 45 degrees, and using the hydraulic push-off device discussed earlier in this chapter. Foam rods were glued at the edges of the top flange of the steel beam to build a dam for the 1 in. (25 mm) thick concrete haunch. Each panel was installed by car- rying it with a spreader beam that supported the panel at 7 points spaced uniformly at 4 ft (1220 mm) and located in the mid distance between adjacent shear pockets. Finally, the shear pockets were filled with SS Mortar mix with no pea gravel, as shown in Figure 86. The grouting of each pocket continued until grout came out from the venting ports. Grouting of Beam 1 went smoothly, with complete filling of the haunch and no recorded problems. For Beam 2, however, after the grout was cured and the foam rods were removed, it was noticed that about a 2 ft (610 mm) long distance between two shear pockets was not completely filled with grout. This 87 Beam #1 B1 B1 Beam #2 B2 B2 S Top layer of reinforcement (6"x6"-W10xW10, Length = 31'-8", Width = 3'-8") "X" Location of a group of studs, "*" Location of a lifting point Bottom layer of reinforcement (6"x6"-W14xW14, Length = 31'-8", Width = 3'-8") 31'-0" 6" 7 spacings @4 ft 2 ft 31'-0" 6" HSS 13x9x5/16 Tube ConfinmentNo. 6 Closed Ties Confinment * ** ** ** * ****** 15 spacings @2 ft = 30 ft 1'-0"1'-0" A2 A2 A1 A1 4'- 0" 3'- 8" 6" 6" 2 ft 4'- 0" 3'- 8" N Figure 82. Arrangement of stud clusters of Beam 1 and Beam 2.

4'-0" Section B1-B1 11 1/2" 5 " 8 " 1 " 1 ' - 7 " 5" 11 1/4" 3 " 1/2" 4 1/2" 1" 3" W18x119 2" 5" 2" 9" 2 " 1 " 2 " 3 " 2 " 7 " HSS 9x7x0.188 9" 3 / 4 " 3/4" HSS 9x7x0.188 1/2" 4 1/2" 1" 3" Section A1-A1 #4 closed ties 1 ' - 1 3 / 4 " 1'-1 3/4" 11 1/2" 4'-0" 5 " 8 " 1 " 1 ' - 7 " 5" 9" 11 1/4" 3 " W18x119 2" 5" 2" 9" 11 1/2" 2 " 1 " 2 " 3 " 2 " 7 " 9 1 / 2 " 3-#4 closed ties with 1 in. clear spacing Figure 83. Sections A1-A1 and B1-B1.

1/2" 4 1/2" 1" 3" Section A2-A2 Section B2-B2 1 ' - 3 1 / 2 " HSS 13x9x5/16 4'-0" 5 " 8 " 1 " 1 ' - 7 " 5" 9" 11 1/4" 3 " 1/2" 4 1/2" 1" 3" W18x119 2" 5" 2" 9" 2 " 3 " 3 " 3 " 2 " 1 ' - 1 " 2 " 1 " HSS 13x9x5/16 W18x119 3-#6 closed ties with 1 in. clear spacing #6 closed ties 11 1/2" 4'-0" 5 " 8 " 1 " 1 ' - 7 " 5" 9" 11 1/4" 3 " 2" 5" 2" 9" 11 1/2" 2 " 3 " 3 " 3 " 2 " 1 ' - 1 " 1 ' - 9 3 / 4 " 1'-1 5/8" 2 " 1 " Figure 84. Sections A2-A2 and B2-B2.

90 0 1 2 3 4 5 6 7 8 9 0 42 6 8 10 12 14 16 18 20 22 24 26 28 Age (days) St re ng th (k si) SS Mortar Grout without pea gravel Panel Concrete Mix Figure 85. Forming and casting of the precast concrete panels.

91 Figure 86. Building the composite beams. was due to the excessive time that elapsed between when the grout was mixed with water and when this area of the beam was grouted. This area was batched by injecting grout directly at the haunch level. Fatigue Testing of the Beam Each beam was loaded with one concentrated load at midspan and subjected to 2,000,000 cycles of fatigue load through a hydraulic actuator. The load setup put all of the shear pockets under the same amount of horizontal load, as shown in Figures 87 and 88. The upper and lower limits of the fatigue load were determined in three steps. Step 1: Calculate Stud Fatigue Resistance, Zr. Using Equations 8 (Equation 6.10.10.2-1 (7)) and 9 (Equation 6.10.10.2-2 (7)) gives Zr = 7.53 × 1.252 = 11.77 kip/stud Step 2: Calculate Vertical Shear Force, Vf. (15) (Equation 6.10.10.1.2-2 (7)) where Vf = vertical shear range due to fatigue load, I = moment of inertia of the composite section, Q = first moment of the area above the interface about the neutral axis, and Vsr = shear flow range due to fatigue load at the interface, which can be determined by: (16) (Equation 6.10.10.1.2-1 (7)) p = spacing of the studs = 24 in. or 48 in., n = number of studs in a cross section = 4 studs for p = 24 in., or 8 studs for p = 48 in. p nZ V r sr ≤ V V Q I sr f =

I and Q were calculated as follows (see Figure 68): Effective slab width (LRFD specifications, Article 4.6.2.6.1): Beff = least of (17) = 48 in. of the slab = 0.189 × 48 = 9.07 in. of the haunch = 0.189 × 10 = 1.89 in. Depth of the NA = 9 07 8 4 1 89 1 8 0 5 35 1 8 1 0 0 5 . . . . . . ×( )( )+ ×( ) +( )+ ( ) + + ×( ) = ×( )+ ×( )+ ( ) = 19 955 66 9 07 8 1 89 1 35 1 109 5 . . . . . 5 8 75= . .in n Beff' n Beff' Modular ratio, n E E c s ' ., . . = = × =33 000 0 150 8 2 51 5 , , . 490 29 000 0 189= B ft ts = = + = × + × 4 48 12 0 5 12 8 0 5 11 in. girder flange. . . . 3 102 1 4 0 25 32 8 96 = = × = = ⎧ ⎨⎪ ⎩⎪ ⎫ ⎬in. span in.ft ft ⎪ ⎭⎪ 92 P fatigue upper limit = 88 kips + 44 kips - 44 kips 15.5 ft 15.5 ft 682 ft-kip Bending moment Shear force Beff. = 4'-0" 8" 19 " W18x119 11 1/4" 1. 06 " n Beff. slab = 9.07" N.A.1" n Beff. haunch = 1.89" 8. 72 " 19 .2 8" 10" 341 ft-kip 341 ft-kip Figure 87. Elastic properties of the composite section. Figure 88. Fatigue test setup.

Q = (9.07 × 8)(8.72 – 4) + (1.89 × 1)(8.72 – 8 – 0.5) = 342.9 in3 Substituting Equation 16 in Equation 15 yields P (concentrated load at midspan) = 2 × 43.2 = 86.4 kip To maintain stability of the test setup, a minimum load of 1.6 kip was provided as the lower limit of the fatigue load. There- fore, the upper limit of the fatigue load = 86.4 + 1.6 = 88 kip. Step 3: Check Stresses at Midspan to Make Sure They Are Within the Elastic Range of Material 0.4 × 8.2 = 3.28 ksi) (fy = 50 ksi) Strain gauges and vertical displacement measuring devices were installed at the quarter-point and midspan sections of fbottom surface of steelbeam = ×( )( )682 12 19 28 7 . ,551 20 90= + <. ksi ftopsurface of steelbeam = ×( ) + −( )682 12 8 1 8 72 7 . ,551 0 30= + . ksi fbottomsurface of slab = − ×( ) −( ) 0 189 682 12 8 72 8 . . 7 551 0 15 , .= − ksi ftopsurface of slab = − ×( )( ) 0 189 682 12 8 72 7 551 . . , = − < =1 79 0 4. ( . 'ksi fc V nZ I pQ f r = = × × × = 4 11 77 7 551 24 342 9 43 2 . , . . kip I = × + ×( ) −( )⎡⎣⎢ ⎤ ⎦⎥ + × 9 07 8 12 9 07 8 8 72 4 1 89 1 3 2 3 . . . . 12 1 89 1 8 72 8 0 5 2 190 35 1 2 + ×( ) − −( )⎡⎣⎢ ⎤ ⎦⎥ + +( . . . , . ) + + × −( )⎡⎣ ⎤⎦ =8 1 0 5 19 8 72 7 5512. . , in4 each beam. The upper limit of the fatigue load was applied as a static load, and the measurements were collected (pre- fatigue records); the beam was then exposed to 2,000,000 cy- cles of the fatigue load at 2,000,000 cycles/sec. Finally, the beam was loaded statically with the upper limit of the fatigue load, and the strain and displacement measurements were collected (postfatigue records). The testing scenario worked well with Beam 1. However, for Beam 2, the hydraulic system of the actuator needed to be re- paired when the beam was exposed to about 1,000,000 cycles. The fatigue test was therefore stopped, the static load was applied, and the measurements were collected at 1,000,000 cycles. After the hydraulic system was repaired, the fatigue test was resumed. The measurements that were planned to be taken at 2,000,000 cycles were not, however, collected because the steel beam fractured as a result of fatigue load close to the midspan section at about 1,950,000 cycles, as shown in Figure 89. The fatigue fracture started at the bottom flange and prop- agated through the web, where it stopped close to the web/top flange junction. As a result of this unexpected failure, a 1⁄2 in. (13 mm) separation between the haunch and the steel beam occurred over a distance of about 2 ft (610 mm) around the failure location. The beam was thoroughly inspected, and no other cracks or signs of distress were detected. It was believed that the fatigue fracture failure occurred because the steel beam was previously subjected to 4,000,000 cycles of fatigue load when it was used for the full-scale bridge test. Also, welding and removing of the 21⁄2 in. (64 mm) diam- eter tubes that were needed in the full-scale bridge test resulted in residual stresses in that flange. Figure 90 shows the pre- and post-fatigue stress distribu- tion at the quarter-point and midspan sections of the beams and compares them with the stresses calculated by the elastic stress theory assuming full-composite action. Figure 91 shows the deflection of the beams at the quarter-point and 93 Figure 89. Fatigue fracture of the steel beam and separation between the haunch and the steel beam.

94 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 -2.000 0.000 2.000 4.000 6.000 8.000 10.000 12.000 Stress (ksi) -2.000 0.000 2.000 4.000 6.000 8.000 10.000 12.000 Stress (ksi) -2.000 0.000 2.000 4.000 6.000 8.000 10.000 12.000 Stress (ksi) -2.000 0.000 2.000 4.000 6.000 8.000 10.000 12.000 Stress (ksi) D ist an ce fr om to p fib er (i n.) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 D ist an ce fr om to p fib er (i n.) North side (Ties) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 D ist an ce fr om to p fib er (i n.) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 -5.000 0.000 5.000 15.00010.000 25.00020.000 Stress (ksi) -5.000 0.000 5.000 15.00010.000 25.00020.000 Stress (ksi) D ist an ce fr om to p fib er (i n.) Theoretical Pre- fatigue Post- fatigue Theoretical Pre- fatigue Post- fatigue Theoretical Pre- fatigue Post- fatigue Midspan Section 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 D ist an ce fr om to p fib er (i n.) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 D ist an ce fr om to p fib er (i n.) South Side (Tube) 2-ft cluster spacing 4-ft cluster spacing Theoretical Pre- fatigue Post- fatigue Theoretical Pre- fatigue Post- fatigue Theoretical Pre- fatigue Post- fatigue Figure 90. Composite section stresses (theoretical and pre- and post-fatigue).

95 0.296 0.431 0.296 0.261 0.450 0.268 0.285 0.437 0.293 0.000 0.100 0.200 0.300 0.400 0.500 0.600 North side (ties) Midspan South side (tube) D ef le ct io n (in .) Theoretical Post- Fatigue Pre- Fatigue Theoretical Post- Fatigue Pre- Fatigue Theoretical Post- Fatigue Pre- Fatigue 2-ft cluster spacing 0.296 0.431 0.2960.306 0.470 0.3010.301 0.480 0.303 0.000 0.100 0.200 0.300 0.400 0.500 0.600 North side (ties) Midspan South side (tube) D ef le ct io n (in .) Theoretical Post- Fatigue Pre- Fatigue Theoretical Post- Fatigue Pre- Fatigue Theoretical Post- Fatigue Pre- Fatigue 4-ft cluster spacing Figure 91. Composite section deflection (theoretical and pre- and post-fatigue).

96 midspan sections. The quarter-point and midspan sections were chosen because they were at the mid distance between two adjacent shear pockets, where it was highly expected that partial composite action would occur. An examination of these figures found the following: • Regardless of the stud cluster spacing and the type of con- finement: – At the quarter-point and midspan sections, pre- and post-fatigue stresses showed almost a linear distribution. – At the midspan section, pre- and post-fatigue stresses at extreme compression fiber of the composite beams were about 20% higher than the stresses calculated by the elastic stress theory. However, the opposite trend occurred at the quarter-point sections. – At the quarter-point and midspan sections, pre- and post-fatigue stresses at extreme tension fiber of the composite beams were about 20% less than the stresses calculated by the elastic stress theory. – At the quarter-point and midspan sections, pre- and post-fatigue stresses at extreme tension and compres- sion fibers of the composite beams were within the elas- tic range of the material. – At the quarter-point and midspan sections, post-fatigue stresses and deflection showed almost no increase when compared with the prefatigue stresses. On the contrary, at some locations the postfatigue stresses were smaller than the prefatigue stresses. – The pre- and post-fatigue deflection levels were in close agreement with deflection calculated using the elastic stress theory. • Regardless of the type of confinement provided around the stud clusters, the stress distribution and deflection meas- urements of Beam 1 and Beam 2 were almost identical. • Regardless of the stud cluster spacing, the stress distri- bution and deflection measurements of the north side of the beam, where ties were used, were almost identical to the stress distribution and deflection measurements of the south side of the same beam, where tubes were used. • Visual inspection of the composite beams before, during, and after applying the 2,000,000 cycles of fatigue load found – no cracks on the top surface of the concrete slab; – no separation between the concrete haunch and the steel beams, except the separation that occurred in Beam 2 around the location of the fatigue fracture failure; – no cracks or signs of distress in the haunch; and – no residual deflection at midspan after removing the load. Based on these observations, the following conclusions were drawn: • It is safe to use Equation 6.10.10.2-1 of the LRFD specifica- tions (7) to determine the fatigue capacity of studs grouped in clusters and spaced as far as 48 in. (1220 mm) apart. • Full composite action between precast concrete panels and steel beams can be maintained up to 48 in. (1220 mm) of spacing between clusters of studs. • The two types of proposed confinement—closed ties and tubes—provide similar behavior due to fatigue load. Ultimate Testing of the Full-Scale Beams To individually investigate the ultimate capacity of the stud clusters for various types of confinement, the 32 ft (9.75 m) span of each beam was divided into two equal spans, as shown in Figures 92 and 93. For Beam 1, an intermediate support was installed exactly at midspan. Then each beam half was tested as a 15.5 ft (4.72 m) simply supported beam under one concentrated load close to the midspan point of that span. This was done by removing the external support of the other half of the span, as shown in Fig- ure 92. The applied concentrated load would thus provide hor- izontal shear forces at the interface only on the stud clusters that existed on the simply supported span. Although the weight of the cantilevered span would provide additional stresses on the studs of the simply supported span, careful checking of these stresses revealed that it would be about 2% of the stresses pro- vided by the concentrated load, which could be ignored. For Beam 2, two intermediate supports were added be- cause the steel beam was fractured close to the midspan point. This arrangement resulted in two simply supported spans that each measured 11 ft (3.35 m) long. To make sure that no continuity existed between the two simply supported spans, the concrete slab was jackhammered at the same loca- tion where the steel beam was fractured to create a real hinge, as shown in Figure 92. Each test setup used two hydraulic jacks, 300 kip (1334 kN) each, supported on a short spreader beam to apply the load as one concentrated load, as shown in Figures 92 and 93. Two modes of failure were checked to determine the possible mode of failure of each simply supported beam. Flexural Capacity of the Composite Beam. The flexural capacity of the composite beam was determined in accor- dance with Article 6.10.7 of the AASHTO LRFD specifica- tions (see Figure 94). Assume that the neutral axis depth is less than the thick- ness of the slab, Dp < 8 in. Ignoring reinforcement in the lon- gitudinal direction of the slab, equilibrium of forces at the plastic stage yields (0.85 × 8.2 × 48)(Dp) = (50)(35.1) Dp = 5.25 in. < 8 in. (inside the concrete slab)

Classification of the composite section (compact versus noncompact): (18) (Equation 6.10.6.2.2-1 (7)) where Dcp = depth of the web in compression at the plastic moment. Therefore, the composite section was compact. Since Dp > [0.1Dt = 0.1(8 + 1 + 19) = 2.8 in.] (Equation 6.10.7.1.2-1 (7)), where Dt = total depth of the composite sec- tion, therefore: (19) (Equation. 6.10.7.1.2-2 (7)) Mp = (0.85 × 8.2 × 48 × 5.25)(8 + 1 + 0.5 × 19 − 0.5 × 5.25) = 27,883 kip-in. = 2,324 kip-ft Ductility can be checked as outlined in the LRFD specifi- cations, Article 6.10.73: Mn = − + + ⎛⎝⎜ ⎞⎠⎟ =( , ) . . . ,2 324 1 07 0 7 5 25 8 1 19 2 182 kip-ft M M D D n p p t = − ⎛ ⎝⎜ ⎞ ⎠⎟1 07 0 7. . 2 2 3 67 3 67 29 000 50 D t t E F cp w w yc = ×⎛ ⎝⎜ ⎞ ⎠⎟ ≤ = zero . . , = ⎛ ⎝⎜ ⎞ ⎠⎟88 4. [Dp = 5.25 in.] < [0.42Dt = 0.42(8 + 1 + 19) = 11.76 in.] OK Therefore, if the 15.5 ft (4.72 m) simply supported spans of Beam 1 were to fail in flexure, it would require a concentrated load P = 2,182/3.799 = 574 kip, and, if the 11.0 ft (3.35 m) simply supported spans of Beam 2 were to fail in flexure, it would require a concentrated load P = 2,182/2.588 = 843 kip, which was beyond the capacity of the hydraulic jacks. Nominal Shear Resistance, Qn. Using Equation 14 (Equation 6.10.10.4.3-1 (7)) where Asc = 1.23 in2, Fu = 64 ksi (1.23 × 64 = 78.72 kip/stud); therefore Qn = 78.72 kip/stud = 78.72 × 4 = 314.9 kip/four-stud cluster = 78.72 × 8 = 629.8 kip/eight-stud cluster Therefore, if Beam 1 were to fail in horizontal shear at the interface, this would require failure of the four shear pockets between the concentrated load and the exterior support, as Qn = × × =0 5 1 23 8 2 5 490 130 5. . . , . kip/stud E w fc c c= = ( ) =33 000 33 000 0 150 8 2 5 4901 5, , . . ,' . ksi, and 97 Figure 92. Ultimate test arrangement of Beam 1.

shown in Figure 92, and a horizontal shear force at the inter- face = 314.9 × 4 = 1,259.6 kip over this distance. The height of the plastic neutral axis that is equivalent to this force The corresponding plastic moment = (1,259.6)(8 + 1 + 0.5 × 19 − 0.5 × 3.77) = 20,928 kip-in. = 1,744 kip-ft. The corresponding concentrated load = 1,744/3.799 = 459 kip = × × = 1 259 6 0 85 8 2 48 3 77 , . . . . in. Applying the same procedure for Beam 2, the concentrated load that would be required to cause horizontal shear failure at the stud cluster between the concentrated load and the in- terior support = 356.1 kip. According to this analysis, it was expected that the four simply supported beams would fail in horizontal shear. Each simply supported beam was provided with one set of strain gauges and a deflection measurement device at the location of the applied concentrated load. The relative horizontal dis- placement between the slab and the steel beam was also recorded at the free end of each beam. The load was applied at 10 kip (44.5 kN) per second until failure occurred or until the hydraulic jacks’ capacity of 600 kip (2669 kN) was reached, whichever came first. Test Results Beam 1-North and Beam 1-South failed in flexure where the top fiber of the concrete slab was crushed in compression, as shown in Figure 95. The applied load at failure was about 600 kip (2669 kN), which was equal to the maximum capac- ity of the hydraulic jacks combined. The flexural failure on Beam 1-North was accompanied with web buckling failure of the steel beam. 98 Figure 93. Ultimate test arrangement of Beam 2. 4'-0" 8" 19 " W18x119 11 1/4" 1. 06 " Plastic NA 1" 10" 5 1/ 4" 0.85x8.2 ksi 50 ksi P co nc re te P st ee l Figure 94. Stress distribution at plastic stage.

Beam 2-North and Beam 2-South did not show any signs of failure in horizontal shear or flexure. Each beam was loaded up to the maximum combined capacity of the hy- draulic jacks—namely, 600 kip (2669 kN). Inspection of the top surface of the concrete slab showed that a longitudinal bursting hair crack was formed exactly over the web location of the steel beam, as shown in Figure 96. The crack covered almost the full length of Beam 1, which was made with 24 in. (610 mm) cluster spacing, while in Beam 2, which was made with 48 in. (1220 mm) cluster spac- ing, the crack covered only the midspan area of the north and south beams. The width of the crack of Beam 1 and Beam 2 was about 0.04 in. (1 mm) and 0.03 in. (3⁄4 mm), respectively. Due to the small width of these cracks, they were not detected until the beams were removed from the supports and set on the ground. However, it is believed that these cracks started to form when the applied moment was about 70% of the plastic flexural capacity of the composite section, when a loud explosion was heard during testing of Beam 1-North and Beam 1-South. Table 11 summarizes the failure mode and the maximum applied load. Figures 97 and 98 show the load-deflection and load- horizontal slip relationships of the full-scale beams. To help in studying the structural behavior of the four beams, the load was replaced by the corresponding applied moment as a percentage of the plastic moment capacity of the com- posite section. 99 (a) Beam #1-North (b) Beam #1-South (c) Beam #2-North (d) Beam #2-South Figure 95. Failure modes of the full-scale beams.

Figures 99 and 100 show the strain distribution of the full- scale Beam 1 and Beam 2, respectively, at various levels of applied load. An analysis of these figures allowed the following conclu- sions to be drawn. Appropriateness of Using the LRFD Specifications for Estimating the Horizontal Shear Capacity of Stud Clusters. Regardless of the stud cluster spacing and the type of confine- ment, all beams were able to develop the stud ultimate capacity given by Equation 6.10.10.4.3-1 of the LRFD specifications (7). The procedure in Article 6.10.7 of the AASHTO LRFD specifications gives a fair estimate of the ultimate flexural ca- pacity of composite sections. Deflection and Horizontal Slip (Figures 97 and 98). The slope of the load-deflection relationship, which is a measure of the composite beam stiffness, is almost the same for the four beams. This means that extending the stud cluster spac- ing to 48 in. (1220 mm) does not reduce the composite beam stiffness. The beams where the stud clusters were confined with HSS tubes showed smaller deflection than the beams where the stud clusters were confined with closed ties. The difference is about 10%. The beams made with 48 in. (1220 mm) cluster spacing showed about a 25% increase in deflection and horizontal slip compared with the beams made with 24 in. (610 mm) cluster spacing. The research team believes that this increase was due to the flexural fatigue failure at midspan that oc- curred during the fatigue test. This can be confirmed from Figure 97, where Beams 2-North and 2-South showed about 0.1 in. (2.54 mm) of deflection once a small amount of load was applied. Also, Figure 98 shows that these beams did not show any horizontal slip for the first period of applying the load (from zero to about 10%). Stress Distribution (Figures 99 and 100). At the same ratio of applied moment-to-plastic flexural capacity, Beams 1 and 2 showed almost the same amount of stresses produced in the concrete slab and steel beam. Beams made with HSS tube confinement showed almost the same amount of stresses as beams made with the closed- ties confinement. Transverse Slab Reinforcement Required To Resist the Transverse Bursting Force. The horizontal shear force at the interface is transferred from the steel beam to the concrete slab by direct bearing of the grout volume on the precast panel. This mechanism is similar to the mechanism of trans- ferring the bearing force of a posttensioned tendon to the end zone of a posttensioned concrete member. According to Arti- cle 5.10.9.3.6 of the LRFD specifications (7), the transverse bursting force is estimated using the following formula: (20) where Ts = the bursting force, Pu = the factored tendon load on an individual anchor (ul- timate horizontal shear force generated by a cluster of studs), a = the anchor plate width (width of the shear pocket), and s = the anchorage spacing (stud cluster spacing). For Beam 1 (2 ft spacing, four studs per cluster): Pu = 478.7 = 314.8 kip, a = 12 in., s = 24 in. 15.74 kip/ft Required conventional reinforcement: For Beam 2 (4 ft spacing, eight studs per cluster): Pu = 8 × 78.7 = 629.6 kip, a = 12 in., s = 48 in. 23.61 kip/ft Required conventional reinforcement: The required conventional reinforcement to resist the bursting force for Beam 1 or Beam 2 is smaller than the re- quired reinforcement determined according to the empirical design method given in Article 9.7.2 of the LRFD specifica- tions (7), which is 0.18 + 0.27 = 0.45 in2/ft/2 layers. Removal of the Precast Panels of the Full-Scale Beams The precast panels were removed by jackhammering the concrete around the shear pockets. The grout around the studs was then removed using a manual driller. Several ob- servations on the condition of the shear studs and the grout surrounding them were made (see Figure 101). • No air pockets were detected in the shear pockets or in the haunch. • No grout crushing was detected at the base of the studs. Also, the grout was fully bonded to the studs. • The studs were almost vertical. The maximum slope that was observed was about 5 degrees. • No cracks were detected at the weld at the base of the studs. A T f s s y = = = 23 61 60 0 39 . . kips ksi in /ft/2 layers2 Ts = × − ⎛⎝⎜ ⎞⎠⎟ = × × =0 2 629 6 1 12 48 0 2 629 6 0 75 94. . . . . .44 kip/4ft = A T f s s y = = = 15 74 60 0 26 . . kips ksi in /ft/2 layers2 Ts = × − ⎛⎝⎜ ⎞⎠⎟ = × × =0 2 314 8 1 12 24 0 2 314 8 0 5 31 4. . . . . . 8 kip/2ft = T P a s s u= − ⎛⎝⎜ ⎞⎠⎟0 2 1. 100

101 Beam #1 (2-ft stud cluster spacing) Beam #2 (4-ft stud cluster spacing) Figure 96. Bursting longitudinal cracks on top surface of the slab. Beam Stud Cluster Spacing (ft) Confinement Type Failure Mode Maximum Applied Load (kip) Load Required to Cause Flexural Failure (kip) Load Required to Cause Horizontal Shear Failure (kip) 1-North 2 Ties Flexural failure/ web buckling, Figure 95-a: Concrete crushing of the top fiber of the concrete slab at the concentrated load location. Also, a vertical crack formed at side surface of the slab at the section of the applied load. Four inclined cracks in the haunch at 45 degrees. One crack at each stud cluster located between the applied load and the exterior support. The web of the steel beam buckled at the exterior support. 588 574 459 1-South 2 Tube Flexural failure, Figure 95-b: Concrete crushing of the top fiber of the concrete slab at the concentrated load location. Four inclined cracks in the haunch at 45 degrees. One crack at each stud cluster located between the applied load and the exterior support. 600 574 459 2-North 4 4 Ties No failure occurred, Figure 95-c: The hydraulic jacks reached their maximum capacity. 600 843 356 2-South Tube No failure occurred, Figure 95-d: The hydraulic jacks reached their maximum capacity. 600 843 356 Table 11. Summary of the full-scale beam ultimate test results.

102 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0.00 0.20 0.40 M a pp lie d / M n (% ) 0.60 0.80 1.00 1.20 1.40 1.60 1.80 Displacement (in.) 2 ft spacing with Ties 2 ft spaci withTu ng be 4 ft spacing withTube 4 ft spacing with Ties Figure 97. Load-deflection relationship of the full-scale beams. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0 0.02 0.04 M a pp lie d / M n (% ) 0.06 0.08 0.1 0.12 0.14 0.16 0.18 Relative Dsiplacement (in.) 2 ft spacing with Ties 4 ft spacing withTube 4 ft spacing with Ties 2 ft spacing withTube Figure 98. Load-horizontal slip relationship of the full-scale beams.

103 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 -2,000 -1,000 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 Strain x 1E+6 D ep th (i n.) Mapplied/Mn = 17% Mapplied/Mn = 35% Mapplied/Mn = 52% Mapplied/Mn = 70% Mapplied/Mn = 87% Mapplied/Mn = 100% 17% 35% 52% 70% 87% 100% 100% 100% Beam #1-North (2ft with Ties) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 -2,000 -1,000 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 Strain x 1E+6 D ep th (i n. ) Mapplied/Mn = 17% Mapplied/Mn = 35% Mapplied/Mn = 52% Mapplied/Mn = 70% Mapplied/Mn = 87% Mapplied/Mn = 100% 17% 35% 52% 70% 87% 100% 100% 100% Beam #1-South (2ft with Tube) Figure 99. Strain distribution of the full-scale Beam 1.

104 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 -1,500 -1,000 -500 0 500 1,000 1,500 2,000 Strain x 1E+6 D ep th (i n.) Mapplied/Mn = 12% Mapplied/Mn = 24% Mapplied/Mn = 36% Mapplied/Mn = 47% Mapplied/Mn = 59% Mapplied/Mn = 71% 12% 24% 36% 47% 71% 71% 71% 59% Beam #2-North (4ft with Ties) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 -1,500 -1,000 -500 0 500 1,000 1,500 2,000 Strain x 1E+6 D ep th (i n.) Mapplied/Mn = 12% Mapplied/Mn = 24% Mapplied/Mn = 36% Mapplied/Mn = 47% Mapplied/Mn = 59% Mapplied/Mn = 71% 12% 24% 36% 47% 71% 71% 71% 59% Beam #2-South (4ft with Tube) Figure 100. Strain distribution of the full-scale Beam 2.

Guidelines for Design, Detailing, Fabrication, and Installation of Full-Depth, Precast Concrete Deck Panel Systems Guidelines for the design, detailing, fabrication, and in- stallation of full-depth precast concrete deck panel systems are given in Appendix C. The guidelines do not cover propri- etary full-depth precast concrete bridge deck panel systems. Individual deck construction projects may have their own unique features and constraints, which may affect the design, fabrication, and construction process. The reader should therefore evaluate the relevance of the provisions in accor- dance with the project requirements. Proposed Revisions to AASHTO LRFD Specifications Proposed revisions to Section 9 of the AASHTO LRFD Bridge Design Specifications (7) are given in Appendix D. The objective of the proposed revisions is to inform designers of the requirements pertaining to use of full-depth precast deck panel systems and thus promote use of these relatively new systems. 105 Figure 101. Shear studs after deck panel removal.