Recommended AASHTO Guide Specifications for ABC Design and Construction (2018)

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.

NCHRP Project 12-102 19 C H A P T E R 4 Technology Synthesis and Knowledge Gaps The project team compiled a significant amount of past research regarding ABC technologies, which made the proposed ABC Guide Specifications technically sound and useful. The results of the findings have been synthesized into major categories and reported in this chapter. Knowledge gaps have been identified within each technology and within the current bridge design specifications. There are two types of knowledge gaps identified and presented in this report. The first are gaps in the current AASHTO LRFD Bridge Design Specifications. This chapter explores these gaps and how the team proposed to fill them. The second type of knowledge gaps are with respect to each individual technology. In the process of reviewing past work, it is inevitable that the researchers uncovered items that need further study. These gaps are presented in each technology section of this chapter, with the hope of identifying potential additional research in the future. 4.1 Current Design and Construction Specifications The current AASHTO LRFD Bridge Design and Construction Specifications do not contain much ABC-specific guidance. There are a few items that are covered in some detail such as Precast Full-Depth Deck Panels and transverse connection recommendations for decked beam elements through the use of post-tensioning (recommendation only). There are also provisions that can be used for connections of elements. This is due to the approach that many ABC designers use a method called emulative design and detailing. Emulative detailing is based on designing connections that “emulate” conventional construction. This can be done through simple closure joints combined with lapped or hooked reinforcing bars. Mechanical reinforcing bar couplers that are also used in emulative detailing are also included in the current AASHTO provisions, which define the strength needs for “Mechanical Connections”. Even with the current provisions, designers are called upon to use engineering judgement when applying current AASHTO provisions in an ABC design. 4.1.1 Approach to Specification Format The goal of this research project to fill the knowledge gaps that exist in the current AASHTO LRFD Bridge Design and Construction Specifications. This was done in several ways: • Gaps supported by research: Significant research has been completed in the field of ABC. This report contains a list of significant research that was used to fill certain gaps in the current specifications. The research team synthesized the research, determined if the technology is mature to the point where it is ready for publication, and then formulated specification language in AASHTO format.

NCHRP Project 12-102 20 • Gaps covered indirectly by current AASHTO specification provisions: In many cases, ABC designs can be executed using existing AASHTO provisions; however, the appropriate provisions are not readily apparent. For example, the development of design connections between adjacent precast concrete full-depth deck panels is subjective. There is no significant research into the forces acting on these connections; however, there are recommendations included in the PCI State of the Art Report on Full Depth Precast Concrete Bridge Deck Panels (2011). These recommendations refer to current AASHTO provisions for deck designs based on engineering judgement and have proven to be sound. The development of ABC Design and Construction Guide Specifications filled both of these gap types. Where adequate research justified specific design and construction guidance, they were formulated into appropriate provisions. Where existing AASHTO specifications are appropriate, guidance was provided for the use of these provisions in an ABC project. There are two potential approaches to the latter: • Reference current AASHTO specifications: This approach includes a simple reference to the applicable AASHTO provisions by article or section number. This is similar to the current format of the AASHTO specifications. The benefit to this approach is that it keeps the size and detail of the provisions to the minimum, but requires that the designer use multiple specifications in a design. This approach can also become problematic if the current AASHTO specifications are reorganized in the future. Re-organization of the existing AASHTO specifications could require a re-write of the ABC Guide Specifications to ensure that the appropriate references are correct. One way to avoid this is to have the ABC Guide Specification reference provisions by name and not specific articles. This approach has been used on other AASHTO specifications. • Re-state the current AASHTO provisions in the ABC Guide Specification: This would involve reiterating current AASHTO provisions in the ABC Guide Specifications in part or whole within the ABC specifications. This approach solves the problem of requiring designers to work with multiple specifications. The problem is that as AASHTO LRFD provisions are modified, the ABC Guide Specification will become outdated, one provision at a time. The research team is recommending the first approach, where the ABC Guide Specification will reference the current AASHTO LRFD provisions by name. Re-organization of the AASHTO LRFD Specifications is rare, however re-writing of specific provisions is more common. Using this approach is also beneficial if in the future, the ABC Guide Specification become incorporated into the AASHTO LRFD Bridge Design and Construction Specifications. The following sections describe the synthesis of the research and documents that were identified in the literature search. The work is broken up into non-seismic research and seismic research. The documents were grouped into common technologies. Within each section, knowledge gaps in the current AASHTO provisions are identified. 4.2 Non-Seismic Technologies 4.2.1 Precast Concrete Full-Depth Deck Panels (FDDP) This technology consists of precast concrete deck elements that are cast to a thickness that is consistent with conventional cast-in-place concrete. The panels are erected onto the bridge girders and connected to the other deck panels and the bridge girders. The intent of these systems is to emulate a cast-in-place

NCHRP Project 12-102 21 concrete deck in most cases. The key difference between these systems and cast-in-place construction is the connection between the deck and the girders and the connections between the panels. The panels are usually cast off-site in a controlled setting and shipped to the site for erection onto the bridge primary members. The setting of the precast panels and the subsequent connections to the bridge girders make this a viable ABC method. This is because it eliminates the time necessary to form, place and tie reinforcing steel, and place and cure the entire concrete bridge deck. The current AASHTO LRFD Bridge Design Specifications address some but not all aspects of the design of full-depth deck panels. There has been significant research regarding FDDP. The database search uncovered a large number of documents (Wipf, et al., 2009, Badie & Tadros, 2008, Carter, et al., 2011, Araújo, et al., 2005, Aaleti, et al., 2013, Phares, et al., 2013, Wipf, et al., 2010, Shim, et al., 2001, SHRP2, (2014), Halverson, et al., 2012, Hanna, et al., 2010, Goodspeed, et al., 2011, Harryson, 2003, Graybeal, 2010, 2012 & 2014, Barr, et al., 2013, Porter, et al., 2010, Kassner, et al., 2007, Wallenfelsz, 2006, Scholz, et al., 2007, and Sullivan, et al., 2008). A review of the documents indicates that a significantly larger body of past work exists. These older research efforts are not readily available within existing databases such as the TRB TRID database. Fortunately, there are a few recent research projects that have synthesized the past work. The major research efforts that accomplish this are: • PCI State of the Art Report on Full-Depth Precast Concrete Bridge Deck Panels (2011) • NCHRP Project 12-65 Full-Depth Precast Concrete Bridge Deck Panel Systems (2008) These documents provide an excellent account of past work and the current state of practice for FDDP. The other research projects and documents found cover specific portions of this technology. For example, a Virginia Tech research project (Scholz, et al., 2007) specifically covers the connection of FDDP to prestressed concrete beams. A review of the research obtained through this study indicates that the design of full-depth deck panels is similar to the design of conventional reinforced concrete decks. The Designer must account for all aspects of the deck design but additionally needs to consider the following: • Panel tolerances and fit-up with respect to bridge cross slope • Connections between panels • Connections to the girders The following sections describe past research for the various aspects of FDDP design and construction. 4.2.1.1 Panel to Panel Connections There have been a number of significant research projects that have investigated the connections between FDDPs. The design of concrete bridge decks is primarily a “one-way” slab design using the AASHTO “strip method” where bending moments are calculated with the deck spanning between beams or girders. Additional reinforcements are added in other directions to distribute the wheel loads within the slab; however, these reinforcing bars are not specifically designed. Instead, a certain percentage of the flexural reinforcing are applied parallel to the supporting beams or girders. These connections can be categorized in two ways. The first is the connection between panels in the “span” or “strength” direction of the panel. This refers to the direction that contains the primary flexural reinforcement of the panel. The second category of connections is typically parallel to the supporting beams or girders and is often referred to as the “distribution” direction.

NCHRP Project 12-102 22 Panels Connected with Post-Tensioning: Post-tensioning has been used extensively for the connection of FDDPs in the distribution direction. Post-tensioning is not typically used for the span direction due to the relatively high bending moments that are generated in the design. The post-tensioning typically consists of Grade 270 pre-stressing strand tendons or threaded post-tensioning rods installed in ducts located at the centroid of the deck panels. Article 9.7.5 of the AASHTO LRFD Bridge Design Specifications contain provisions for the design of precast deck slabs on girders. Included in these provisions are recommendations for post-tensioning of deck panels in the distribution direction. It is believed that the genesis of these provisions comes from research on the connection of prestressed double tee bridges (Shahawy, et al., 1992). Post-tensioned systems require very specific steps in order to complete the construction. The basic steps are as follows: • Erect the deck panels to the specified line and grade. • Connect and seal the post-tensioning ducts between the panels. • Loosely place the post-tensioning tendons in the ducts. • Grout the transverse joints between the panels and wait for the grout to attain a strength that will allow for application of the post-tensioning forces. • Stress the post-tensioning tendons to achieve the desired level of stress in the joints. • Grout the post-tensioning tendon ducts. • Establish the composite connection of the panel to the supporting beams or girders. This is typically done with shear connectors in grouted pockets cast into the panels. It is important to precisely follow this procedure and apply the post-tensioning to the system prior to making the composite connection to the beam or girder. Otherwise, the post-tensioning force will induce a positive bending moment in the supporting beam or girder, potentially affecting the load-carrying capacity of the girder. There has been several research projects regarding the post-tensioning of FDDP (Wipf, et al., 2009 & 2010, Carter, et al., 2011, Hanna, et al., 2010, Goodspeed, et al., 2011, Barr, et al., 2013, and Porter, et al., 2010). Two projects investigated the use of tendons set externally to the deck panels and placed over the top flange (Wipf, et al., 2009 and Hanna, et al., 2010). This system was developed at the University of Nebraska and is referred to as “NUDECK”. The NUDECK system has an innovative open space over the top flanges of the supporting girders. Pre-stressing is used in the transverse direction, and mild reinforcing steel is used to act as compression struts to maintain the pre-stressing across the joints. Several generations of NUDECKs have been developed and tested. The longitudinal post-tensioning is external to most panels and is placed in the field directly over the top of the supporting girders. The post- tendons are anchored in the end panels. These systems do not require grouting of the post-tensioning ducts, which can expedite the construction. This system also allows for easier installation of shear connectors on steel beams, since the top flange of the girders are exposed. The design of anchorage zones for post-tensioning is similar to any post-tensioned structure. Anchorages for tendons are proprietary devices that are embedded in the ends of the deck system. It requires the designer to design the general zone reinforcing and check local zone reinforcing submitted by the contractor when the final post-tensioning system is submitted. The design of these anchorage zone areas is included in the AASHTO LRFD Bridge Design Specifications. Anchorage systems available on the market are specifically designed to accommodate a certain number of strands. Each system has a thickness that needs to be considered when starting the design, as the cover

NCHRP Project 12-102 23 requirements around the anchorage will need to be met. This may require that the slab thickness be increased above what is typically required for cast-in-place concrete. Additional details to consider are hand holes required in longitudinally post-tensioned deck panel systems. This allows for the post-tensioning conduits to be connected and wrapped keeping the grout in place in the post-tensioning conduits. Panels Connected with Mild Reinforcement (Distribution Direction): NCHRP Project 12-65 (Badie, et al., 2008) studied ways to connect deck panels in the distribution direction using simple shear keys combined with mild reinforcing bars, thereby eliminating the need for post-tensioning. The panel to panel connection developed under this project consists of steel bars placed into slots between adjacent panels. The slot is made with steel embeds along the edges of the panel that form slots in the top surface of the panel. Once the panels are set, reinforcing bar dowels are set in the slots, and the joint and slots are filled with non-shrink grout. This approach eliminates several of the construction steps required for post-tensioned connections, which makes this technology attractive. The Ontario Ministry of Transportation has also completed research on panel to panel connections (Au, et al., 2011). The researchers concluded that a variety of hooked and lapped bars could provide a strong and durable connection. They investigated hooked bars, short lapped bars with spiral confinement, and high early strength concrete. The only connection that did not perform well had short lapped bars with transverse tie bars. Panels Connected with Mild Reinforcement (Strength Direction): The key to the successful performance of a deck panel connection is the continuity of the reinforcing. There are number or ways to connect panels including the use of lapped or hooked reinforcing bars designed according to the provisions of the AASHTO LRFD Bridge Design Specifications. There is a growing movement to reduce the width of the closure pours using higher strength concrete materials. NCHRP Project 10-71 (French, et al., 2011) studied various ways to connect precast deck elements. Various mild reinforcement strategies were studied including the use of headed reinforcing bars combined with high strength grout. Other work has investigated the use of UHPC, which is a mix design with a large amount of steel fiber that produces a compressive strength in excess of 20 ksi (Aaleti, et al., 2013; SHRP2, 2014; Harryson, 2003; Parant, et al., 2006; Graybeal, 2010) (2). Using this material, the development and lap length of traditional reinforcing bars can be greatly reduced, thereby reducing the width of reinforced closure pours. The Federal Highway Administration (FHWA) has synthesized this work into a concise document entitled Design and Construction of Field-Cast UHPC Connections (Graybeal, 2014). This document is an excellent source of design and construction guidance and will most likely be the basis for the development of the ABC Guide Specifications on this subject. 4.2.1.2 Panel Connections to Beams or Girders The AASHTO Bridge Design Specification does cover the design of the connection of full-depth deck panels to steel or concrete girders, but not in significant detail. The specification recommends the placement of shear connectors or shear reinforcement within discrete pockets in the panels. Panels are typically cast flat and placed on top of girders along the desired cross slope of the roadway. Leveling devices have been used to distribute panel dead loads to girders and to establish correct deck

NCHRP Project 12-102 24 grades. The detailing of forming of the haunches (gap between the girder and bottom of deck) has varied across the country. Some designers detail the forming technique and others do not. It is the opinion of the project team that this issue falls under “contractor means and methods” (similar to other forming details). If there is a need to limit the forming techniques, the plans can indicate these limitations. Designers can also show a recommended forming technique, but allow for contractor modifications to the forming details shown. The design of the connectors (size and spacing) can be found within the pertinent sections of the AASHTO LRFD Bridge Design Specifications. Designers typically place the shear connectors within pockets at fixed intervals. The shear resistance is adjusted by varying the number of connectors within a pocket. The composite connection to the girders has been studied under several projects (Badie, et al., 2008, Carter, et al., 2011, Araújo, et al., 2005, Wipf, et al., 2010, Shim, et al., 2001, Graybeal, 2012, Barr, et al., 2013, Wallenfelz, 2006, Scholz, et al., 2007, and Sullivan, et al., 2008). These studies have included connections to both steel and concrete girders. Several different pocket configurations have been studied. Most bridges built to date have had full-depth pockets. There is concern in the industry regarding the rideability and durability of full-depth pockets on bridges with exposed bridge decks. NCHRP Project 12-65 (Badie, et al., 2008) verified that it is possible to use partial-depth pockets that do not penetrate the top surface of the deck. FHWA studied partial-depth pockets that are continuous over the girder flange (Graybeal, 2012). The pockets were filled with UHPC. The general conclusion that can be drawn from all of this work is that full composite action can be attained and that the current AASHTO LRFD Bridge Design Specifications can be used for the design of the shear connectors. 4.2.1.3 Overlay Systems Many agencies use overlay systems on precast bridge decks and decked beam elements. The overlays are used to protect the deck from exposure to deicing agents, and/or to provide a smooth finished riding surface over the relatively rough surface of the constructed precast deck. The roughness of a deck is a function of the fabrication and erection tolerances of the panels, the chorded nature of building a vertical curve using flat panels, and excess grout used in joints and pockets. If a precast deck is to be left bare, or if a thin overlay is proposed, deck grinding is in order, especially for a bridge deck subjected to higher vehicle speeds. Driving over a rough deck may lead to premature deterioration of the deck and increased dynamic load on the girders. It is possible to use thick overlays without grinding, since the overlay system can smooth out the roughness in the constructed deck. There has only been one study on deck overlays placed on precast deck panels (Pantelides, et al., 2011). This study investigated five thin overlay systems. The Utah DOT has experienced problems with deck leakage and thin overlay delaminations on several bridges. This study investigated long-term fatigue loading on deck panel joints covered with thin overlays. The conclusion of the study indicated that the thin overlays could function well, provided that the substrate was properly prepared prior to application of the overlays.

NCHRP Project 12-102 25 4.2.1.4 Knowledge Gaps 1. Design of FDDP: The process for designing FDDPs is not clearly indicated in the AASHTO LRFD Bridge Design Specifications. To date, the design of FDDPs has been based on an emulative approach. The PCI State of the Art Report on Full Depth Precast Concrete Bridge Deck Panels (Carter, et al., 2011) recommends designing the panel reinforcing using standard AASHTO deck design provisions. Future research could confirm the applicability of the emulative approach to deck panel design. 2. Losses in Post-tensioning: The current AASHTO provisions specify a minimum effective prestress of 250 psi across a grouted joint. The provisions do not include specifications or guidance on the calculation of long-term losses in the deck panels. The composite action of the deck and the girder greatly complicate this calculation. Some designers have neglected long-term losses and had satisfactory results. The PCI State of the Art Report on Full Depth Precast Concrete Bridge Deck Panels (Carter, et al., 2011) recommends a simple lump sum loss approach. Technically, a detailed time step analysis could be done in order to account for the age of the panels and the potential transfer of post-tensioning force due to creep of the panels. This approach seems onerous and not conducive to promoting the use of these systems. Through future research, it may be viable to develop reasonable lump sum losses. 3. Effect of Negative Moment Regions on FDDP Connections: Negative moment regions can lead to live load tension in typical bridge decks. The current AASHTO provisions for FDDP design do not clearly address how to accommodate these stresses. Applying a net post-tensioning compressive stress of 0.250 ksi in a continuous girder bridge often results in unacceptable levels of post- tensioning. There is a need to study the effect of continuity on post-tensioned connections. In theory, the live load required to generate the maximum negative moment places the trucks away from these joints. Based on this approach, it may be reasonable to reduce the allowable stress on the post- tensioned joints in the negative moment regions. This issue could benefit from further research. Some of the mild reinforced connections studied were based on simple span bridge layouts. The effects of negative moments on these joints has not been thoroughly studied. As with the post- tensioned connections, the effect of negative moments on these joints should be investigated further. 4. Duct Friction and Intermediate Post-tensioning Anchorages: There are practical limits for the length of post-tensioning systems, especially for the smaller duct sizes typically used in precast FDDPs. The tolerances of the duct connections at each joint will inevitably lead to slight misalignments which cause friction between the duct and the tendons. The authors are aware of one project in New Hampshire, where the duct friction was excessive, leading to a project delay. There have not been any studies on the effect of duct misalignments on these friction forces. The AASHTO LRFD Bridge Design Specifications for duct friction are based on match casting. Additional guidance on this issue is needed. The issues with duct friction may lead to the need for intermediate anchorages for deck post-tensioning systems. The effect of these anchorages on the stresses in the deck and the supporting girders is not well known. Long-term creep of the panels may result in a re-distribution of the anchorage forces into the girders, which may or may not be detrimental. 5. Effect of Skew on Post-tensioned Panels: The AASHTO LRFD Bridge Design Specifications limit the skew of reinforcing in bridge decks to 25 degrees. This requirement can limit the use of skewed deck panels. For bridges with skews higher than 25 degrees, the panels are normally made perpendicular to the supporting girders. The effect of skew on post-tensioned panels has not been studied in great detail. There are two concerns with this situation: • Loss of perpendicular prestress in the panel joints. • Failure of the joint grout during post-tensioning. 6. The first issue can be overcome by increasing the post-tensioning force to account for the skew. The minimum prestress of 0.250 ksi should be stated to be perpendicular to the joint. The second issue is more complicated. The concern is that the panels might slip during post-tensioning due to the

NCHRP Project 12-102 26 prestress parallel to the joint. To the knowledge of the team, no skewed panels have slipped to date; however this has never been studied in detail. 4.2.2 ABC Substructures The typical prefabricated substructures consist of precast elements that are joined together in the field to comprise the finished structure. Emulative detailing is the most common approach used for this model. This involves the emulation of cast-in-place concrete using precast concrete elements. By using emulation, the design of the elements is essentially the same as with cast-in-place construction. The connection of the various elements makes up the various technologies that are in use. The section below on Seismic Technologies gives detailed descriptions of these connections as they relate to seismic considerations. The most common connections as they relate to non-seismic technologies are described below. It should be noted that it is perfectly acceptable to use seismic connection in a non-seismic location. The research completed to date on the connections identified in this report as “seismic connections” was centered on seismic performance. 4.2.2.1 Bar Coupler Connections Methods for joining precast elements such as pier caps, columns, footings and walls often require splicing steel reinforcing. Prefabricated elements are often connected by means of mechanical reinforcing splicers. The AASHTO LRFD Bridge Design Specifications define a mechanical coupler as a device that can resist 125% of the specified yield strength of the connected bar. There are many proprietary mechanical connectors in the market; however, only a few are easily adaptable to precast construction. Threaded couplers are an example of a device that does not work well in a precast element. The turning of the bar is simply not possible unless there is a significant closure pour at the joint. Couplers with a combination of grouting and threaded connections can work well with precast elements if the threaded portion of the coupler is embedded in one of the joined elements. The most common type of bar coupler is the grouted splice coupler. Grouted splice couplers rely on a grout-filled device to transfer load between two pieces of reinforcing steel. Grouted splice couplers are proprietary ductile iron castings that have bars inserted on each end. The transfer of bar force is akin to a lap splice in that the forces are transferred from the bar to the coupler and then back to the bar. One precast element acts as the host and the other is the element to be joined. The sleeve is cast into the host element with the reinforcing to be spliced inside the sleeve. The sleeve is positioned within the host element with an opening at the face of the concrete to receive protruding reinforcing from the element to be joined. The connection between the host element and the element to be joined is made when the protruding reinforcing from the element to be joined is inserted into the sleeve during construction and the sleeve is filled with grout. Grouting can be done prior to positioning the element to be joined, which may be referred to as “pre-grout”, or it can be done after placing the element to be joined by means of a grout pump. There have only been a few research projects involving the testing of grouted couplers. More research is on-going, specifically for use in seismic applications. One manufacturer performed testing of their grouted splice couplers on partial scale columns (NMB NISSCO, (undated). Classic column plastic hinging was studied. Several joint configurations were studied as well as a control specimen without couplers. The results of this study indicate that connections with the couplers can emulate traditional cast-in-place construction.

NCHRP Project 12-102 27 4.2.2.2 Grouted Post-Tensioning Connections Grouted post-tensioning duct connections are similar to grouted couplers; however, there are several distinct differences. The ducts are cast into the end of receiving element. Protruding bars from the second element are inserted into the ducts, and then the remaining void in the duct is filled with grout. The primary difference between grouted post-tension duct connections and grouted splice couplers is that reinforcing steel is not being spliced; it is simply being developed into the host element. Grouted couplers transfer the bar force to the coupler, which is a structural element. Grouted ducts are non- structural and simply used to transfer the force from the bar to the surrounding concrete. This works well for connecting into mass concrete (pier caps or footings for example), however connecting to parallel reinforcing bars is more difficult (pier column). This requires adding splice bars outside of the coupler. There have been several research projects that investigated the use of grouted post-tensioning ducts for connections. The earliest testing was completed at the University of Texas under several projects. The first project was entitled Development of a Precast Bent Cap System (Matsumoto, et al., 2001). The second was entitled Anchorage Requirements for Grouted Vertical-Duct Connectors in Precast Bent Cap Systems (Brenes, et al., 2006). This research demonstrated that it was feasible to reduce the development length of reinforcing bars in mass concrete using ducts. In addition to reducing development lengths, the ducts helped to reduce the cracking in the element at ultimate load. Details were developed for precast pier caps using these details; however, the testing was not intended for seismic applications. 4.2.2.3 Pocket Connections The connection of a precast element into a larger mass concrete element can be achieved through the use of pocket connections. Pockets or voids are formed into the larger element (pier cap, footing, integral abutment stem). The connected element is typically a column with protruding reinforcing bars, piles, or drilled shafts with protruding reinforcing bars. In this type of connection, the development of the connected element into the pocket is achieved through the placement of grout or concrete into the pocket after setting the elements in position. There are two common types of pockets in use, formed pockets and corrugated metal pipe pockets. Formed Pockets: Formed pockets are typically used to provide a connection between precast pier columns and precast pier caps. The Texas DOT has developed and tested several details that are applicable to bent cap construction including a project entitled Development of a Precast Bent Cap System (Matsumoto, et al., 2001). This research showed the viability of pocket connections. Recommendations for development length of the embedded reinforcing into the pockets was provided. Corrugated Metal Pipe (CMP) Pockets: CMP pockets are similar to formed pockets in that reinforcing steel or pile are inserted into a void and connected by means of concrete or grout. The pocket is formed by using a standard corrugated metal drainage pipe (galvanized steel), which is beneficial for several reasons. The CMP can transfer forces between the elements through the corrugations by action that is similar to aggregate interlock, and the pipe can help to confine the reinforcing steel in the connection. Iowa State University studied the use of CMP voids for connection of piles to integral abutment stems (Wipf, et al., 2009). Testing indicated that these connections are capable of resisting very large loads. Another important finding was that the CMP pocket did not have a measurable effect on the design of the element in which it was embedded.

NCHRP Project 12-102 28 The use of CMP pockets was further studied for seismic applications, and the performance was found to be very good. Additional confinement reinforcement is recommended in order to develop adequate seismic performance of the connection. 4.2.2.4 Geosynthetic Reinforced Soil/Integrated Bridge System (GRS/IBS) GRS/IBS is a bridge system that was developed by the FHWA Turner Fairbanks Research Center (Adams, et al., 2012). It is both cost-effective and fast to construct, which makes it an ideal ABC technology. The semi-integral abutment system can best be described as a flexible soil mass that can support the bridge superstructure and accommodate reasonable thermal movements without expansion devices (joints and bearings). The soil mass is akin to a large elastomeric bearing in that it flexes with thermal movements of the bridge without the need for a sliding interface between the superstructure and the substructure. The advantages of GRS/IBS are as follows: • Construction can be accomplished quickly with very little equipment. • The use of site cast concrete is often not required. • The integrated soil mass supports both the superstructure and the approach embankment, therefore approach slabs are not required. • Deck end expansion joints and expansion bearings are not required. FHWA has developed design and construction guidelines. These will be used for the development of the ABC Guide Specifications for this technology. 4.2.2.5 Knowledge Gaps 1. Current AASHTO Provisions for Mechanical Couplers: The current AASHTO LRFD Bridge Design Specifications include provisions for mechanical reinforcing couplers based on a resistance of 125% of the specified yield strength of the connected bars. There are also limitations on the use of these couplers in columns (staggering, spacing, location). The Building Code Requirements for Structural Concrete (ACI-318) also has these same provisions for couplers of this strength (referred to as Type 1); however, it also includes provisions for higher strength couplers that can develop 100% of the specified tensile strength of the bars (referred to as Type 2). ACI-318 lifts the use limitations for the Type 2 couplers. Similar provisions should be considered for the AASHTO LRFD Bridge Design Specifications. Under this project, specifications will be written to be consistent with ACI -318. Two types of couplers can be specified. The current restrictions on mechanical couplers can be kept for the Type 1 couplers. The Type 2 couplers can be specified to have unrestricted use. Special considerations for seismic performance will also be added. 1. AASHTO Provisions for CMP Pipe Void Connections: The development of the CMP pipe void for integral abutments was based on limited testing at Iowa State University (Wipf, et al., 2009). The testing was based on a specific bridge project. The CMP void connection was not tested to failure, and the testing was limited to a few specimens. Design provisions were not developed under this research. There is a need to perform more testing and to develop design provisions for this popular connection. The research team recommended the use of interface shear or punching theory for this connection.

NCHRP Project 12-102 29 4.2.3 Decked Beam Elements The typical decked beam element system consists of beam elements that are fabricated with an integral deck. The decked beams are joined together in the field to create the finished superstructure. The critical feature of these elements is the connection of the beams at the deck level. The primary advantage to decked beam elements is the elimination of the complicated forming of the deck at the bridge site. The following is a listing and brief description of the various decked beam elements that are in use: Deck Bulb Tee Beams: A deck bulb tee beam combines the stem and bottom flange of a typical bulb tee beam and with an extended top flange that is the full depth of the deck. The ends of the flanges are connected with a closure pour or a grouted shear key combined with welded tabs. The deck bulb tee beam has the advantage of being an efficient load-carrying section with an integral precast deck. Modular Steel Deck Beams with an Integral Concrete Deck: Modular Steel Deck Beams are made by casting the concrete onto beams at an off-site location. Typically a steel decked beam element will have two beams in order to facilitate shipping and handling. A single beam decked element can also be used. In this case, the approach would be similar to a deck bulb tee beam. These prefabricated modular steel beam elements can then be shipped and lifted into place at the bridge site. After placement, the joints between the deck portions of the modular steel beam elements are connected with closure pours. Steel beams can offer an advantage over precast decked beams since they can be cambered to achieve the vertical profile required by the roadway. This can be done with precast; however, it is more difficult. Steel decked beam elements are fairly wide. This can lead to increased shipping and handling weight, which can limit their practical use to short and moderate span bridges. Precast Multi-Stem Beams (Double Tee or Triple Tee): Precast multi-stem beams have been widely used in parking structures for their efficient sections and constructability. However, the loads supported by typical parking structures are significantly lower than highway bridges and also have shorter span ranges than highway bridges. With the interest in ABC technology, several new tee beam sections have been developed including Florida Double Tee beams, Washington State Double Tee and Triple Tee Beams, and the Northeast Extreme Tee beams (NEXT beams). 4.2.3.1 Deck Edge Connections: Decked Beam elements require a connection at the edge of the deck, where the adjacent beams meet. Several different types of deck edge connections are currently in use. Some have been tested, while others were designed based on the Strip Method in the AASHTO LRFD Bridge Design Specifications, assuming the deck portion of the beam to be akin to a cast-in-place concrete deck. The following sections outline the more common approaches used for these connections. Conventional Reinforced Concrete Closure Joints with Conventional Reinforcing: Closure pours can be made by placing reinforcing so that it protrudes from the side of the precast element into the joint. After the precast elements are positioned, the void is filled with concrete. The joints are structurally analyzed and reinforced for all applicable loads, such as shear, moment and axial loads, as well as minimum reinforcing requirements. Texas A&M University completed a research project for the Texas DOT on various conventionally reinforced concrete connections (Brush, 2004). This project looked at several configurations of lapped

NCHRP Project 12-102 30 and hooked bars for decked beam connections. Very short lap lengths were used for hooked bars, which did not perform well. Shear failures were also found near the interface between the closure joint and the precast panel. The study indicated the need for shear keys at closure joints and a need to properly detail continuity of reinforcing bars. The Ontario Ministry of Transportation has completed research on precast panel to panel connections (Au, et al., 2011), which is also applicable to this connection. The researchers concluded that a variety of hooked and lapped bars could provide a strong and durable connection. They investigated hooked bars, short lapped bars with spiral confinement, and high early strength concrete. The only connection that did not perform well had short lapped bars with transverse tie bars. Construction methods play a critical part in the performance of closure pours. Because the material used for closure is placed in a confined space, between two connecting elements, shrinkage of the material and subsequent cracking has been a major concern. Grouted Reinforced Closure Pours: These connections differ somewhat from the concrete closure pours in that they make use of high performance grout materials. These materials allow for smaller width joints and shorted reinforcing bar lap lengths. The major types of these connections include: Grout with Headed Bars or Hooked Bars: The connection of the deck edges has been studied under several projects. The most significant work was included in NCHRP Project 10-71 entitled Cast-in-Place Concrete Connections for Precast Deck Systems (French, et al., 2011) and NCHRP Project 12-79 entitled Design and Construction Guidelines for Long-Span Decked Precast, Prestressed Concrete Girder Bridges (Oesterle, et al., 2009). This work investigated a number of connection types including the use of headed and hooked bars placed in a narrow grouted joints. The research showed that the bars could be fully developed as a moment connection. The results of these studies indicate that it is possible to develop a #5 reinforcing bar lap splice in approximately 6” using headed reinforcing bars or hooked bars. The headed bars are widely and readily available, as well as cost-effective; however, the diameter of the head on the bar can lead to reduced concrete cover. UHPC with Straight Bars: UHPC is a concrete mix containing steel fibers that have both a high compressive and high tensile strength compared to conventional concrete. UHPC can develop reinforcing bars in very short lengths reducing lap lengths. The use of UHPC has the advantage of the simplicity of a closure pour with straight bars, with a reduction in the required width of the joint due to shorter lap splices. A further advantage to UHPC is that it provides a more durable connection than those using materials that are more conventional. Research on UHPC joints was not limited to decked beam elements. The testing on precast full-depth deck panels is consistent with the needs for decked beam elements. Welded Plate Connections: Welded plate connections are similar to double tee connections that are used in vertical construction structures such as precast parking structures. Details have been developed for bridge applications that combine the welded tabs with a grouted joints. These structures have typically been used on low volume roads. There has been one major study of welded tab connections undertaken at Texas A&M University entitled Lateral Connections for Double Tee Bridges (Jones 2001). This study was undertaken by the

NCHRP Project 12-102 31 Texas DOT to investigate better connections for precast double tee bridges, which have had a history of cracked joints and leakage. The study tested several details. The recommended detail includes the use of sloped plates cast in the edges of the double tee flanges at 5 foot intervals. When the beams are erected, these plates form a V-shaped groove. The connection is made by welding a round steel rod between the two plates. The shear keys between the plates are then grouted to complete the connection. Similar to closure pours, construction methods play a critical role in the performance of welded plate connections. Adequate grouting of the key is required to provide the mechanical interlock of the joint. Additionally, overlay and waterproofing systems or sealing of cracks can be utilized to reduce problems with leakage through the grouted joints. Concern has been expressed by several agencies over the potential for fatigue cracking of the welded tabs. Some agencies such as Texas DOT and Washington State DOT only use these types of connections on bridges located on secondary roads, which are low- volume secondary highways, county roads and local roads. Transverse Post-tensioning: Transverse Post-tensioning used in conjunction with grouted shear keys is a system used to connect precast bridge elements. Use of post-tensioning for decked beam elements is allowed by the AASHTO LRFD Bridge Design Specifications to ensure that the units act together, and a minimum prestress value of 0.25 ksi is recommended. This recommendation appears to have come from research done in the 1990s on transversely prestressed double tees by the Florida Department of Transportation (Shahawy, et al., 1992). From the results of their load testing, they concluded that maintaining 0.20 ksi prestress helps the units act together and provides good punching shear resistance. Camber Differential Effect on Deck Connections: There are three sources of potential differential elevations across a deck edge connection. Fabrication tolerances can lead to differential elevations across the joints. Camber differentials due to camber growth in prestressed girders can be more difficult to accommodate, since there is more variation from design calculations. Skew effects can also lead to camber differential across joints. Camber differential can lead to fit-up problems, lack of cover on reinforcing bars and uneven riding surfaces. Pre-loading of beams can be recommended, however this will lock in long-term stresses. The designer needs to study any proposed preload operation and investigate the stresses that will be locked into element. Limited grinding of the deck after completion of the superstructure construction may be acceptable, provided that there is ample cover on the deck reinforcement. Overlays are another option for producing a smooth deck cross slope. 4.2.3.2 Multi-Span Structures: It is possible to design prefabricated multi-span structures using decked beam elements without the need for deck expansion joints at the interior supports. In most cases, these structures are not designed or detailed to resist dead load forces via continuous beam action. The prefabricated units are set on the substructures and connected in various fashions to form a jointless structure. Two basic concepts include: Simple Span for Dead Load, Continuous for Live Load: This method has been employed by the precast concrete industry for decades. In conventional construction, the beams are set and the deck cast without a connection of the beams at interior supports.

NCHRP Project 12-102 32 At that time, a connection is made, and the resulting superstructure is able to carry live load via continuous beam action. This concept is easily applied to prefabricated decked beam elements, since the deck is cast prior to erection. Connections between the prefabricated beam elements are made via closure pours after erection to form a continuous superstructure. There has been much research on this connection for precast prestressed beams. This approach is well documented in the current AASHTO LRFD Bridge Design Specifications, therefore it is not included in this project. There have been several research project that studied this connection for steel girder bridges. The first is entitled Development of a Steel Bridge System - Simple for Dead Load and Continuous for Live Load (Azizinamini, et al., 2005) that was undertaken at the University of Nebraska. This research investigated making this connection through a combination of butted plates and a cast-in-place reinforced concrete diaphragm. Several variations of the connection were tested. One detail was included in an actual bridge construction project in Nebraska. Another detail was developed under the Strategic Highway Research Program (SHRP2 R04) project entitled Innovative Bridge Designs for Rapid Renewal (SHRP2, 2014). This detail includes a combination of bolted splice plates and a continuous deck. A threaded rod was added near the top flange of the girders to reduce deck cracking induced by live load negative moments. Link Slabs: Link slabs are a technology that was developed by the precast industry in the 1990’s. It is a method to make simple spans jointless without introducing a moment connection at the supports. The link slabs run across the pier joints and are designed to accommodate composite dead load and live load rotation without generating significant moments and cracking. The first research regarding link slabs was undertaken by the North Carolina State University and published in an article in the PCI Journal entitled Behavior and Design of Link Slab for Jointless Bridge Decks (Caner, et al., 1998). The approach for detailing link slabs is that a portion of the normal deck slab composite connection is detached (left unbonded) near the interior support in order to accommodate beam rotations without significant cracking. The debonding is achieved by disconnecting the slab from the beam through the plane at the top of the beam surface. This is typically achieved by a thin layer of polyethylene. The unbonded zone was recommended to be approximately 5 percent of the adjacent spans. Several other research projects were undertaken by the Michigan DOT. The University of Michigan studied link slabs under a project entitled Field Demonstration of Durable Link Slabs for Jointless Bridge Decks (LI, et al., 2005). This study investigated high performance concrete materials for the link slab regions called Engineered Cementitious Composite (ECC). Western Michigan University studied the effects of skew on link slabs under a project entitled High Skew Link Slab Bridge System (Aktan, et al., 2011). Texas A&M University also investigated the use of link slabs for decked beam bridges as part of a research project for the Texas DOT entitled Connection of Modular Steel Beam Precast Slab Units with Cast-In-Place Closure Pour Slabs (Brush 2004). FHWA and the New York State DOT have investigated the use of thin link slabs made with UHPC (Graybeal 2014). It is believed that this has only been used on one bridge in New York. All studies have shown that link slabs are a viable system that can create a jointless bridge structure with very simple details. Simple details are normally less expensive and faster to build; therefore, this technology shows significant promise.

NCHRP Project 12-102 33 4.2.3.3 Knowledge Gaps 1. Non-Shrink Grouts for Closure Joints: There is a need to have better materials specifications for non- shrink grout. Current ASTM specifications are not adequate for the uses that are in place. Most non- shrink grouts were developed for placement under machine bases. Using grouts for larger and more voluminous spaces is simply not the same. Research at Virginia Tech studied the performance of certain grouts (Sholtz, et al., 2007). FHWA has also been working on a performance specification for grouts. It may be possible to develop specification acceptance criteria based on the work at Virginia Tech and FHWA. The team will work with the FHWA Turner Fairbanks Research Center to develop a reasonable grout specification. 2. Link Slabs: There currently are no provisions for the design and construction of Link Slabs in the AASHTO LRFD Bridge Design or Construction Specifications. Using the research noted above, the project team has developed specification language to fill this gap. 3. High Early Strength Concrete: The use of high early strength concrete is becoming common in ABC projects made with prefabricated bridge elements. In order to achieve high early strength, many admixtures are employed. Some high early strength concretes have a tendency to crack excessively, thereby reducing the long-term performance of the closure pour concrete. As with grouts, performance specifications would be helpful to designers and contractors. The current AASHTO LRFD Construction Specifications do not address high early strength concretes. The team has developed performance specifications for this type of concrete based on existing agency specifications that have been successfully used. 4.2.4 Bridge Systems Bridge systems consist of two major technologies, SPMTs and Lateral Sliding. Several SPMT and Lateral Slide manuals were found during the literature search including the Utah DOT Structures Design and Detailing Manual (2015), the FHWA Manual on Use of Self-Propelled Modular Transporters to Remove and Replace Bridges (2007), the SHRP2 Report ABC Standard Concepts: The Lateral Slide(2015). These documents represent the current state of practice in bridge systems. No significant actual research projects were discovered. 4.2.4.1 Knowledge Gaps 1 Bridge systems represent a construction methodology as opposed to a design methodology. The design of the actual bridge in a Bridges System is essentially the same as with conventional construction. The documents cited above cover methodology of bridge systems, however they do not cover the dynamic forces imparted on the bridge and falsework during the installation. 2 NCHRP had identified the need to fill this knowledge gap with the issuance of NCHRP Project 12- 98. This project includes two sub-projects that are applicable to bridge systems. The first sub-project involves the study of dynamic effects on bridge systems. This work involved the investigation of the dynamic loads generated during SPMT transporter movements. The hypothesis for this work was that an SPMT transporter creates forces that are similar to seismic events. The top of the transporter is akin to ground motions during seismic events. Accelerations and decelerations occur during starting and stopping and when traversing rough terrain. The testing has proven the hypothesis. The second sub-project in NCHRP 12-98 involves the study of sliding friction forces in lateral slide bridge systems. The work included friction testing of various commonly used sliding systems including both static and dynamic friction. The research team has developed guidelines for the design of dynamic forces for bridge systems entitled “Guidelines for Dynamic Effects of Bridge Systems” that is scheduled to be published in 2017. The development of the guide specifications for NCHRP Project 12-102 will reference this document.

NCHRP Project 12-102 34 3 NCHRP Project 12-98 is being completed in parallel with Project 12-102. Fortunately, the Principal Investigator for Project 12-98 is Michael P. Culmo, the Principal Investigator for Project 12-102. Through this connection, the development of guide specifications can account for the work in Project 12-98. The project panel has decided to incorporate the Project 12-98 guidelines into the proposed guide specifications through reference. 4.2.5 Lightweight Concrete Lightweight concrete offers several potential benefits to ABC. First, it can be used to reduce the weight of individual precast elements. This can reduce costs by requiring smaller cranes and falsework. The second potential benefit is accelerated curing of closure joint concrete. There is a need on some ABC projects to place closure joint concrete and quickly re-open the structure to traffic. The lack of significant curing can lead to a reduction in the long-term performance of the concrete. There has been significant research on materials such as lightweight fine aggregate to provide internal curing. The theory is that porous lightweight aggregates can store water that can be used to supply water to the concrete matrix during the hydration process, thereby curing the concrete internally. This material combined with other admixtures may have the potential to create mixes that need little or no curing beyond strength gain. There have been several research studies regarding the use of lightweight fine aggregates for internal curing (De La Varga, et al., 2010), Ideker, et al., 2013), and Rangaraju, et al., 2014). This approach to accelerated curing shows promise to improve durability of closure joint concrete mixes. 4.2.5.1 Knowledge Gaps 1. The full benefit of accelerated curing for closure joints through the use of lightweight fine aggregates needs to be further investigated. The amount of lightweight aggregate needs to be identified through materials testing research. 4.2.6 Precast Segmental Construction The research team is aware of several projects where segmental technology was employed for portions of the structure other than the superstructure. Post-tensioned segmental piers have been built in a number of locations including Florida, New York, and Washington State. Precast segmental construction is a well-established ABC technology. Provisions for design and construction of segmental construction are already included in the AASHTO LRFD Bridge Design Specifications and the AASHTO LRFD Bridge Construction Specifications. 4.3 Seismic Technologies 4.3.1 Bar Coupler Connections A bar coupler joins two reinforcing bars. In most cases the bars are collinear, so the coupler forms a mechanical butt-splice, although recently a coupler has been developed that consists of a lap splice within a confining sleeve. Chapter 21 of ACI 318-11 defines two splices: Type I and Type II. Only Type II, which must be able to develop the specified ultimate strength of the bar, is allowed in the plastic hinge zone. AASHTO LRFD permits mechanical couplers but imposes a slip criterion for service loads, and the AASHTO Seismic Guide Specification contains no limitations. Caltrans does not now allow bar couplers in the plastic hinge region.

NCHRP Project 12-102 35 Most couplers use one of the following technologies to transmit tension forces from one bar to the other. They are: • Grouted sleeves • Screw Threads • Headed bars with threaded couplers • External clamping screws. Combinations of these technologies may also be used, such as grouted sleeve for one bar and a threaded connection for the other. Source: Pantelides et al (2014) Figure 4.3.1-1 Typical Bar Coupler Application Source: Tazarv and Saiidi-Bar (2016) Figure 4.3.1-2 Mechanical Bar Coupler Examples

NCHRP Project 12-102 36 4.3.1.1 Types of Bar Couplers A grouted sleeve coupler is composed of a steel sleeve enclosing two bars placed end to end and filled with high strength grout. Tension is transferred from bar to sleeve to bar via the bond of the bars to the grout. The sleeve has internal lugs to improve its bond. A recently developed sleeve consists of two lapped bars within a grouted confining sleeve. A threaded sleeve connection is constructed with male threaded bars screwed into female threaded sleeves. Threads may be tapered to reduce the number of turns required and to reduce the stress concentration caused by threading. Headed bars with mating sleeves are connected via a threaded coupling piece. A shim is sometimes used to ensure bearing between the two bars. External clamping screw sleeves consists of a steel sleeve enclosing two bars placed end to end with screws driven radially through the sleeve into the bars. Tension is transferred from bar to sleeve to bar via shear in the screws. Swaged couplers are not considered in this document. 4.3.1.2 Design Considerations The use of bar couplers has an impact on the design of precast elements that are not encountered in traditional cast-in-place construction, which does not have a need for them. The coupler is inevitably larger, and therefore stiffer, than the bar, so its presence affects the distribution of strains along the bar. The performance of the connection, and particularly its ductility capacity, are likely to be affected by the location, e.g. in the column or footing/cap beam. The performance of the connection is also likely to be impacted by the orientation of the coupler. Placement of the coupler in the footing/cap beam leads to greater ductility capacity, but more column damage, than placement in the column. Another design consideration is the requirement of local debonding. Local debonding of the bar next to the coupler may be necessary to avoid strain concentrations and premature bar fracture. Additionally, if the coupler is placed within a column, a step change must be made in the dimensions of the spiral reinforcement to accommodate the extra width of the couplers. Care must be taken in design and construction to avoid the occurrence of failure at the discontinuity between spiral diameters. Additionally, research has shown that staggering couplers within a plastic hinge region contributes to adjacent bar buckling (Phillippi, et al., 2013). Grouted Sleeves: Grouted sleeves have greater ductility and lower strength than threaded connections made with tapered threads. Thus, the strength of a combination coupler will be governed by the weaker element, i.e. the grouted end, and the ductility will lie between that of the two types. (Jansson 2008) and Ameli, et al., 2014). Design should also consider measures to improve performance. The performance of grouted sleeves embedded in the column above the footing can be improved by raising them on a CIP pedestal of height at least one half the column diameter and debonding reinforcement through the pedestal (Tazarv, et al., 2014). Better performance is also observed as the sleeve diameter decreases relative to the bar size, due to improved confinement within the sleeve (Alias, et al., 2013).

NCHRP Project 12-102 37 Threaded Sleeves: Threaded sleeves can provide essentially the full tension strength of bars up to grade 75. The deformation capacity, i.e. strain or ductility ratio, is high, but depends on the way in which it is defined (Rowell, et al., 2010). Headed Bars with Mating Sleeves: Design of bar coupler connections are meant to emulate cast-in-place connections and it is helpful to the designer to know how they compare to cast-in-place connections. Tests have shown the following: the average performance of headed bars with mating sleeves relative to an A615 Grade 60 control bar achieved 90% of ultimate strength (129.6 ksi), 50% of maximum strain (12.2%), and 49% of ductility ratio (30.2) (Rowell, et al., 2010). The numbers provided in parentheses are the control case. The designer should also be aware that the data for structural performance for headed bars with mating sleeves is inconsistent. However, some research shows that the behavior of a connection featuring headed couplers and A706 grade 60 bars without a pedestal performed essentially the same as a cast-in-place connection (Haber, et al., 2013). While others, using A615 grade 60 bars and high strain rates, found that the coupler achieved 90% of the bar’s strength and about 50% of its deformation capacity (Rowell, et al., 2010). Additionally, no improvement to the ductility or reduction of the moment demand was observed with the presence of pedestals (Haber, et al., 2013). This finding is not surprising, however, because the coupler without a pedestal already performed as well as a cast-in-place column. Finally, initial design parameters and detailing for precast columns with mechanical splices can be reasonably done with moment-curvature and lumped-plasticity model (Haber, et al., 2013). External Clamping Shear Screws: Another design consideration includes the test results for externally clamping shear screws. In tests using A615 Grade 60 bars, the coupler achieved 76% of the bar’s tension strength and 15 to 20% of its deformation capacity (Rowell, et al., 2010). The deformation capacity is significantly affected by the stress concentrations in the bar caused by the screws. While not directly related to seismic loading, other research has shown that the shear screw couplers are able to develop reinforcement beyond yield strength and to maintain ductility under blast loading. More tests are needed to determine if full ultimate strength is developed under blast loading (Holland, et al., 2010). 4.3.1.3 Construction Considerations Couplers almost invariably lead to tight construction tolerances, which is an important aspect of ABC. In addition, the coupler has a diameter that is larger than that of the bar, so the column transverse reinforcement must change size to accommodate that difference. Cover requirements are likely to prove critical at the coupler, so the cover to the bars themselves is likely to be larger than normal. In some cases, debonding is critical to the performance of the connection. Debonding of the reinforcing bars may be achieved by wrapping them with tape or by encasing them in a plastic sleeve. Another construction consideration is the time to construct the connection. The time needed to complete a multi-coupler connection is related to the number of couplers used. If that number is large, the time advantage of using a precast element over cast-in-place construction will diminish. Grouted Sleeves:

NCHRP Project 12-102 38 Construction tolerances are tight, though less stringent than for other bar coupler types, e.g threaded sleeves. A grouted sleeve connection between two precast elements can be completed without the need for a final concrete pour around the connection. This is beneficial for both time savings and maintaining good surface finishes. Threaded Sleeves: Alignment of the two parts of these connectors must be extremely accurate if the coupler is used to join two precast components and the threads are to match up. However, if the threaded part is used in a combination coupler, and the threaded bar-to-sleeve part is embedded in the precast member, the threaded connection is made before the coupler is cast into the member and the alignment constraint is no longer relevant. Two precast elements cannot be joined using a “double-ended” threaded sleeve connection, because one component has to be turned. A final construction consideration for threaded sleeves is that machined threads are sensitive to accidental damage on site, and must be adequately protected. Headed Bars with Mating Sleeves: A connection made with headed bars and mating sleeves requires a final concrete pour around the connection hardware, which cannot be completely embedded within the precast members. Additionally, the bar ends must be aligned accurately for the coupler components to be connected. When used in Caltrans’ Headed Reinforcing Corporation (HRC) Connections, replacing fuse bars will require lining up a fuse with threads after large deformations have occurred during a seismic event. This may prove to be a difficult repair to make. External Clamping Shear Screws: Alignment of the two parts of these connectors must be extremely accurate if the coupler is used to join two precast components in order to fit the bars within the coupler. The coupler must also be exposed during construction in order to install the screws. This will require the use of a grouted or concreted recess. 4.3.1.4 Knowledge Gaps 1. Bar Couplers: More information about location, orientation of the couplers and debonding of the reinforcement is desirable. While recent tests have significantly added to the knowledge base, sufficient information to create a detailed design methodology is still not available. The capacity of existing couplers with higher strength rebar (80 – 100 ksi) should also be investigated. 2. Threaded Sleeves: Topics for investigation related to threaded sleeves include: • Inelastic cyclic performance – drift capacity • Influence of the coupler on bar strain distribution • Influence of the coupler location and orientation on inelastic performance and ease of construction • Influence on deformation capacity of local debonding of reinforcing bar near coupler 3. Headed Bars with Mating Sleeves: There are no identified knowledge gaps specific to this coupler type at this time. 4. Externally Clamping Shear Screws: Topics for investigation related to externally clamping shear screws include: • Inelastic cyclic performance – drift capacity • Influence of coupler on bar strain distribution • Influence of coupler location and orientation on inelastic performance and ease of construction

NCHRP Project 12-102 39 • Influence of local debonding of reinforcing near coupler 4.3.2 Grouted Duct Connections Grouted duct connections are accomplished by grouting a single bar that projects from one member into a duct in another member. Force transfer occurs from the bar into the grout, from the grout into the duct, and then from the duct into the surrounding concrete. Transfer can also be accomplished through lap splicing ducts with adjacent rebar in the receiving member. Grouted ducts can be used in column-cap beam, column-footing, and pile-pile cap connections. Grouted ducts are also typically used in capacity-protected manner, such as the connection between segments of a column. Figure 4.3.2-1 Grouted Duct Connection 4.3.2.1 Design Considerations There are many design considerations in the use of grouted duct connections. Below are considerations the designer should be aware of. Grouted Duct Location and Geometry: Grouted ducts may be used either in the capacity-protected region (Pang, et al., 2008) or in the plastic hinge zone (Belleri, et al., 2012). If they are used in the plastic hinge zone, the ducts inhibit bar buckling and the ductility capacity may be higher than that of a comparable cast-in-place system. Local debonding of the bar adjacent to the duct also influences the ductility capacity. Steuck, Eberhard, and Stanton 2009 found that a bar can be developed in 6.5 bar diameters under static loading and 10 bar diameters under cyclic loading, using corrugated steel ducts. Thus, the connection can be short enough to fit easily within the depth of a typical cap beam, even if large bars (up to #18) are used.

NCHRP Project 12-102 40 Ultimate duct bond strength is decreased when the duct size is increased, though a minor effect is seen in initial bond behavior (Tazarv., et al., 2014). This occurs because a cone of grout adheres to the loaded end of the bar and separates from the reminder of the grout, thereby shortening the length available for bond. Grout Characteristics: Research has shown that the presence of fibers in the grout does not improve monotonic pull out of bars in ducts and decreases the strength of the grout (Steuck, et al., 2009). The effective friction coefficient of at least 0.65 is justifiable under cyclic loading for use in interface and joint shear calculation between concrete and grout. Extending longitudinal bars into the diaphragm enlarges the joint area and thus reduce the stress (Belleri, et al., 2012). Duct Materials: The designer should be aware that the use of plastic duct material provides inferior bond strength relative to steel and requires 30% additional embedment length. This conclusion was determined using #11 bars in a constant duct size, and therefore, did not account for effects of bar size and duct size (Brenes, et al., 2006). Corrugated galvanized steel is typically used for the duct material, smooth is not recommended due to inferior bond strength. It has been shown that with conventional concrete, the critical bond surface is the bar-grout surface (7Steuck, et al., 2009). Debonding: Debonding of reinforcing bars within the plastic hinge region allows for approximately uniform strain distribution in the debonded region and allows the strain profile of the column to develop similarly to a CIP column (Pang, et al., 2008). Cyclic loading causes some additional debonding to occur beyond that intentionally applied by taping or sleeving the bar. The amount of additional debonding depends on the nature of the cyclic loading, but is likely to be at least one bar diameter at each end of the debonded region (Raynor, et al., 2002). Connection Strength: The strength of a grouted duct connection decreases when a bar is placed eccentrically within a duct (Brenes, et al., 2006). Also, individual grouted duct strength decreases when multiple grouted ducts are loaded in tension together without adequate spacing due to a group pullout effect (Brenes, et al., 2006). Grouted duct response to pullout testing shows insensitivity to the presence of epoxy coating, embedment depth of the connector, clear spacing between ducts, and the presence of transverse reinforcement (Brenes, et al., 2006).Transverse reinforcement may play a greater role if the cover over the duct is small and splitting is a potential mode of failure. 4.3.2.2 Construction Considerations Grouting Methods: The duct may be grouted either by pouring from above, or by pressure grouting from below. In both cases, the fluidity of the grout and the experience of the contractor are important in filling the ducts successfully. Depending on the regulating body, the construction may require a pressurized grouting procedure as described above, and there may be difficulty in finding a contractor with experience in performing pressurized grouting (Alderson 2005).

NCHRP Project 12-102 41 Layout and Geometry: During construction, the bars in one member and the ducts in the other must be properly aligned. Use of the same template or two matching templates on connecting members during fabrication helps achieve this. (Marsh, et al., 2011). The use of larger ducts simplifies fit-up, but takes more space in the reinforcing cage. It also slightly reduces the pullout strength. If the coupler is made within a column, a step must be made for spiral reinforcement. The use of larger ducts creates a need for a bigger step. Care must be taken in design of proper details and in construction to avoid the occurrence of failure at the discontinuity between spiral diameters. Care must be taken in determining which connecting member should house the grouted duct. For example, there is less construction risk in a column-footing connection if the duct is housed in the column and transportation is simpler in a column-cap beam connection if the duct is housed in the cap beam. Another consideration is that using a smaller quantity of large bars, i.e. #18’s, instead of a larger quantity of small bars, provides greater construction tolerances and reduces congestion. Also, bars should be placed as near to the center of the duct as possible per current research (Pang, et al., 2008). Serviceability: Another notable aspect of grouted duct connections is that in a column-footing connection with the ducts in the column, the rigidity of the ducts causes the damage to be concentrated in the grout pad beneath the column. Repair is then simpler than for a cast-in-place or socket system. (Belleri, et al., 2012). 4.3.2.3 Knowledge Gaps 1. Effect of duct size on anchorage length: The ratio of the bar diameter to the duct diameter could potentially have an effect on development length. 2. Influence of duct location on cyclic performance: The AASHTO prohibition for mechanical splices in one plane should be studied including the potential to place the splices in the hinge region. 3. Implication of lap splicing column bars to connection bars: The transfer of force from the bars within the duct to the duct and to the surrounding bars could be studied further. 4. Impact of additional longitudinal bars on plastic hinge region cyclic performance: The impact of these connections on plastic hinge length and plastic hinge region criteria could be studied. 5. Impact of transverse bars on cyclic performance of grouted ducts: Transverse bars around the ducts could have an impact on the overall connection performance, including how the confinement steel steps from around the coupler to being around the longitudinal reinforcement. 6. Influence of duct material, off-center bar, group pullout effects, bedding layer reinforcement in conjunction with bar and duct size: All of these parameters could have an impact on the performance of these connections, therefore they could be studied. 4.3.3 Pocket Connections Pocket connections are made by grouting extended reinforcement, sometimes referred to as “downstands,” from one connecting member into a precast pocket within a receiving member. This is illustrated in Figure 4.3.3-1. The transfer of forces between the two members occurs at the pocket perimeter. The column-cap beam connection is the most common application of the pocket connection. It is possible to use pocket connections in column-footing and pile-pile cap connections as well. However,

NCHRP Project 12-102 42 difficulties arise in the grouting of the pocket when a column-footing connection is used, as access to the pocket is limited. Figure 4.3.3-1 Pocket Connection 4.3.3.1 Design Considerations Connection Reinforcing: One design condition is that the use of a steel duct removes the need for much of the vertical shear steel in the joint region. Where it is impossible to use a steel duct, the lack of confinement results in more significant joint shear cracking. Supplementary hoops should be provided where there is a lack of a duct to provide adequate confinement of the joint region. Typically, this is above and below the corrugated confining duct within cap beams, where longitudinal steel must be allowed to pass through (Restrepo, et al., 2011). Another design consideration is that additional longitudinal reinforcement is necessary, as strains in the cap beam bottom longitudinal bars experience greater strains than in cap beams formed with cast-in-place connections (Matsumoto, 2009). Connection Behavior: When designed to reach full ductility with the 2009 LRFD, SGS columns with pocket connections exhibit behavior closely emulating cast-in-place concrete. It is also noteworthy that a cap pocket designed for limited ductility satisfies the requirements set forth for Seismic Design Category B using 2006 specifications (Restrepo, et al., 2011). 4.3.3.2 Construction Considerations Layout and Geometry: One construction consideration is that a grout or concrete bedding joint can be used to provide adjustability. In addition, a steel duct between the receiving members top and bottom longitudinal reinforcing may be used for the dual purpose of providing formwork for the pocket and confinement to the joint. It also frees the joint region of congestion, as less vertical shear reinforcement is needed. Where the steel duct is impractical because of longitudinal reinforcing, formwork for the pocket region in the receiving member may be accomplished with a cardboard concrete form tube. The cardboard concrete form tube may be perforated to allow the longitudinal bars to pass through. With consideration to constructability, headed bars can be used to reduce the development length of bars, and remove 90 and 180 degree bar bends, which results in a less congested pocket region. In

NCHRP Project 12-102 43 addition, self-consolidating concrete may be pumped into heavily congested pocket regions (Karapiperis, et al., 2010). Connection Materials: During construction either shrinkage compensating grout should be used in the pocket region to limit shrinkage cracks or epoxy-coated reinforcement should be used in the joint region to prevent corrosion when shrinkage cracks do form (Mastumoto, 2009). Consideration should be taken for the reduced bond and subsequent longer development lengths required when epoxy is used. Finally, the use of high early strength concrete or grout within the pocket region reduces the curing time of the joint region and is a consideration for accelerated construction (Marsh, et al., 2011). 4.3.3.3 Knowledge Gaps In the area of pocket connection research, the following knowledge gaps exist: • Joint behavior and joint performance limit states. • Pipe effects and crack patterns (Restrepo, et al., 2011). • Optimal duct thickness in pocket region. 4.3.4 Socket Connections In a socket connection, the embedded member, typically a column or a pile, is placed in an opening in the receiving member, typically a footing, pile cap, cap beam, or drilled shaft. Fig 4.3.4-1 illustrates the principle configuration of a socket connection. The embedded member may be constructed from concrete, steel or a combination of both, but is always fabricated before the connection is made. The receiving member may be precast with an opening, in which case the connection is completed by grouting the annular space (Fig 4.3.4-1a). Alternately, it may be cast in place around the embedded member (Fig 4.3.4-1b). In all cases, the primary source of force transfer is by shear across the surfaces of the socket. Figure 4.3.4-1 Socket Connection Examples of socket connections include column-footing connections, column-to-beam connections, column-to-drilled shaft, and pile-to-pile cap connections. (a) (b)

NCHRP Project 12-102 44 4.3.4.1 Types of Socket Connections Socket Connection with CIP Receiving Element: This type of socket connection is accomplished by casting in place the receiving element around a prefabricated embedded element. An opening is made in the reinforcing cage of the receiving element to accommodate the embedded element. The embedded element is then placed within the opening and the receiving element is cast in place around the element. This connection is sometimes referred to as a “wet socket”. This system has several advantages as well as drawbacks. These include: • The system is better suited to footings than cap beams, because a cast-in-place cap beam fails to take advantage of the time savings inherent in a precast cap beam. • In a footing, the cast-in-place option is likely to be preferable, because precasting is likely to save little time, and may lead to higher crane costs if the footing is large. • The cast-in-place receiving element allows for very large placement tolerances. • The lack of steel crossing the interface between the two elements simplifies handling and transportation of the embedded element. • The embedded element may be constructed using precast concrete, a concrete-filled tube, or some other system. Connection details vary with the type of embedded element. Concrete-filled tubes do not inherently accelerate construction, but as part of a properly planned ABC project, they may provide construction convenience that is consistent with ABC goals. • The embedded element may be fabricated in a plant or on site. Socket Connection with a Precast Receiving Element This type of socket connection is accomplished by placing an embedded element into a pocket formed in the precast receiving element and grouting around the space between them. 4.3.4.2 Design Considerations General Consideration for CIP Receiving Elements: The system should typically be designed so that the inelastic action occurs in the embedded member, and not in the receiving member. The thickness of the receiving member may be controlled by this requirement. The cyclic response of embedded members using socket connections emulates cast-in-place behavior closely, and thus the embedded members can be designed using current cast-in-place design methods (Haraldsson, et al., 2013). The connection must be designed so the axial load from the embedded member can be transferred to the receiving member, or vice versa. For precast concrete columns, this may be achieved by suitable roughening of the surface. Corrugations using wooden strips have proven successful (Haraldsson, et al., 2013). For other systems, discrete shear connectors may be needed. Examples of shear connectors include rings or flanges welded to a steel Concrete-Filled Tubes (CFT) column or pile. Joint shear within the connection region must be addressed. The depth of embedment affects joint shear demand, and capacity is affected by the method of anchorage of the column reinforcement. The use of straight, headed bars has been shown to be greatly superior to the use of bars bent outwards into the receiving member. Research found the AASHTO specified minimum transverse steel requirements in a footing to be unnecessary when headed bars were used in a column (Haraldsson, et al., 2013).

NCHRP Project 12-102 45 The bottom mat of receiving member steel does not have to pass under/over the column. The bars that would normally be directly beneath/above the column may be moved aside and placed adjacent to it (Haraldsson, et al., 2013). Considerations for CFTs: The embedment length of the tube into the footing has the greatest effect on performance. Many researchers have tried embedment depths less than the column diameter, and have provided recommendations for the critical depth needed to give satisfactory performance. The critical depth depends on several characteristics. (Moon, et al., 2013), Stephens, et al., 2012), Lai 2010), and Osanai, et al., (1996). CFSTs (Concrete-Filled Steel Tubes) can lead to ductile response when the steel yields. Unlike CFSTs, CFFTs (Concrete-Filled Fiber Tubes) do not exhibit a yielding phenomenon, but the tube can partially slip out of the footing under cyclic loading, which provides behavior somewhat similar to yielding. However, when this occurs, the strength drops significantly below the peak value. Considerations for Precast Receiving Members: Considerations for precast receiving members are generally similar to socket connections for cast-in- place receiving members with a few additional considerations. If the socket diameter is significantly larger than that of the embedded column, for example in a footing application, the reinforcement in the receiving member will necessarily be placed further from the column face. This may detract from the system performance. Additionally, if the connection is used between a column and cap beam, the size of the opening may be limited by the width of the cap beam. 4.3.4.3 Construction Considerations General Considerations for CIP Receiving Members: If the connection is used at a footing, the weight of the column must be supported while the footing concrete is cast. A “rat slab” may be needed under the column if the ground is unable to support the local contact pressure caused by the leveling devices. Leveling may be done using shims or other leveling devices. The embedded portion must also be braced. The footing reinforcement may be installed before or after the column is placed. Pre-fabricating the footing cage is likely to save time on site. Considerations for Precast Receiving Members: During construction of socketed connections with precast receiving members, the corrugated tube or other device used to form the opening in the receiving member must be carefully placed to ensure that the column can be accurately positioned. This is particularly important in footing applications. Two site operation, footing casting and subsequent grouting, are required. Comparatively, socket connections with cast-in-place receiving members only require one operation. The designer also needs to be aware that grout dams are required for grouting column to cap beam connections. For some of these connections, shoring may be required depending on the size and weight of the members and the detailing of the cap beam. In one testing program, the researchers used a connection in which the column diameter was reduced in the region of embedment in the cap beam. This detail both reduced the size of opening needed and provided a “shoulder” on which the cap beam could be set during

NCHRP Project 12-102 46 erection. This configuration had the advantage of eliminating temporary support from shoring or column clamps. 4.3.4.4 Knowledge Gaps 1. Alternative methods of roughening the surface of a precast column should be studied. The use of wood strips to form corrugations is effective in transferring shear stress, but is labor-intensive. Generation of a rough, exposed aggregate finish using retarder has shown promise. 2. The diameter-to-thickness ratio for CFSTs is generally no greater than 100, which is equivalent to a RC column with 4% steel, so the steel cost is relatively high. However, thinner tubes risk premature buckling. More work is needed to clarify the relationships between tube diameter-to-thickness ratio, steel strength and transfer of forces. 3. More work is needed to achieve a better understanding of the cyclic behavior of CFFT connections and FRP tubes. 4.3.5 Integral Connections Integral connections are used to provide moment continuity across a joint interface and improve behavior of the bridge when subjected to longitudinal seismic forces. The most common application of integral connections is between the cap beam, diaphragm, and girders. These integral connections are typically required in areas of high seismicity unless a special analysis is performed. They offer the advantage of reducing the moment at the base, and therefore permit the use of smaller and more economical foundations. Because the ability to inspect integral connections is limited and repair is difficult, these connections are typically designed as capacity protected with any energy dissipation designed to occur outside of the connection region. a) Non-integral (pinned) bent cap b) Integral (fixed) bent cap Figure 4.3.5-1 System Response to Longitudinal Excitation

NCHRP Project 12-102 47 Figure 4.3.5-2 Integral Precast Lower-Stage Cap with Precast Girders (NCHRP 12-74) While integral connections are not themselves an ABC connection style, this section will serve as a discussion of utilizing ABC techniques to achieve integral connections. An example of such techniques include utilizing stay-in-place formwork in the integral region and monolithically casting the connecting members together. Ideally, connecting bridge members is achieved in one of three ways: • Bolted directly to the formwork for steel members • Inserted entirely into formwork and cast monolithically for concrete members • Connected by inserting individual connection bars The last connection listed requires shear lugs or studs on both sides of the formwork for secure shear transfer. Positive and negative moment continuity of the girders may be achieved by extending longitudinal post-tensioning or mild reinforcement dowels into the integral joint region. Bar couplers, grouted ducts, socket connections or pocket connections may be used to connect the columns into the cap beam. 4.3.5.1 Design Considerations Connection Behavior: When detailed appropriately, integral cap beam connections featuring concrete column and steel girder composite decking remain elastic through seismic loading, with a plastic hinge forming within the column and outside of the connection region. The large confinement provided by the steel formwork may allow the reduction of confining reinforcement within the connections (Wassef, 2004). Connection Reinforcement Detailing: An integral connection featuring longitudinally spliced bulb tees and bathtub girders allows for essentially elastic performance within the bent cap and superstructure. In this connection, column reinforcement should be extended into the bent cap as far as possible to achieve favorable bond-strut angles. Additionally, bent cap pre-stressing reduces congestion and improves seismic performance (Holombo, et al., 2000). Both the as-built and newly developed girder to cap beam connection tested at Iowa State University performed as a fixed connection. Prior to this test, the as-built connection was treated as a pinned

NCHRP Project 12-102 48 connection. This also means that the top of columns require confinement, as a plastic hinge will develop in this region (Khaleghi 2013). 4.3.5.2 Construction Considerations Stay-in-place Formwork: Stay-in-place formwork may be made of steel, precast concrete, or composite materials. The stay-in- place formwork will require shoring unless it is designed to support itself and connecting bridge members. The formwork must contain dowels or lugs to ensure composite action between the formwork and the concrete cast within it. When steel girders are used, stay-in-place formwork made of steel may be used. They are to be constructed by a steel fabricator, and in most cases certified for bridge work. Ideally, reinforcement would be installed into the formwork prior to on-site installation. Construction tolerances must be accounted for in formwork fabrication. When steel girders are used and connected via bolts to the stay- in-place formwork, tolerances are tight. Connection Materials: Concrete must have a high enough slump or be self-consolidating so it flows enough to get into the forms. Adequate consolidation may be difficult to ensure around congested steel or beneath steel plates. Oversized holes must be sealed before pumping the concrete. 4.3.5.3 Knowledge Gaps 1. Specific anchorage detailing, for both bars and strands, in the joint regions to provide adequate positive moment continuity. 2. Joint shear and anchorage detailing considerations for two-stage cap beams. 3. Torsional stiffness and strength of two-stage cap beams. 4.3.6 Re-Centering Connections The project panel has decided to not include re-centering connections in the ABC Guide Specifications. This decision was made after the completion of the literature search. The information contained in this section is for future reference should these technologies gain common acceptance with the AASHTO SCOBS Technical Committee T-3 – Seismic Design. This section is included to capture advancements that are currently being investigated, researched, and implemented in demonstration projects as the upcoming solution to performance enhancement of bridges in high seismic regions. These advancements are, for the most part, not developed sufficiently to be adopted into the AASHTO Guide Specification for LRFD Seismic Bridge Design. Currently there are several NCHRP projects that are directed to the advancement of these technologies, with more to come. The primary purpose of these advancements is to improve ductility, reduce damage, reduce permanent set (displacement), and dissipate energy. Some of these advancements may be incorporated into a seismic restraining system, such as a prefabricated column, that can be termed accelerated. Inclusion of one of these advancements into the ABC design and/or construction specification should not be a means to bypass the acceptance of the AASHTO SCOBS T-3 technical committee for implementation. Most of these systems require analysis that is not currently defined in the current specifications. Once these design principles and analysis techniques are incorporated into the AASHTO Guide Specification for LRFD Seismic Bridge Design, then the advancements should be considered for ABC. The importance of

NCHRP Project 12-102 49 identifying this work for the report(s) of this NCHRP project is essential, as this is the next generation for bridge seismic engineering. 4.3.6.1 Technology Description Re-centering systems constitute a class of structural systems that return to their unstressed geometry when unloaded, even after undergoing inelastic deformations. In bridges, re-centering methods are used to bring the columns back to vertical after large seismic displacements. These systems are primarily intended to improve seismic performance, and they have no inherent ABC properties. However, many of them can be incorporated into an ABC program in which case it is possible to achieve simultaneously both faster construction and better seismic performance. The two most common methods used to achieve re-centering are: • Systems that use linear elastic materials in a nonlinear geometric context. Examples are: columns that rock as rigid bodies and are brought back to vertical by unbonded pre- or post-tensioned reinforcement. • Systems that use materials that are nonlinear elastic. Shape Memory Alloy (SMA) reinforcing bars are an example. In both cases the load-deflection curve has the shape idealized in Fig 4.3.6.1-1(a). Its most important feature is that it passes through the origin so that, when the external force is zero, the displacement is also zero. This feature differs significantly from the goals of traditional seismic systems, in which fat hysteresis loops were preferred because their high energy dissipation provided damping that limited the peak displacements during the earthquake. However, an inevitable consequence of such a loop is that when the force is zero, the displacement is not, so such systems are prone to residual “plastic” displacements after the ground motion stops. Figure 4.3.6.1-1 Load-Deflection Curves for Re-Centering Systems In a re-centering system based on unbonded post-tensioning in a single-column bent, the column rocks as a rigid body and a crack opens at the interface with the footing. As the crack opens, the stiffness drops significantly, as shown in Fig 4.3.6.1-1(b). When the load is removed, the column unloads back along its

NCHRP Project 12-102 50 loading path. Such an arrangement provides no damping, so bonded reinforcing bars may be added to it. They yield alternately in tension and compression and display the load-deflection response shown in Fig. 4.3.6.1-1(c). An optimal system has as much energy dissipation as possible without jeopardizing the goal of no residual displacements. The resulting energy dissipation is less than that dissipated in a conventional system, so the peak displacement may be slightly higher, although this depends on other characteristics, such as the structure’s natural period. No bridges have yet been built using these “hybrid” systems that contain both types of reinforcement, but tests, including shaking table tests, have been conducted and have proven successful in providing both complete re-centering and very low damage. About a dozen buildings have been built using the same principles. SMAs have stress-strain curves that resemble the load-displacement curves shown in Fig 4.3.6.1-1(a). When they are used to reinforce a column, its load-deflection characteristics also resemble those shown in Fig 4.3.6.1-1(a). The first SMA bars were based on nickel-titanium alloys, and have two major drawbacks: the material is very expensive and very hard to process. The high cost leads to designs that minimize the use of the SMA and include only short “fuse” bars in the plastic hinge region. That arrangement implies coupling to conventional reinforcing bars, which in turn implies machining or other processing, which is difficult and expensive. Alternative alloys, based on copper, are less expensive and are easier to process, but their stress-strain properties are not as useful as those of the nickel-titanium bars. SMA-based systems have not yet been deployed on a commercial basis. In this section, re-centering systems have been broken up into three categories: Pre-tensioned pre- stressing strands, post-tensioned pre-stressing strands, and methods that do not involve any pre-stressing. 4.3.6.2 Pre-Tensioned Systems Currently, there is only one pre-tensioned system that has been developed. The column features a socket into CIP connection at the footing, with a steel confinement jacket at the base and headed bars, and a partial socket into precast member connection combined with grouted duct connections at the cap beam. Debonded Pre-tensioned pre-stressing strand is used concentrically throughout the length of the column for rocking, with bonding allowed at the ends for anchorage. Mild reinforcement extends over the whole length of the column and is bonded everywhere except locally at the rocking interface, where it may be debonded locally to avoid strain concentrations. Additional, fully bonded rebar is placed in the body of the column between the two confinement tubes to reinforce the concrete near the interface and to inhibit damage there when the column rocks. Design Considerations: • The footing connection was the “wet socket” tested by University of Washington and used in the bridge over I-5 described in (Khaleghi, et al., 2012). The top connection was a combination of a socket connection into the precast cap beam, achieved by reducing the column diameter locally within the depth of the cap beam, and grouted ducts for perimeter rebar. • To handle interface stress concentration at the footing, using a steel confinement tube at the plastic hinge region and debonding longitudinal reinforcing across the footing surface provides the best performance compared to using conventional concrete only or a Hybrid FRC shell in the region (Stanton, et al., 2014). • Strand yielded at 3% drift. Rebar fractured at 6% drift. Both of these were as expected and could be changed by design. No concrete or footing damage was observed up to 10% drift. Peak drift was measured at 13%, with residual drift less than 0.2% and essentially no damage to the concrete (Stanton, et al., 2014).

NCHRP Project 12-102 51 • The transverse reinforcement required in the footing by AASHTO is unnecessary if headed bars are used at the end of the column within the socket (Stanton, et al., 2014). • Diagonal shear friction steel used in the tests is not needed within the footing (Stanton, et al., 2014). • The vertical capacity of the wet socket connection was confirmed by loading the column to 3.5 times its factored vertical load without a crack in the footing (Stanton, et al., 2014). During that loading, the column had no support directly beneath it. Construction Considerations: • Unbonding of strands can be achieved by use of a plastic sleeve along the debonded length. • The specimens were easy to construct in the lab. Knowledge Gaps: 1. Bond and anchorage of the strands under cyclic loading. 4.3.6.3 Post-Tensioned Systems Post-tensioned re-centering systems behave in the same way as their pre-tensioned counterparts, with the key difference being that the pre-stressing takes place on site. At this time there are many different manifestations of post-tensioned re-centering columns. The typical features include a concentric post- tensioning tendon through the middle of either a continuously precast column or through jointed column segments. Bar tendons offer construction advantages over strands in that the connections are easier to make. However, they have only about half the strain capacity of strand so a supplementary source of flexibility, such as a polymer washer, may be needed in non-slender columns. One variation of the segmental technique allows slipping at the segmental joints to allow energy dissipation. Another variant features a dual steel shell, which provide confinement, shear reinforcement, and allows for a reduction in weight, and therefore longer precast members. Design Considerations: • On-site construction requires more steps than with the pre-tensioned system. • Corrosion of the PT anchorages is a potential concern. • Higher strain capacity is needed for stocky members. • Columns reinforced with both deformed bar and pre-stressing tendons provide less hysteretic damping than a comparable fully bonded system, and thus a larger maximum displacement will be experienced. Also, due to higher compressive stresses in the column due to the pre-stressing, premature crushing may occur in the columns at lower drift levels (Kwan, et al., 2003). • Conventional strand wedge anchors create stress concentrations that reduce PT tendon’s cyclic strain capacity below that of the bare tendon (Walsh, et al., 2012). • Stainless steel bars may be used for energy dissipation. They feature more ductility and corrosion resistance (Restrepo, et al., 2011). • An ultimate deformation capacity of 6% drift ratio was achieved by all three unbonded PT columns tested as part of NCHRP 681. Damage was observed in the grouting bed at 2% drift and may be mitigated with the use of fiber-reinforced grout (Restrepo, et al., 2011). • In a large-scale test of square hollow box segmental columns, a higher ratio of energy dissipating steel to post-tensioning steel increased system energy dissipation and residual drift (Ou, et al., 2010). This is consistent with theoretical behavior. • Modifying the wedge geometry is proposed to improve anchorage strength and ductility when used with certain anchorages (Abramson, 2013).

NCHRP Project 12-102 52 • An FRP tube adds confinement, which reduces concrete damage and increases strength. However, if the tube slips relative to the concrete a load drop will occur. • The increase in displacement ductility demand was greatest for structures with periods less than 1 second and low to medium yield strength ratios between 0.1 and 0.5. Increasing energy dissipation capacity counteracts this effect (Chou, et al., 2008). • A slip-dominant segmental column demonstrated capacity to control applied seismic forces, provide energy dissipation, and provide moderate re-centering properties. Joint sliding did not cause any damage beyond spalling at joint edges and residual sliding can be restored by mechanical means (Sideris, et al., 2014) (2). • Measured yield and maximum strains were greater for 0.5 in diameter strands than 0.6 in diameter ones (Walsh, et al., 2010). • Seating loss equations for unbonded PT strands under inelastic cyclic loading are presented in (Bruce, et al., 2014). • The limit on peak PT strain varies by chuck type used and initial prestress levels, with older chucks displaying limits of 2% and 1% strain for 20% and 80% of guaranteed ultimate tensile strength respectively. For a newer chuck type, limiting strains were 3.5% and 2.9% respectively for the same loading conditions. • The unbonded portion of the energy dissipating reinforcement can be placed in a tube to prevent buckling under compression. (Chou, et al., 2008). • Using CFFT with unbonded PT allows for 9.2% lateral drift without significant damage or residual displacement. Including energy dissipaters allows the same 9.2%, but with some damage. These compare to a reference CIP failure at 6.9% and substantial damage and residual displacement (El Gawady, et al., 2010). • An unbonded PT column featuring hybrid FRC and armored headed rebars in the plastic hinge region accumulated only 0.4% residual drift ratio with minimal damage from the same seven ground motions that caused 6.8% residual drift ratio with substantial damage in a reference CIP column. The peak drift ratio was 8.0% for the unbonded PT column compared to 10.8% for the reference (Trono, et al., 2014). • Unbonding the mild bar in the plastic hinge region results in a shorter plastic hinge region, a slightly larger maximum displacement, and a slightly larger residual displacement (Jeong, et al., 2008). • A confining steel jacket sheathing and unbonded mild reinforcement prevented any observable damage from occurring from any test. The residual drift index was less than 0.1% at design level seismic loading, and 0.6% at the maximum seismic load applied (Jeong, et al., 2008). • The controlled damage re-centering design featuring a member socket connection experienced concrete spalling under cyclic loading when no cover confinement is provided. After repair of this concrete and the addition of GFRP cover, a significant increase in column performance occurred, with no spalling until 3.25% drift (White, 2014). • When external dissipaters were used in a repaired controlled damage member socket member, no damage was seen in the dissipaters themselves up to drifts of 7.8%, though there was some slipping of the mounting and pull out of the dissipaters partly due to prior damage before the repair (White, 2014). Construction Considerations: • Unbonding of strands can be achieved by use of a plastic sleeve along the debonded length. • Corrosion protection for the strand must be made. Strands are more susceptible to corrosion than bars because of their higher surface-to-volume ratio. Anchorages pose the greatest threat. • If a hollow box pier column is used, reparability is improved if access to the post-tensioning ducts is provided in the interior.

NCHRP Project 12-102 53 • More precautions must be taken around prestressed strands than mild reinforcement. • A U-shaped tendon with two top anchorages provides better inspectability for anchorages than other PT anchorages. This anchorage is accomplished by casting a U-shaped duct in the footing that lines up with two ducts in the column. PT strand is then fed through the duct and back up the other side on site and the bottom U-shaped portion of the duct is grouted. • Accomplishing an anchorage beneath the footing would be very difficult due to limited access in most applications. Corrosion requirements would also more extensive if this were to be done. • A tendon or tendon anchorage may be cast straight into the footing. If column segments are used, they can be threaded over the tendon one by one. If a whole column piece is used, then it is possible to feed the strand through before lifting and placing the column. • One possible corrosion protection method would be to grout the tendon into an interior ducts, which is free to slide inside an exterior duct that is bonded to the surrounding concrete. • Epoxy-coated strand offers an advantage for corrosion protection, but often contains grit meant to improve bond, which is not desirable in this application as friction between the strand and plastic will be increased during stressing. The coating also risks being broken if the tendon follows a U- shaped profile, with two anchorages at the top of the column. Knowledge Gaps: 1. Development of a cost-effective solution that allows for adequate anchorage, corrosion protection, and ease of inspection and repair. 4.3.6.4 Non-Prestressed Systems In this section, coverage is provided for systems that do not fit into any of the previously covered sections. Included in this section are the SPER System, originated in Japan, and systems using SMA bars. Design Considerations: SPER (Sumitumo Precast form for resisting Earthquake and for Rapid construction) System: • There are two versions of this stay-in-place formwork system, a 40 ft maximum method and a 160 ft maximum method. For the 40 ft method, cross ties are pre-installed into square stackable sections. As they are stacked, epoxy is used along the vertical reinforcement. For the 160 ft method, an inner and outer formwork are used. They are shipped in two pieces and coupled together before they are stacked. Additionally, the cross ties are installed on site (Barthes, et al., 2012). SMA Bar Systems: • The nonlinear elastic SMA bars return to their original shape after stress is removed or heat is applied. • External screw bar couplers can be used to splice SMA bars to steel bars and an adhesive method using epoxy is needed to splice SMA bars with FRP bars. If this is done, the FRP bars require 18 bar diameters for development lengths (Alam, et al., 2012). • As expected, tests show that using SMA bars reduces residual displacements and crack widths. Beams with SMA take a higher normalized ultimate load and strain (Abdulridha, et al., 2013). • A beam reinforced with SMA bars dissipated comparable energy to a CIP beam on initial loading, but only roughly half as much in the reverse cycle (Abdulridha, et al., 2013). Construction Considerations: SPER System: • Time savings relative to CIP is estimated to be about 50% for the 40 ft maximum system and 30% for the 160 ft maximum system (Barthes, et al., 2012). SMA Bar Systems:

NCHRP Project 12-102 54 • SMA bars are restrictively expensive in their current state. Knowledge Gaps: 1. SPER System: • Usage of post-tensioning with the SPER System. 2. SMA Bar Systems: • A cost-effective material/ method in order to accomplish the shape re-forming capabilities. 4.3.6.5 Major Tests Performed University of Washington (Stanton, et al., 2014). Seismic ABC Research at University of Washington Preliminary test results for a 25% scale prototype bridge test are presented. The prototype bridge featured a pre-tensioned re-centering column utilizing a socket into CIP at the footing and a mix of grouted ducts and a socket into a precast member at the cap beam. University of Canterbury (White, 2014). Controlled Damage Rocking Systems for ABC Two columns featuring controlled damage connection types, which aim to provide adequate performance in moderate to high seismic regions with less up-front cost. One column was circular and featured a member socket connection with post tension and cover confinement provided at the connection interface. Threaded inserts are also including for the installation of an external energy dissipater. The second column was square and used replaceable segments of longitudinal bar connected to threaded studs using threaded bar couplers. Conventional mild steel stirrups were used for confinement of the base region after the installation of the energy dissipating system on site. University of California, Berkeley (Trono, et al., 2014). Seismic Response of a Damage-Resistant Re- Centering Post-Tensioned-Hybrid Fiber-Reinforced Concrete (HYFRC) Bridge Column Two columns were tested under seismic loading on a shake table. One was a reference CIP column and the other was an unbonded PT rocking columns with headed bars and HYFRC in the vicinity of the rocking plane. University at Buffalo, State University of New York (Sideris, et al., 2014). Large-Scale Seismic Testing of a Hybrid Sliding-Rocking Post-tensioned Segmental Bridge System A two phase test of hybrid sliding-rocking (HSR) members was conducted. HSR joints can be designed to either slide (slip-dominant, SD) or rock (rocking-dominant, RD). A series of shake table tests were performed on a large-scale bridge specimen featuring a RD superstructure and two single-column SD piers quasi-static testing of just the substructure of the bridge specimen as well. University at Buffalo, State University of New York (Sideris, et al., 2014). Quasi-Static Cyclic Testing of a Large-Scale HSR Segmental Column with Slip-Dominant Joints Summarizes the major findings from the quasi-static test described in (Sideris, et al., 2014). University at Buffalo, State University of New York (Sideris, et al., 2014). Effects of Anchorage Hardware on the Cyclic Tensile Response of Unbonded Monostrands Cyclic tensile testing was performed on 53 in long strand samples of 270 ksi 0.5 in and 0.6 in diameter monostrand and associated anchorage systems. A numerical model was adapted from the experiment. Results differed as some samples fractured one at a time as opposed to all at once. Virginia Tech University (Bruce, et al., 2014). Behavior of Post-Tensioning Strand Systems Subjected to Inelastic Cyclic Loading

NCHRP Project 12-102 55 Seating losses, deformation capacity prior to initial wire fracture, additional deformation capacity after initial fracture, and overall load-deformation behavior of unbonded PT strands were explored through monotonic and cyclic testing. Two types of strands, two types of anchorages, and varying levels of pre- stressing were applied. Twenty-four 6 ft long PT strands with varying combinations of the above parameters were performed. An equation that provides a reduction in PT strand force due to seating losses and suggested strain limits to avoid strand fracture are provided. University of Minnesota (Abramson, 2013). Comprehensive Evaluation of Multistrand Post-Tensioning Anchorage Systems for Seismic Resilient Rocking Wall Structures Two anchorage types and four wedge types were tested in various combinations to observe the performance of PT anchors in an unbonded rocking system. National Chiao Tung University (Chou, et al., 2008). Design and Tests of Post-Tensioned Structural Systems for Seismic Resistance: from Segmental Bridge Columns to Dual-Core Self-Centering Braces Three unbonded post-tensioned precast concrete segmental bridge column specimens were tested, two with mild steel bars crossing the column at different column heights. The purpose was to explore the effects on position of anchorage on energy dissipation. University of California, Berkeley (Barthes, 2012). Design of Earthquake Resistant Bridges Using Rocking Columns Numerical models are used to explore the performance of rocking columns. University of Notre Dame (Walsh, et al., 2012). Effects of Loading Conditions on the Behavior of Unbonded Post-Tensioning Strand-Anchorage Systems Cyclic tests were performed on 0.5 in and 0.6 in diameter 270 ksi unbonded PT strand-anchorage systems while considering loading rate, eccentricity between strand ends, post-yield cyclic loading, and initial pre-stressing. New stress-strain relationships are presented based on the results. University of California, San Diego (Restrepo, et al., 2011). Development of a Precast Bent Cap System for Seismic Regions As part of the testing done under NCHRP 12-74, three re-centering columns were tested. The first of these columns had conventional re-centering detailing with conventional spiral confinement reinforcement and full length mild reinforcement. A second column was a concrete-filled pipe, which used the full length steel shell for confinement and only required mild reinforcement at the joint region for energy dissipation. The third re-centering column used outer and inner full length shells which acted as confinement and shear reinforcement. Mild reinforcement was only used at the joint for energy dissipation. This is similar to the second column, but was detailed with the inner shell to reduce the weight and make constructing precast columns of longer lengths feasible. Design specification were developed for the re-centering connections tested as well. Stanford University. (Lee, et al., 2011). Performance‐Based Earthquake Engineering Assessment of a Self‐Centering, Post‐Tensioned Concrete Bridge System Two analytical bridge models, identical except for the type of column used, were developed to compare the performance of a conventional bridge with an unbonded post-tensioned bridge under the established PEER PBEE framework. National Taiwan University (Ou, et al., 2010). Large-Scale Experimental Study of Precast Segmental Unbonded Post-tensioned Concrete Bridge Columns for Seismic Regions

NCHRP Project 12-102 56 Four large-scale specimens were designed and tested with lateral cyclic loading. Post-tensioning was run in the hollow core of the columns and energy dissipating bars were used in the plastic hinge length with an unbonded region. University of Washington (Stanton, 2010). Unbonded Pre-tensioned Hybrid Concept Two pre-tensioned columns, one with a column-footing connection and the other with a column-cap beam connection, were tested under self-reacting frame, providing axial and lateral loading. Washington State University (El Gawady, et al., 2010). Seismic Behavior of Self-Centering Precast Segmental Bridge Bents Five columns were tested under cyclic loading. One served as a CIP reference case and the other four were self-centering with varying details including external energy dissipaters and neoprene isolation. University at Buffalo, State University of New York (Roh, et al., 2009). Nonlinear Static Analysis of Structures With Rocking Columns A nonlinear analytical model is developed for rocking columns using equivalent flexibility approach considering rigid body rotations and flexural deformations. Three 1/3 scale rocking columns, with slenderness ratios of 6.85, were subjected to quasi-static testing. National Chiao Tung University (Chou, et al., 2008). Hysteretic Model Development and Seismic Response of Unbonded Post‐Tensioned Precast CFT Segmental Bridge Columns Two unbonded PT CFT segmental columns were tested under cyclic loading to develop an understanding of their hysteresis behavior. Efforts were then made to create an SDFS model that predicts inelastic stiffness reductions and cyclic response. National Taiwan University (Jeong, et al., 2008). Shaking Table Tests and Numerical Investigation of Self-Centering Reinforced Concrete Bridge Columns Two series of shaking table tests were performed. First, four cantilever columns with partial prestessing and differing details were tested under bidirectional earthquake bending. Second, one bridge specimen with a two-column bent was evaluated. Analytical investigations were then conducted based on the experiments. Finally, a series of parametric studies for self-centering columns were performed that showed that local unbonding and steel jackets are beneficial. University of Washington (Cohagen, et al., 2008). A Precast Concrete Bridge Bent Designed to Re- Center After an Earthquake Two 40% scale specimens featuring large bar grouted duct connections and a central unbonded post- tensioned strand were tested under pseudo static loading to observe residual displacements and re- centering properties. National Kaohsiung First University of Science and Technology (Cheng, 2007). Energy Dissipation in Rocking Bridge Piers Under Free Vibration Tests Four columns were constructed to investigate the dynamic characteristic of rocking bridge columns subjected to quick release loads. Research parameters included rocking interface material, area of anchor bars, and aspect ratio and size of columns. University of California, Berkeley (Mahin, et al., 2006). Use of Partially Prestressed Reinforced Concrete Columns to Reduce Post-Earthquake Residual Displacements of Bridges Six 22% models were tested to test the effects of unbonding longitudinal mild reinforcement, providing a steel jacket, and varying the amount of pre-stressing in columns featuring post-tensioned strands and

NCHRP Project 12-102 57 mild steel. One served as a CIP reference, two served as similar base PT cases, one featured a steel jacket, one featured unbonded mild reinforcing bars, and one had a larger pre-stressing force. University of Washington (Hieber, et al., 2005). Precast Concrete Pier Systems for Rapid Construction of Bridges in Seismic Regions A parametric study of two precast bridge systems was performed using nonlinear finite element models to investigate the global response of various frame configurations. One system was of an emulative precast system with mild steel continuous throughout the members. Stanford University (Billington, et al., 2004). Behavior of Unbonded Post-Tensioning Monostrand Anchorage Systems Under Monotonic Tensile Loading Seven 1/6th scale cantilever columns featuring unbonded post-tensioning connections column segments and ductile fiber-reinforced cement-based composites (DFRCC) in the precast segments and plastic hinging regions were tested. Four were short columns and three were tall. One column in each height grouping featuring no DFRCC and the rest did. The purpose of the study was to understand the global response of post-tensioned rectangular concrete columns, investigate the cyclic response of columns with two aspect rations, investigate the effects of the DFRCC in the hinge region, and to study the energy dissipation of embedding a segment in the footing.