Testing of Connections between Prestressed Concrete Piles and Precast Concrete Bent Caps

Transcription

1 Testing of Connections between Prestressed Concrete Piles and Precast Concrete Bent Caps Dr. Paul H. Ziehl Dr. Juan M. Caicedo Dr. Dimitris Rizos Dr. Timothy Mays Aaron Larosche Mohamed ElBatanouny Brad Mustain Submitted to: Federal Highway Administration And South Carolina Department of Transportation July 2011 FY2009 Innovative Bridge Research and Deployment (IBRD) Program: Replacement of Road S-31 Bridge over Waccamaw River Swamp in Horry County Department of Civil and Environmental Engineering 300 Main Street Columbia, SC (803) This research was sponsored by the Federal Highway Administration and the South Carolina Department of Transportation. The opinions, findings and conclusions expressed in this report are those of the authors and not necessarily those of the FHWA or SCDOT. This report does not comprise a standard, specification or regulation. Federal Project No. BR26 (011) CEE

2 Copyright 2011 i

3 DISCLAIMER The contents of this report do not necessarily reflect the views of the Federal Highway Administration or the South Carolina Department of Transportation. This report does not constitute a standard, specification, or regulation. ii

4 ACKNOWLEDGEMENTS The authors thank the FHWA and SCDOT for sponsorship of this research and especially the timely and helpful input of Barry Bowers, Lucero Mesa and Ken Johnson. This work is related to an ongoing investigation of Cast In Place (CIP) bent caps, SCDOT Project 672. The input of the Steering Committee for this project is appreciated including Barry Bowers, Lucero Mesa, Ken Johnson (FHWA), and Charles Matthews. The input of the Steering Committee for Project 672 is likewise appreciated including Lucero Mesa, Barry Bowers, Ken Johnson (FHWA), Hongfen Li, Merrill Zwanka, Bill Mattison, and Ted Gettis. The authors also thank all the companies that contributed materials and time to make this project possible. This includes Florence Concrete Products of Sumter, South Carolina, Holcim Cement Company, Hardaway Concrete, CMC Rebar, Steel Specialties of Mississippi, and Glasscock Sand. The technical input of Eddie Deaver of Holcim Cement Co. and JR Parimuha of Florence Concrete Products is gratefully acknowledged. The authors thank their associates Avery Fox, William McIntosh, Shawn Sweigart, Jesè Mangual, and Daniel Learn for their hard work and help throughout the duration of this project. iii

5 ABSTRACT Currently, the South Carolina Department of Transportation (SCDOT) specifies cast-in-place (CIP) bent caps for bridge construction. However, the SCDOT recognizes the potential merit of precast bent caps and is interested in using these caps as part of the Road S-31 Bridge Replacement Project in Horry County South Carolina. The use of a precast bent cap as opposed to a CIP cap could aid in the construction time and cost of this project as well as others in the future. As a result, the University of South Carolina (U.SC) has concluded an investigation related to the viability and performance of a precast bent cap system. South Carolina is considered to be a high seismicity area. When subjected to large seismic events the ductility of bridge systems is of extreme importance. Although ongoing research of CIP caps has shown elastic response and adequate capacity protection of these elements, similar performance of precast caps is uncertain in the absence of additional research. Pile installation tolerances require large openings in the precast caps and result in precast cap dimensions that are larger than those used for typical CIP projects. Reinforcing details can be similar to those used in CIP projects but bar placement near the pile is also difficult due to construction tolerances. Finally, closure pours as required to tie the precast caps to the pile heads are typically made with non-shrink or low shrink grout and may result in less confinement of pile heads as compared to CIP projects. This report presents the results of a research project related to a prestressed concrete pile to precast bent cap connection detail. The detail achieved the necessary ductility capacity via pile hinging while protecting the cap under reverse cyclic loading. iv

6 TABLE OF CONTENTS COPYRIGHT... i DISCLAIMER... ii ACKNOWLEDGEMENTS... iii ABSTRACT... iv TABLE OF CONTENTS... v LIST OF TABLES AND FIGURES... vii CHAPTER 1 - INTRODUCTION... 1 CHAPTER 2 - LITERATURE REVIEW Introduction Precast Caps in Non-Seismic Regions Precast Caps in Seismic Regions... 5 CHAPTER 3 - SPECIMEN FABRICATION AND INTERNAL INSTRUMENTATION Introduction Precast Bent Cap Fabrication Pile Fabrication Connection Fabrication CHAPTER 4 - EXPERIMENTAL SETUP AND LOADING PROCEDURE Introduction Loading Procedure Experimental Setup Interior Specimen Instrumentation Interior Specimen Experimental Setup - Exterior Specimen Instrumentation Exterior Specimen CHAPTER 5 - FINITE ELEMENT MODELING Introduction Stressing of Reinforcement Creating the Pile Creating the Cap Material Models Load Cases Results of Analysis CHAPTER 6 RESULTS Introduction Material Performance Interior Specimen Performance Exterior Specimen Performance v

7 CHAPTER 7- RECOMMENDATIONS AND CONCLUSIONS Literature Review Proposal of Connection Detail Determining Acceptability of Proposed Detail Development of Computer Models Full-Scale Laboratory Testing Conclusions and Recommendations for Future Research REFERENCES CITED vi

8 LIST OF TABLES AND FIGURES Table 3.1 Fresh Properties of Precast Elements Table 3.2 Initial Vibrating Wire Strain Gage Readings Table 3.3 Connection Mix Design Table 4.1 Displacement Cycles Table 4.2 Instrumentation Interior Specimen Table 4.3 Instrumentation Exterior Specimen Table 5.1 Measured Concrete Compressive Strength Table 6.1 Test Specimen Strength Data Table 6.2 Test Specimen Modulus Data Table 6.3 Pile Vibrating Wire Strain Gage Measurements Figure 2.1 Connections Used by Matsumoto (2002)... 4 Figure 2.2 Connection used at St. George Island Replacement (Zendegui, 2005)... 5 Figure 2.3 Tenon and Mortice Connection (Stanton, 2006)... 6 Figure 2.4 Grouted Duct Connection (NCHRP Report 681)... 9 Figure 2.5 Cap Pocket Plan View (NCHRP Report 681) Figure 2u.sc.6 Columns used in Emulative Specimens (NCHRP Report 681) Figure 3.1 Reinforcement Interior Specimen Figure 3.2 Precast Bent Caps Prior to Casting Figure 3.3 Reinforcement Exterior Specimen Figure 3.4 Reinforcement Prestressed Piles Figure 3.5 Piles during Fabrication Figure 3.6 Connection Fabrication Interior Specimen Figure 3.7 Precast Caps upon delivery to U.SC Structures Laboratory Figure 3.8 Strain Gage Locations within Caps Figure 3.9 Strain Gage Locations within Piles Figure 3.10 Vibrating Wire Strain Gage Figure 4.1 Hydraulic Actuator Figure 4.2 Data Acquisition Systems used during Testing Figure 4.3 Displacement versus Time Figure 4.4 Experimental Setup - Interior Specimen (Schematic) Figure 4.5 Experimental Setup - Interior Specimen (Photographs) Figure 4.6 Plan View of Experimental Setup - Exterior Specimen Figure 4.7 Experimental Setup - Exterior Specimen (Photograph) Figure 4.8 Steel Plate Assembly Figure 4.9 String Potentiometers at 150 in. - Interior Specimen Figure 4.10 Joint Shear String Potentiometers Figure 4.11 Layout of Curvature Sensors Figure 4.12 Cap Rotation Sensors Figure 4.13 Actuator Connection Exterior Specimen vii

9 Figure 4.14 Actuator Connection and Shear Plate Assembly - Exterior Specimen Figure 5.1 Pile Strand Pattern Figure 5.2 Extruded Reinforcement Figure 5.3 Steps and Deformation of Pile in Creation Figure 5.4 Modeled Corrugated Pipe Figure 5.5 Modeled Exterior Specimen Reinforcement Figure 5.6 Modeled Exterior Cap Specimen Figure 5.7 Interior and Exterior Specimen Modeled Geometry Figure 5.8 Pile Concrete Stress Strain Curve Figure 5.9 Bent Cap and Connection Concrete Stress Strain Curves Figure 5.10 Modeled Boundary Conditions Figure 5.11 Pile Cracking Moment - Interior Specimen Figure 5.12 Tensile Stress in Pile at Pile to Bent Cap Connection-Interior Specimen Figure 5.13 Tensile Stress in Pile at Pile to Bent Cap Connection-Interior Specimen Figure 5.14 Pile Crushing Moment - Interior Specimen Figure 5.15 Compressive Stress in Pile at Pile to Bent Cap Connection-Interior Specimen Figure 5.16 Compressive Stress in Pile at Pile to Bent Cap Connection-Interior Specimen Figure 5.17 Yield Moment - Interior Specimen Figure 5.18 Pile Cracking Moment - Exterior Specimen Figure 5.19 Tensile Stress in Pile at Pile to Bent Cap Connection-Exterior Specimen Figure 5.20 Tensile Stress in Pile at Pile to Bent Cap Connection-Exterior Specimen Figure 5.21 Pile Crushing Moment - Exterior Specimen Figure 5.22 Compressive Stress in Pile at Pile to Bent Cap Connection-Exterior Specimen Figure 5.23 Compressive Stress in Pile at Pile to Bent Cap Connection-Exterior Specimen Figure 5.24 Yield Moment - Exterior Specimen Figure 5.25 Stress Progression in Pile - Exterior Specimen Figure 5.26 Localized Stress - Exterior Cap Figure 6.1 Unconfined Concrete Stress (ksi) Strain Curve Figure 6.2 Confined Concrete Stress (ksi) Strain Curve Figure 6.3 Pile Cross-Section and Mesh Figure 6.4 Lateral Force vs. Displacement - Interior Specimen Figure 6.5 Moment vs. Rotation - Interior Specimen Figure 6.6 Moment vs. Displacement - Interior Specimen Figure 6.7 Crack Locations - Interior Specimen Pile Figure 6.8 Moment vs. Curvature - Interior Specimen Pile Figure 6.9 Joint Region following a Displacement of 6.0 in. - Interior Specimen Figure 6.10 Equations used to Calculate Joint Shear Stress Figure 6.11 Lateral Force vs. Displacement - Exterior Specimen Figure 6.12 Moment vs. Rotation - Exterior Specimen Figure 6.13 Moment vs. Displacement - Exterior Specimen Figure 6.14 Crack Locations - Exterior Specimen Pile Figure 6.15 Moment vs. Curvature - Exterior Specimen Pile Figure 6.16 Joint Region following a Displacement of 7.5 in. - Exterior Specimen viii

10 CHAPTER 1 - INTRODUCTION Currently the South Carolina Department of Transportation (SCDOT) specifies cast-in-place (CIP) bent caps for bridge construction. However, the SCDOT recognizes the potential merit of precast bent caps and is interested in the viability of using them. It is well known that the use of precast bent caps can deliver time and economic advantages when compared to CIP caps. Therefore, the SCDOT is exploring the possibility of implementing the additional use of a precast bent cap system to their current bridge design practices and standards. The incorporation of precast bent caps has numerous advantages over traditional CIP bents. Most notable is the cost savings due to an accelerated construction schedule. With precast bent caps, contractors are able to order the cap systems in advance with delivery at the time construction begins. This reduces the time needed to construct form work, place reinforcement, pour concrete, allow curing to take place and finally remove formwork. During the replacement of a State Highway 66 bridge near Dallas, TX the use of precast bents saved approximately six months of construction time as the contractor was able to place a bent per day (Matsumoto et al., 2008). The decreased time of construction allows for less traffic interruption. Federal Highway Administration (FHWA) reports further benefits in the way of safety. Traditional CIP bent caps require workers to perform in somewhat dangerous locations, above waterways, and at varying heights for lengthened periods of time (Fouad et al., 2006). Environmental hazards are also reduced on the jobsite as the use of precast elements reduces the amount of concrete cast on site thus reducing the possibility for spills. Finally, the use of precast elements allows for more stringent construction tolerances as elements are cast offsite in an environment that promotes better construction techniques and quality control. The potential for a significant earthquake in South Carolina is of concern. As a result, it is critical that bridge substructures provide ductility for the bridge system while maintaining adequate moment and axial load capacity. The transportation system within the state is heavily reliant on bridges and therefore it is paramount that bridges remain reliable. Although several states have begun to investigate the implementation of a precast bent cap system, little of this effort has been completed in areas of high seismic demand. The SCDOT is currently involved in a project to replace the Road S-31 Bridge in Horry County SC. The interest for a precast bent cap system arose in this project as a result of the accelerated construction schedule needed. The University of South Carolina (U.SC) along with the Citadel has been asked to assess the viability of such a system. In addition to the assessment of viability, U.SC has also been asked to test the system under reverse cyclic lateral loading, to provide commentary on the detailing of a precast bent cap system, and to provide recommendations to the SCDOT regarding findings. As a part of this research, the University of South Carolina has undergone several tasks to efficiently and accurately provide guidelines to design a precast bent cap system in accordance with the SCDOT standards and specifications. It is desirable that the behavior of a precast bent cap system achieve full ductility capacity while protecting the cap from a design earthquake by means of forming a plastic hinge within the pile. The efforts described in this report were coordinated with an ongoing research project aimed at optimizing the connection design between precast prestressed concrete piles and CIP bent caps. 1

11 Since beginning work on this project, the University has consulted with industry partners aiding in the process to design a precast cap suitable for withstanding the seismic events expected in South Carolina. In conjunction with input from local industry members, U.SC has developed computer models to simulate the design of the cap subjected to seismic forces. The precast caps themselves were designed by a regional consulting firm. Upon completion of the assembly and instrumentation of the precast caps and piles, U.SC completed the testing of two full-scale single pile bent cap specimens. Both a single pile interior and single pile exterior specimen were fabricated and tested at the University of South Carolina s structures laboratory. The precast caps and prestressed piles were both fabricated by Florence Concrete Products of Sumter, South Carolina. The design of these experiments was intentionally similar to that of the ongoing CIP bent cap research project s preliminary results of which are reported by Larosche et al.,

12 CHAPTER 2 - LITERATURE REVIEW 2.1 Introduction In past several, several state transportation departments in cooperation with National Cooperative Highway Research Program (NCHRP) and Audit Subcommittee with assistance from the American Association of State Highway and Transportation Officials (AASHTO) have begun experimenting with the increased use of precast elements for bridge construction (NCHRP Report 472, 2004). While the use of precast bridge girders, and to a lesser extent precast bridge decks, have been in use for quite some time, the use of precast bent caps has risen nationwide only recently. Precast construction has proven to be both cost and time efficient. This has led to some organizations proposing the implementation of bridges constructed almost entirely of precast concrete elements. Such practices are already common in other areas of the world where local labor and materials are not sufficient to promote more cast in place construction. This chapter presents a literature review of precast construction practices and research both in areas with and without appreciable seismic activity. Please refer to SCDOT project 672 for additional literature addressing the behavior of the connection between precast prestressed piles to cast in place concrete bent caps. 2.2 Precast Caps in Non-Seismic Regions Recently, both Florida and Texas have begun employing the use of precast bent cap systems. Though the level of seismic activity in these states is relatively low, the work is still significant. Work performed at the University of Texas by Matsumoto et al. (2002 and 2008) concentrated on the development of a bent cap system without the need to withstand seismic loads. The first phase of this research focused on two connection types between bridge columns and caps (see Figure 2.1). Both connections were made by dowels extending from the columns into voids cast into the bent caps. The voids were then filled with grout to secure the connection. For the round column to precast cap connection, two or more tapered voids of rectangular cross section were cast into caps at each connection location. The cap also incorporated the use of a spiral reinforcement detail to better confine the connection. For the square column to precast cap connection, four cylindrical voids were cast through the depth of the cap. These voids were lined with a corrugated pipe, and the four together were confined by spiral reinforcement. In the first phase of research, the authors conducted thirty two pullout tests between the two connections using both standard and headed epoxy coated dowels. The research showed that little slip occurred in either of the two connections, confining spiral reinforcement was of little consequence, and that grouts used with compressive strength equal to that of the precast cap provided satisfactory anchorage for the dowels. This work is also reported on by Brenes et al

13 Figure 2.1 Connections Used by Matsumoto (2002) Additional research by Billington et al. (1999) proposed the use of a segmental precast bridge substructure. The authors proposed that precast bent caps be constructed with reduced weight by use of post-tensioning, pre-tensioning, or some combination of the two. The cap system recommended in the work is an inverted T configuration connected to segmented precast columns. It is recommended that this connection be made via a looped PT strand. This strand is looped through the bent, segmented column, and foundation being terminated at both ends in recessed areas of the bent cap. Although ship impact and not seismic activity controlled the design of the St. George Island Bridge in Florida, a fixed connection between bent cap and piles was still desired as a result of significant lateral forces. Completed in 2004 this bridge spans 4.1 miles. The implementation of a precast bent cap system aided contractors by reducing environmental impact and decreasing the time of construction. As reported by Zendegui et al. (2005), precast bent caps were connected to spun cast cylindrical driven piles with a suspended structural steel pipe (see Figure 2.2). Once driven, bent caps were placed atop the 54 in. diameter spun cast piles. A thin layer of grout connected the two. At this point a steel pipe was suspended within the bent cap extending a distance into the hollow pile. The void was then filled with cast in place concrete securing the connection. Although this work does not report any testing of the connection, the connection itself is worth noting due to its similarity to that desired by the SCDOT. 4

14 Figure 2.2 Connection used at St. George Island Replacement (Zendegui, 2005) The State of Alabama has also expressed interest in the use of precast bent caps as evidenced by the report composed by the University Transportation Center for Alabama. The report aimed to recommend and design a segmental bridge by maximizing the use of precast elements. The report selected the use of bent caps that would be placed onto the top of piles or columns through the use of a template similar to that proposed by Billington et al. (1999). Reinforcement extending from the columns was inserted into the cap though grouted sleeves similar to a connection tested by Matsumoto et al. (2002). In this test as with several others presented, experimental work was recommended though not performed. 2.3 Precast and Cast in Place Caps in Seismic Regions Stanton et al. (2006) reported Rapid Construction Details for Bridges in Seismic Zones in which proposed systems and connections of bridge bents in seismic regions are presented. Four separate bridge bent systems along with five connection types are outlined. The different bent systems are divided by the amount of column that will be precast as an integral part of the cap itself. The first system presented is referred to by the authors as a Post and Beam or PB system. The authors present this system as similar to that desired by the SCDOT. Here the bent cap is cast separate from columns or piles. Some type of connection is required in the field. The main advantage to the use of this system is listed as the ease of transportation. The authors state that the use of this system requires the connection to be designed properly. The second system referred to as a mid-column connection system is such that about half of the columns are to be precast with the connection to the bent cap made at the pre-casting plant. These types of caps are then joined at about mid-height of the columns in-situ. The connection is made with grouted ducts and dowels between the two elements or through the use of a steel sleeve between the two. The authors state that because this connection is at mid-height of the columns the forces at the connection are minimized. 5

15 Although interesting in concept for concrete columns supported on a foundation system, such systems cannot be used in practice for precast pile bent caps where alignment of precast pile splices would not be practical with the cap already connected. A third bent system is presented as being a portion of the bent cap cast with each column to which it provides support. With this system the connection between individual portions of the cap and alignment of the caps are of chief concern. These portions of the bent cap are then connected together in the field with small cast in place pours. The fourth and final system is the entire bent cap and columns being cast together. Of the available connections presented what the authors call a Tenon and Mortice connection is most similar to that used in the studies completed in this report (see Figure 2.3). This connection is much like the grouted sleeves or pockets presented earlier in this review. Unlike grouted pockets or sleeves which incorporated the connection between several bars in this connection, a single large element is used to make the connection. In the case of this project the large element is the pile itself. In other cases steel piles and large steel pipes have been used to make the connection. The advantage of this connection is listed in the work as its simplicity and low time of construction. Figure 2.3 Tenon and Mortice Connection (Stanton, 2006) The state of Washington has also reported on the use of precast bent cap systems (Hieber et al., 2005). In the referenced report, a brief review of completed work is accompanied by recommendations of a bent cap system. The authors propose the use of a grouted duct and post tensioned hybrid system for use in the seismic areas of Washington State. Though no experimental testing is reported, the authors were able to detail the use of a pushover analysis which was the basis of their findings. Similar to the work presented by Zendegui et al., Asnaashari et al. (2005) also details the construction of a bridge using precast bent caps. The authors present the construction processes during the construction of the San Mateo-Hayward bridge widening project spanning the San Francisco Bay, a location of known high seismic activity. During this project bent caps were fabricated to a depth of 2 ft. with a circular void to allow reinforcement from driven piles to penetrate the cap. Once placed over the piles with pile reinforcement extending through the cap, traditional CIP concrete was placed to achieve a desired depth of the cap of 7 ft. This connection is similar to that reported in Matsumoto et al. (2002). 6

16 Steunenberg et al. (1998) show results of precast concrete caps connected to steel piles. This work was performed in conjunction with the Ministry of Transportation and Highways of British Columbia as an investigation of the performance of presently implemented modular bridge construction. A precast cap 67 in. long x 30 in. wide x 32 in. deep were cast with a steel plate of 26 in. x 24 in. x 2 in. cast flush into the bottom of the cap. A steel pile was then welded to the plate to form the connection. The specimen was subjected to lateral displacement cycles to simulate seismic loads. No axial load was applied to the pile. For experimental data coupled with linear and nonlinear models, the authors found that the connection was adequate for strength, failure (by formation of a plastic hinge within the pile), and ductility requirements. The National Cooperative Highway Research Program has recently published Report 681 (NCHRP Report 681, 2011). In addition to providing a summary of implemented precast bent cap systems at the time of the report, the authors designed and tested seven different connection types. The different connection types were classified as either emulative or hybrid with respect to a traditional cast in place bent cap of which one was constructed and tested to serve as a control. Emulative specimens were designed for their performance to emulate that of a CIP bent cap. These connection types are designed to form a plastic hinge mechanism just below the soffit. Hybrid specimens are designed to dissipate energy through rocking at the location of a joint just below the soffit. These connections are also often classified by their hysteretic response and damage mechanisms. Emulative connections form complete plastic hinges in the piles, result in large and full hysteretic curves, and have significant concrete spalling of the pile over a significant distance (i.e., up to one pile dimension typical) away from the soffit. Hybrid connections result in pinched hysteretic curves and more concentrated damage at the soffit interface usually as a result of rocking combined with strain penetration across the interface and into the pile cap as well. The concentrated damage is usually manifested in one large crack at the interface (i.e., just below the soffit) with minor cracking away from the soffit up to one pile dimension typical. The seven experimental specimens were created at 42% full-scale and are listed below: 1. Grouted Duct Specimen: A precast bent cap with sixteen (16) 3 in. diameter ducts, in which longitudinal reinforcement from a pile or column would be fit. These ducts are then filled or pumped with grout to complete the connection. A thin layer of grout is also used for connection at the interface between the column and cap. This detail is seen in all specimens. 2. Cap Pocket (Full Ductility): This specimen was designed with a single 18 in. diameter void cast into the cap at the connection location. This void was constructed with corrugated pipe between top and bottom layers of longitudinal reinforcement. Sonotube dams were then placed into the remaining sections. Longitudinal bars from a column then fit into this void. Traditional CIP concrete was cast into the void. Similar to the grouted duct specimen a thin layer of grout was cast at the interface between the cap and column. In this specimen a single hoop was also cast into the grout layer. 3. Cap Pocket (Limited Ductility): This specimen was designed similar to the Cap Pocket (Full Ductility) specimen with the exception of a reduced area of longitudinal reinforcement as well as a reduced amount of reinforcement at the joint region. The similarity between this and the full ductility specimen allowed for direct comparison between the two, especially in the joint region. 7

17 4. Conventional Hybrid: This specimen was designed similar to that of the grouted duct specimen. As in the grouted duct specimen reinforcement from the pile extended into eight (8) 1.75 in. diameter ducts. In addition to the grouted ducts though the cap, the cap was post tensioned to the column via a single post-tensioned cable extending through the center of the connection region. The cable was anchored at a recessed region at the center of the cap. 5. Concrete filled pipe hybrid: The bent cap in this specimen was the same as that used in the conventional hybrid specimen. The difference in this specimen is in the column. The column used in this specimen was constructed with a 1 ft.-8 in. steel tube containing concrete. 6. Dual steel shell hybrid: The cap of this specimen was designed to be the same as the previous two hybrid systems. As with the Concrete Filled Pipe specimen, the column creates the difference between this and other specimens. The column in this specimen was similar to that of the concrete filled pipe specimen with the addition of a second steel pipe within the column. The second pipe was constructed of a corrugated steel pipe. This construction detail led to an essentially hollow column. 7. Integral System: In addition to the six other specimens reported an Integral System was also created and designed with girders post-tensioned to the bent cap. Testing and construction of this specimen was drastically different to that of the other specimens. However, this work is similar to that presented by Holombo et al. (2000). In construction of specimens, a non-shrink grout was designed to attain a strength of at least 500 psi greater than that of the surrounding cap. In the case of the cap pocket specimens a normal weight concrete was used for the connection as opposed to the non-shrink grout. The normal weight concrete used in these specimens was also designed to attain a strength of 500 psi greater than the surrounding concrete. This was in an effort to ensure failure of the cap prior to that of any grout within the individual specimen. Testing of each of the specimens was performed with force controlled cyclic loading to an estimated point of yield, at which point lateral load was applied with displacement controlled cycles. Short stub columns were connected to caps given the specific detail of the specimen. Caps were then inverted and tested with a combination of two actuators. The first actuator placed in a vertical position applied a representative dead load while the second placed horizontally applied lateral load and displacement. Of the three emulative specimens tested, all three were reported to have performed with stable hysteretic behavior without a significant strength reduction. Further, through testing, each of the three developed the desired plastic hinge at a location just below the connection. Also similar to each specimen was an achieved displacement ductility ratio of 8. The failure mechanism in each of the three specimens was also similar. In each specimen, failure was attributed to the fracture of two reinforcing bars which in part comprised the connection. Although significant spalling was reported in each specimen, only the Limited Ductility Cap Pocket specimen was reported to have significant cracking at the joint region. 8

18 Though the Limited Ductility specimen exhibited larger joint strains, this softening of the joint (due to a limited amount of reinforcement in the region) lead to a delay in the onset of spalling at the same location. The softer joint region also led to a higher amount of bar slip during testing, though no significant loss of anchorage was reported. These specimens are compared with the traditional CIP specimen tested in the same manner. The cast in place specimen was reported to have developed the desired plastic hinge region and achieved a displacement ductility of 10 while performing in a similar manner to the emulative specimens in terms of hysteretic behavior (stable without significant degradation). Similar to the emulative specimens spalling did occur in the column, while spalling was not observed at the joint region. The specimens were concluded to have behavior similar to that of the traditional cast in place specimen created as a benchmark. As such, the authors have proceeded to recommend slight changes to current design practice in order to account for differences between CIP and precast bent caps. The report goes on to state that following the recommended practices presented within the report a functional precast bent cap can be designed and implemented within any area of seismic activity. Figure 2.4 Grouted Duct Connection (NCHRP Report 681) 9

19 Figure 2.5 Cap Pocket Plan View (NCHRP Report 681) Figure 2.6 Columns used in Emulative Specimens (NCHRP Report 681) 10

20 CHAPTER 3 - SPECIMEN FABRICATION AND INTERNAL INSTRUMENTATION 3.1 Introduction This research effort is part of an Innovative Bridge Research and Design (IRBD) project led by the SCDOT and FHWA. As a part of the project a precast bent cap connection detail was developed in accordance with the SCDOT Seismic Design Specifications for Highway Bridges (SCDOT, 2008) and AASHTO Guide Specifications for LRFD Seismic Bridge Design (AASHTO, 2007). The precast bent cap was designed by a regional consulting firm with input from the research team. Guidance and recommendations with individuals from industry was taken into account for many aspects of the project. Following the finalization of a design, U.SC project members oversaw fabrication of both an interior and an exterior single pile bent cap specimen for later testing at the U.SC structures laboratory. Both bent caps and piles were fabricated at Florence Concrete Products in Sumter, SC between June 4, 2010 and June 7, Among other responsibilities, project members created test cylinders, gathered data of the fresh concrete properties, and installed instrumentation. Fresh properties related to both the piles and bent caps are shown in Table 3.1. Both the precast caps and prestressed piles were delivered to the U.SC structures lab 28 days following casting on July 5, Table 3.1 Fresh Properties of Precast Elements Parameter Bent Cap Pile Date 6/6/2010 6/6/2010 Slump (in.) Air Content (%) Temp. Ambient ( F) Temp. Concrete ( F) Unit Weight (lb/ft 3 ) Precast Bent Cap Fabrication Precast bent caps were designed by STV Incorporated with many details of the design being adopted from NCHRP Project Once finalized, precasting beds at Florence Concrete products were modified for specimen casting. All reinforcing within the caps was type ASTM A706. Bent caps were center of the bars. No. 5 stirrups were positioned along the length of each cap with a 5 in. center to center spacing between reinforcing. Note that subsequent figures from the test setup show the top and bottom bars reversed due to the orientation of the test set cast June 6, 2010 using a class 4,000 design mix. Project members from U.SC were on site during casting to inspect that all reinforcement was assembled properly, to document casting, install instrumentation, and to create test cylinders for future laboratory testing. 11

21 Measurements of slump and air entrainment are shown in Table 3.1. At 28 days the bent caps achieved a compressive strength of approximately 6,400 psi. To accommodate later pile embedment the dimensions of each cap were significantly larger than that of CIP caps previously tested at U.SC. Each cap was cast with a 3 ft. diameter combination corrugated pipe and Sonotube combination void through the cap, centered about the point of pile embedment. During casting, this cavity was sealed with a Styrofoam cover to prevent concrete from entering the cavity. Reinforcement of the caps consisted of fourteen (14) No. 9 bars with standard 90 hooks for longitudinal reinforcement. Eight of these bars were positioned at the top of the cap. The remaining six (6) bars were positioned at the bottom of the cap with a 3 ½ in. edge distance measured to the up. Interior Specimen The single pile interior bent cap specimen was fabricated with the following dimensions: Length: 6 ft in. Width: 4 ft. 6 in. Depth: 3 ft. 6 in. This specimen was designed to be representative of a typical interior bent connection. Reinforcement design of the specimen is shown in Figure 3.1. To accommodate a ½ in. plate between the laboratory floor and the specimen during testing, the typical 7 ft. length of the specimen was reduced by ½ in. The 3 ft. diameter cavity was located in the center of the cap. The interior specimen was fit with six (6) CEA-WA strain gages manufactured by Vishay Micro-Measurements. These strain gages were welded to longitudinal reinforcement at a position centered about the embedment of the pile (Figure 3.5). A photo of the specimen prior to casting is shown in Figure

22 Figure 3.1 Reinforcement Interior Specimen 13

23 Figure 3.2 Precast Bent Caps Prior to Casting Exterior Specimen The single pile exterior bent cap was constructed with the following dimensions: Length: 6 ft. Width: 4 ft. 6 in. Depth: 3ft. 6in. Reinforcement of this specimen was kept consistent with that of the interior specimen with the exception of typical SCDOT end reinforcement as well as a reduced stirrup spacing of 3.75 in. at the exterior end. Reinforcement of this specimen is seen in Figure 3.3. The blocked out section or cavity was centered at a point 3 ft. from the interior end of the specimen. Figure 3.7 shows a photo of the bent caps at the time of their delivery to the University of South Carolina Structures Laboratory. As in the interior specimen six WA strain gages (Vishay Micro- Measurements) were welded to longitudinal reinforcement. The location of these strain gages is shown in Figure

24 Figure 3.3 Reinforcement Exterior Specimen 15

25 3.3 Pile Fabrication Four piles were fabricated at Florence Concrete Products on June 6, Each of the four piles was 16 ft. 6 in. long with an 18 in. square cross-section. Each pile contained nine (9) ½ in. diameter low relaxation prestressing strands with guaranteed ultimate stress capacity of 270 ksi. These strands were encased in 70 turns of W6 spiral wire with 5 turns at 1 in. at each end of the pile and 60 turns at a 3 in. pitch between. Pile dimensions, strand layout, and spiral layout are shown in Figure 3.6. Figure 3.7 shows a photo of pile reinforcement. Each pile was fit with seven CEA Vishay strain gages as well as a Model 4,200 Geokon vibrating wire strain gage. Strain gages were placed to maximize recorded strains during testing. The location of these strain gages is shown in Figure 3.9. Three strain gages were mounted to selected strands following stressing at 13 in. from the embedded end of the pile. Three additional strain gages were placed at a distance of 29 in. from the same end on the same selected strands. Also, one epoxied strain gage was placed on the spiral reinforcement also at a distance of 29 in. from the embedded end of the pile. The vibrating wire strain gages were oriented so that they would be parallel with the direction of displacement during testing. A photo detailing the placement of the vibrating wire strain gages is seen in Figure Vibrating wire strain gage measurements are shown in Table

26 Figure 3.4 Reinforcement Prestressed Piles 17

27 Figure 3.5 Piles during Fabrication Table 3.2 Initial Vibrating Wire Strain Gage Readings Prior to Pour Gage # u-ε Hz u-s C Following Pour Gage # u-ε Hz u-s C Prior to Strand Release Gage # u-ε Hz u-s C Following Strand Release Gage # u-ε Hz u-s C Connection Fabrication Mix Design Several tests were conducted and input from local industry and the SCDOT was used to determine a proper mix design that could be used for the connection region between the precast bent cap and the precast prestressed pile. Non-shrink grout was originally planned for the connection region. However, it was determined that the use of this product would not be economical or well-suited for field use. One of the primary purposes of this project is to reduce the economic needs of constructing a bent cap system. Therefore a mix design was developed to adequately form the connection between the two elements. 18

28 The mix design was based on a standard class 4,000 mix design employed by Florence Concrete Products of Sumter, South Carolina. This mix incorporated the use of 611 lb./yd 3 of type 1 cement acquired from Holcim Cement. Approximately 1,100 lb./yd 3 and 1,300 lb./yd 3 of fine and coarse aggregate respectively were used and received from Hardaway Concrete of Columbia, SC. From this baseline a number of modifications were made. Most importantly the coarse aggregate used was reduced in size from a number 57 (nominal maximum aggregate size 1 in.) to a number 67 aggregate (nominal maximum aggregate size ¾ in.). The reduction of the maximum aggregate size allowed for better workability and a higher level of confidence in the placement of the concrete. Several admixtures were also added to the mix design. Both high and low range water reducers were added. Glenium 7,500 and Pozzolith 80 were used as the high and low range water reducers respectively. Eclipse 4500 was also added to the mix as a shrink reducing admixture. The mix design used for the connection between elements also incorporated the use of BASF 300R, a retarding admixture to increase set time providing an increased allowable time between mixes made in the laboratory. Due to the constraints encountered in the laboratory the retarder was introduced to prevent the mix from setting up in the time it would take to add more concrete to the connection region. Quantities of all constituents of the mix can be seen in Table 3.3. Table 3.3 Connection Mix Design Cement 610 lb/yd 3 Fine Aggregate 1120 lb/yd 3 Coarse Aggregate 2080 lb/yd 3 Water 27.8 gal/yd 3 Eclipse gal/yd 3 BASF 300 R Retarder 720 ml/yd 3 BASF AE-90 Air Entrainer 90 ml/yd 3 BASF Pozzolith ml/yd 3 BASF Glenium ml/yd 3 Water to Cement Ratio 0.38 Connection Assembly The connection between the bent caps and piles was performed in such a way that the system was inverted for casting. Bent caps were placed in a level position in the soils pit within the structures lab. Piles were then placed and centered vertically into the cavity within the bent cap to a depth of 26 in. With the unsymmetrical strand layout of the piles each pile was rotated 90 to avoid a top bar effect. This position was maintained with steel bracing. BEI Duncan linear displacement sensors were then placed over exposed strand ends to record slipping that may occur during testing of the specimen. Due to limits of size within the soils pit the connection of the two specimens were cast at separate times, with the interior specimen being cast first. Casting was completed in the laboratory with a 4,000 psi design mix created by project members with input from SCDOT as well as local industry experts. The specimens were air cured for a period of two weeks before being rotated into a testing position. The interior specimen connection concrete reached a 28 day compressive strength of 5,800 psi, the exterior specimen connection concrete achieved 5,500 psi at 28 days. 19

29 Figure 3.6 Connection Fabrication Interior Specimen Figure 3.7 Precast Caps upon Delivery to U.SC Structures Laboratory 20

30 CEA-06-W250A-120 Strain Gage Figure 3.8 Strain Gage Locations within Caps 21

31 29 1 Gage to be placed on Gages to be placed on 29 and 13 1 Vibrating Strain Gage to be placed in the direction of Displacement 3 Gages to be placed on shown strands at 13 and 29 from embedded end Figure 3.9 Strain Gage Locations within Piles Figure 3.10 Vibrating Wire Strain Gage 22

32 CHAPTER 4 - EXPERIMENTAL SETUP AND LOADING PROCEDURE 4.1 Introduction Experimentation for this research project consisted of two full scale specimens: one single pile interior specimen (T-joint) and one single pile exterior specimen (knee joint). The experimental setup used for these tests was similar to that used in a concurrent research project investigating the connection between precast prestressed piles and cast in place (CIP) bent caps. Loading of the specimens was applied with a hydraulic actuator (manufactured by Shorewestern) capable of applying 135 kips in tension and compression. The actuator used during testing is seen in Figure 4.1. In both tests data was acquired with a Vishay series 7,000 data acquisition system along with a data acquisition system associated with the actuator. Both data acquisition systems are shown in Figure 4.2. Figure 4.1 Hydraulic Actuator Figure 4.2 Data Acquisition Systems used during Testing 23

33 Displacement (in.) 4.2 Loading Procedure Both tests were performed under displacement control. A full load history for both specimens is shown in Figure 4.3. Each specimen test was divided into a series of six subtests. Displacement cycles of ±0.1 in. were employed to begin each test through displacements of ±0.6 in. At this point cycles were increased to ±0.2 in. through ±1.4 in. Again displacement increments were increased at this point to ±0.3 in. through ±2.6 in. The remainder of the tests was completed with displacement cycles of ±0.5 in. from ±3.5 in. through ±8.0 in. At this level of displacement, corresponding to drift of approximately 5%, the test was concluded. Individual subtests and displacement cycles are shown in Table 4.1. Table 4.1 Displacement Cycles Subtest # Step Size (in.) Beginning Displ. (in.) Ending Displ. (in.) ±0.1 ± ±0.4 ± ±0.8 ± ±1.7 ± ±3.0 ± ±5.5 ± Time (Unscaled) Figure 4.3 Displacement versus Time 4.3 Experimental Setup Interior Specimen As previously mentioned, the experimental setup in this test is similar to that incorporated for SCDOT project 672. The interior bent cap specimen was tested in an orientation such that the longitudinal axis of the cap was perpendicular to the laboratory floor. The cap was held in this position by four threaded steel rods running through the cap which were attached to a modified reaction frame. The threaded rods were inserted through the cap by means of 2 in. PVC pipe which was cast into the specimen. The PVC was placed at 1 ft. from either end of the cap. 24

34 In order to avoid a distributed reaction between the specimen and the reaction frame steel plates were inserted between the cap and the reaction frame at the points where rods ran through the cap. Also, a ½ steel plate was placed between the specimen and the laboratory floor. The reaction frame was constructed with arms on either side of the cap at a length equal to the depth of the cap. Arms were centered about the centerline of the pile in the specimen assembly. Pivots on each arm were utilized to maintain the position of an applied axial load of 50 kips through a full range of displacements during testing. This axial load was applied as a representative dead load which would normally be imposed by a bridge superstructure. The axial load was applied through two hollow core hydraulic rams via a modified steel shape. The experimental setup of this specimen is shown schematically in Figure 4.4 and photographs are shown in Figure 4.5. Steel loading frame Stiffened steel angle Load Curvature Sensors at opposite sides of pile 50 k Axial Load Laboratory Floor ½ in. Steel spacers Joint Shear Sensors Figure 4.4 Experimental Setup - Interior Specimen (Schematic) Figure 4.5 Experimental Setup - Interior Specimen (Photographs) 25

35 To achieve desired displacements, load was applied perpendicularly to the pile with the hydraulic actuator at a distance of 140 in. from the connection between the pile and the bent cap (referred to as the soffit). The actuator was hung from a steel loading frame constructed about the point of loading and attached at the opposite end to the pile by means of a steel plate assembly. This assembly, shown in Figure 4.8, was comprised of steel plates placed both above and below the pile with thin neoprene sheets placed between the plates and the pile to prevent slipping. 4.4 Instrumentation Interior Specimen Data taken during the testing of the interior specimen was gathered through a number of sensors. Each sensor and the parameter of interest are described here and listed in Table 4.2. Two Vishay CDS-20 String Potentiometers with a range of ±10.0 in. were used to measure pile displacements at a distance of 156 in. from the soffit. Additional String Potentiometers of the same model were used to monitor joint shear by mounting two of the sensors to the cap in an X orientation in the embedment region of the pile. Four Vishay HS-50 series plunger type linear transducers with a range of ±1 in. were mounted to the pile in series to collect curvature data in the plane of displacement (top and bottom sides of the pile). The same model sensors were also used to monitor rotation of the cap during testing. These sensors were mounted between the cap and the reaction frame at points 1 ft. above and below each end of the cap. BEI Duncan 9,615 linear transducers were mounted to exposed strands at the end of the pile prior to the connection being cast. These sensors were for monitoring slippage in the strands during testing. Additionally, strain gages were monitored during testing. The locations of the strain gages are reported in Chapter 3. Photographs of sensor positions can be seen in Figures

36 Table 4.2 Instrumentation Interior Specimen Gage Type Gage Purpose/Measurement Location Model String Potentiometer Joint Shear Embedment Region on Cap CDS-20 String Potentiometer Joint Shear Embedment Region on Cap CDS-20 String Potentiometer Displacement Pile 156 in. from Soffit CDS-20 String Potentiometer Displacement Pile 156 in. from Soffit CDS-20 Linear Transducer Displacement Pile 156 in. from Soffit HS-100 LVDT Displacement Pile 146 in. from Soffit Shorewestern Actuator Load Cell Load Pile 146 in. from Soffit Shorewestern Actuator Linear Transducer Pile Curvature (4) In Series from Soffit Pile s Top Face HS-50 Linear Transducer Pile Curvature (4) In Series from Soffit Pile s Bottom Face HS-50 Linear Transducer Cap Rotation * 1ft. From Top of Cap Between Cap and Reaction HS-50 Frame Linear Transducer Cap Rotation (3) at Prestressing Strand 13 in. from Embedded end C2A LW-120 Uniaxial Strain Gage Strand Slippage (3) at Prestressing Strand 29 in. from Embedded end C2A LW-120 Linear Transducer Strand Slippage (2) At Exposed Strands Embedded end of Pile BEI 9615 Uniaxial Strain Gage Spiral Strain Pile Spiral 29 in. from Embedded end C2A LW-120 Uniaxial Strain Gage Reinforcement Strain (6) Longitudinal Reinforcement within Cap CEA-06-W250A-120 Vibrating Wire Strain Gage Confining Strain Center of Pile 9 in. from Embedded end Geokon M4000 * Refers to Gage Location while in a Testing Position 4.5 Experimental Setup - Exterior Specimen The experimental setup of the exterior specimen differed from that of the interior specimen. Through the results of a parametric study performed for SCDOT Project 672 it was found that the piles at exterior ends of a bent cap could be subjected to both compressive and tensile axial loads (Mays and Mulliken, 2008). The experimental setup used for the interior specimen was designed to impose only axial compressive loads as described above. The actuator used in the interior specimen test was mounted to the exterior specimen such that both axial tension and compression resulted due to the self-reacting nature of the test setup. Although the experimental setup differed between the interior and exterior specimens the displacement cycles were identical. The exterior specimen was rotated so that both the longitudinal axis of the cap and that of the pile were parallel with the laboratory floor. The testing orientation of the specimen is pictured in Figure 4.6 and Figure 4.7. The actuator was connected at one end to the interior face of the cap. The other end of the actuator was connected to the pile at 92 in. from the soffit. The same plate assembly used to connect the actuator to the pile during testing of the interior specimen was utilized in the testing of the exterior specimen. Connecting the actuator to the interior end 27

37 of the specimen was completed by bolting a modified W-shape 36 in. in length to the cap through twelve (12) 1 ¼ in. epoxied steel rods embedded to a depth of 18 in. A photograph of this connection is shown in Figure Steel angles of dimensions 4 in. x 3 in. x ¾ in. were placed at either side of the actuator connection to minimize slipping at this point. Throughout testing this site was monitored for any gaps between the steel angles and the actuator. In the few instances that gaps did occur they were promptly filled with aluminum shims. These shear plates are pictured in Figure With the actuator connected to the specimen in this way testing was able to achieve both axial tension and compression at positive and negative displacement cycles respectively. Further, this test setup allows the test to be self-reacting thus negating external sources of unwanted force. In this test configuration the end of the pile was supported by a specially designed roller plate resting on a lubricated steel plate. This allowed for rotation of the pile without the weight of the pile affecting the test results. This configuration also mitigated friction at the end of the pile. Hydraulic Actuator Steel Column bolted to strong floor Cap Rotation Sensors Bent Cap Curvature Sensors at opposite sides of pile Pile Actuator Load Cell Displacement Sensors at 156 Joint Shear Sensors Figure 4.6 Plan View of Experimental Setup - Exterior Specimen 28

38 Figure 4.7 Experimental Setup - Exterior Specimen (Photograph) 4.6 Instrumentation Exterior Specimen Instrumentation of the exterior specimen was similar to that of the interior specimen because the data of interest between both tests was similar. In this test, four Vishay CDS-20 string potentiometers were used to collect displacement data at 156 in. from the connection. The additional sensors allowed any out of plane rotation of the pile to be monitored. Two of the same string potentiometers were used to monitor displacements of the pile at a distance of 86 in. from the connection in order to accurately describe the motion and position of the actuator so that the axial loads could be precisely calculated. As in the interior test string potentiometers were used to measure joint shear in the connection region. Cap rotations were monitored with a set of string potentiometers placed at both the interior and exterior ends of the top of the cap. Similar to the interior specimen HS-50 series sensors were used to measure curvature. In the case of the exterior specimen these sensors were mounted to either side of the pile as the direction of displacement was changed due to the orientation. As in the interior specimen BEI Duncan linear transducers were fit to exposed strands before embedding the pile. The strain gages listed in Chapter 3 were also monitored. 29

39 Table 4.3 Instrumentation Exterior Specimen Gage Type Gage Purpose/Measurement Location Model String Potentiometer Joint Shear Embedment Region on Cap CDS-20 String Potentiometer Joint Shear Embedment Region on Cap CDS-20 String Potentiometer Displacement Pile 156 in. from Soffit CDS-20 String Potentiometer Displacement Pile 156 in. from Soffit CDS-20 String Potentiometer Displacement Pile 86 in. from Soffit CDS-20 String Potentiometer Displacement Pile 86 in. from Soffit CDS-20 LVDT Displacement Pile 92 in. from Soffit Shorewestern Actuator Load Cell Load Pile 92 in. from Soffit Shorewestern Actuator Linear Transducer Pile Curvature (4) In Series from Soffit Each side of Pile HS-50 String Potentiometer Cap Rotation * (2) Back of Cap 1 ft. from each end CDS-20 Uniaxial Strain Gage Strand Slippage (3) at Prestressing Strand 13 in. from Embedded end C2A LW-120 Uniaxial Strain Gage Strand Slippage (3) at Prestressing Strand 29 in. from Embedded end C2A LW-120 Linear Transducer Strand Slippage (2) At Exposed Strands Embedded end of Pile BEI 9615 Uniaxial Strain Gage Spiral Strain Pile Spiral 29 in. from Embedded end C2A LW-120 Uniaxial Strain Gage Reinforcement Strain (6) Longitudinal Reinforcement within Cap CEA-06-W250A-120 Vibrating Wire Strain Gage Confining Strain Center of Pile 9 in. from Embedded end Geokon M4000 * Refers to Gage Location while in a Testing Position Figure 4.8 Steel Plate Assembly 30

40 Figure 4.9 String Potentiometers at 150 in. - Interior Specimen Figure 4.10 Joint Shear String Potentiometers 31

41 Figure 4.11 Layout of Curvature Sensors Figure 4.12 Cap Rotation Sensors 32

42 Figure 4.13 Actuator Connection Exterior Specimen Figure 4.14 Actuator Connection and Shear Plate Assembly - Exterior Specimen 33

43 5.1 Introduction CHAPTER 5 - FINITE ELEMENT MODELING A finite element analysis was performed on the interior and exterior precast bent cap and pile specimens using ABAQUS FEA version The following section describes how the model was created and presents results obtained. The computer model for the cap and pile was created much the same as in the lab. This process consists of three separate jobs, stressing the prestressing strands within the pile, releasing this stress onto the pile, and then mating the prestressed pile to the cap. 5.2 Stressing of Reinforcement The first and simplest job of the analysis was to create the strands used to prestress the concrete pile created in the next step. This consists of nine strands of 0.5 in. diameter evenly spaced in a circular pattern as shown in Figure 5.1 below. The geometry was modeled using 3- D deformable solid extrusion parts arranged as shown in Figure 5.2. Each strand was extruded to 210 in. and the elastic properties of steel were applied. At this point it was not necessary to include plastic steel properties since this job does not require the material to go beyond the elastic region. This material was redefined in the final job where plastic deformation was expected to occur. A 1.5 hex mesh was then applied to the geometry using a swept advancing front pattern. After the mesh was created, a fixed boundary condition was applied to one end and a pressure load was applied to the other. This pressure load was used to pull the strands to the specified tension. The resulting deformed configuration is shown in Figure 5.3. Figure 5.1 Pile Strand Pattern Figure 5.2 Extruded Reinforcement 34

44 5.3 Creating the Pile The second step was to create the prestressed pile that was used in the analysis. The first part of this task was to import the deformed geometry created in the first job. Once this was completed a predefined field was created for the part which includes all of the material data and stresses that were created by the applied stress from the first job. Next, the spiral wire is created using a 3-D deformable solid revolved shape. The wire revolves around with a 13 in. interior diameter with a variable pitch as shown previously. Next, an 18 in. by 18 in. 3-D deformable solid square was extruded to 210 in. to represent the concrete section of the pile. Again elastic properties were used for concrete and steel to reduce computation time because the materials were not expected to approach their elastic limits. The plastic properties were added in the final step where inelastic behavior is a possibility. Again a 1.5 in. mesh was applied to both the spiral and pile using a swept advancing front pattern. The stressed prestressing strands do not need to be meshed again as the imported geometry consists of only the orphan mesh and associated data. Next, the steel spiral and strand configuration are combined with the concrete pile using an embedded element constraint. Boundary conditions were then applied in the center of both ends of the pile that allow for compression deformation in the axial direction. The forces that created the stress in the strands from the first job were not imported therefore the strands were able to transmit a compression force on the other elements of the pile to create the prestressed pile. Each individual part used in this job and the deformed configuration can be seen in Figure 5.3. The prestressed pile structure has now been simulated and is available for importing into a new job for analysis. 5.4 Creating the Cap Figure 5.3 Steps and Deformation of Pile in Creation Since the interior and exterior cap specimens have different geometries, separate jobs were used for the creation of each. Both caps incorporate the use of a corrugated pipe, the creation of which is discussed first. This part was created using a 3-D deformable revolved shell with a 0.1 in. thickness as shown below in Figure 5.4. A 1.5 in. mesh was applied to the part in the same manner as all previous parts. 35

45 Following the building of the corrugated pipe section, the reinforcement within the bent caps were created. The locations and dimensions of the steel reinforcement for each of the caps are as described previously in this report. This reinforcement was created by making 3-D deformable solid extrusions of each piece that makes up the cage, arranging them in the proper locations and applying the usual 1.5 in mesh. Once created individual meshes were combined creating the new part. Figure 5.5 shows the reinforcing cage created for the bent cap of the exterior specimen (hooks are not shown as slipping is prevented within the model). A similar cage was created representing the reinforcement design of the interior specimen. The concrete of the bent cap was created in three separate sections since different concrete material properties were applied to the sections of the cap inside and outside the pipe section. The three sections were created to represent the bent cap surrounding the corrugated pipe in each cap, as well as the concrete within the corrugated pipe in each specimen. The material properties of these sections were taken from measured properties of the physical specimens. The first section consisted of a rectangle using the outer dimensions of each cap with a 3 ft. diameter hole cut into it where the pipe section would be applied. A circular 3 ft. diameter part was then extruded to 42 in. with an 18 in. by 18 in. hole cut all the way through it. Finally, an 18 in. by 18 in. square was extruded to 16 in. to effectively plug the end of the cap that the pile does not enter. This third part was created to aid in the mesh convergence of the entire model, and not to signify a material change. These three parts were then combined into one using the assembly module and meshed together using a 1.5 in. medial axis mesh. The completed cap part is shown in Figure 5.6. As for the pile reinforcement, both the cage reinforcement and the corrugated pipe section were embedded into the cap to complete the construction of the reinforced cap. After the cap geometry was completed all of the parts from the pile construction were imported into the final model and assigned predefined fields as previously done with the prestressing strands. Surfaces were then created for the areas on the cap and pile where the two elements were mated. After this was done a tie constraint was added binding the two surfaces together with the pile being the master surface and the cap being the slave. When the analysis was started, ABAQUS adjusted the slave nodes to meet the master surface which nullified the slight deformation caused by the previous job. Once all geometries were created material models and forces were applied to complete the analysis. The total geometry of the interior and exterior specimens is shown below in Figure 5.7. Figure 5.4 Modeled Corrugated Pipe 36

46 Figure 5.5 Modeled Exterior Specimen Reinforcement Figure 5.6 Modeled Exterior Cap Specimen Figure 5.7 Interior and Exterior Specimen Modeled Geometry 37

47 5.5 Material Models In this job plasticity was included in the material models to accommodate non-linearity following yield. The model used for all of the mild steel reinforcement in both the cap and pile was created as described in SCDOT Seismic Design Specifications for Highway Bridges section Nonlinear Reinforcement Steel Model for Ductile Reinforced Concrete Members (SCDOT, 2008). In ABAQUS the plasticity was modeled by simply adding the inelastic strain along with the corresponding stress. A second steel material model for the prestressing strands used in the pile creation followed the model described in Nonlinear Prestressing Strand Model. All of the concrete parts were modeled using the Mander model as described in section of the SCDOT Seismic Design Specifications for Highway Bridges, Unconfined Concrete Model (SCDOT, 2008). The confined concrete model described in the same specification was not utilized in this analysis. This was not required as a review of the results gathered from the model showed that stresses within the confined section of the pile did not warrant the use of this model. Concrete models were based on measured properties that were obtained near the time of specimen testing. These measured values are shown in Table 5.1. The concrete stress-strain curves used in the pile and bent cap can be seen below in Figures 5.8 and 5.9, respectively. In ABAQUS the Concrete Damage Plasticity model was used to simulate the behavior of concrete. This model was designed to properly reproduce the very different responses of concrete in compression and tension. Additionally, a damage parameter was used within the Concrete Damage Plasticity model to improve the non-linear effects. These parameters (cracking for tension and crushing for compression) controlled the evolution of the yield surface and the degradation of the elastic stiffness during the analysis. The stress-strain relationships used can be seen in Figure 5.8 and 5.9 below for the pile and cap concrete respectively. 38

48 Stress(psi) Stress(psi) Table 5.1 Measured Concrete Compressive Strength Cap Part f'c (psi) Pile 8,230 Outside Pipe 7,170 Inside Pipe(Interior) 6,290 Inside Pipe(Exterior) 5, Strain(in/in) Figure 5.8 Pile Concrete Stress Strain Curve Strain(in/in) Bent Cap Concrete Connection Concrete Interior Specimen Connection Concrete Exterior Specimen Figure 5.9 Bent Cap and Connection Concrete Stress Strain Curves 39

49 5.6 Load Cases For both the interior and exterior specimens a fixed boundary condition was placed over the entire top face of the cap. The load case for the interior specimen consisted of an axial load of 50 kips applied at the end of the pile as well as a pressure load applied over a 36 in. square area centered at a distance of 146 in. from the face of the connection between pile and cap. The pressure load was increased in increments of 10 psi per step to achieve pile deflections. Due to the cyclic axial load applied to the exterior specimen another part was created for application of this load in the model. A 36 in. square, rigid plate with a depth of ¼ in. was created and tied to the pile at the point of load application (92 in. from the soffit). Also connected to the plate was an arm set at an angle similar to that used in the experimental setup. A pressure load is then applied to the end of the arm which increased by approximately 300 psi per step. This gave the model the ability to simulate loading similar to that used during experimentation. The boundary conditions and loads for each specimen are shown below in Figure Results of Analysis Interior Specimen Figure 5.10 Modeled Boundary Conditions Strain data was compiled on the pile at the pile to bent cap connection to determine when tensile cracking and crushing would occur. Data was collected on both the tensile and compressive faces of the pile. As can be seen from the graph in Figure 5.11, the pile is expected to experience cracking on the tensile face when the applied moment approaches 1,400 k-in. 40

50 Moment(k-in.) Figures 5.12 and 5.13 graphically display the stresses in the pile at the pile to bent cap connection. As can be seen in Figure 5.14 the compressive face of the pile begins to experience crushing at approximately 2,000 k-in. The stresses in the pile are shown graphically in Figures 5.15 and 5.16 at the point when crushing begins. A graph was then created to show the displacement of a point 156 in. above the pile cap connection as a function of the applied moment. As can be seen from Figure 5.17, yielding can be expected to occur at a displacement of 0.4 in. 3,000 2,500 2,000 1,500 1,400 1, E E E E E E-03 Strain (in./in.) Figure 5.11 Pile Cracking Moment - Interior Specimen Figure 5.12 Tensile Stress in Pile at Pile to Bent Cap Connection - Interior Specimen 41

51 Moment(k-in.) Figure 5.13 Tensile Stress in Pile at Pile to Bent Cap Connection - Interior Specimen 6,000 5,000 4,000 3,000 2,000 1,800 1, E E E E E E-04 Strain (in./in.) Figure 5.14 Pile Crushing Moment - Interior Specimen 42

52 Moment (k-in.) Figure 5.15 Compressive Stress in Pile at Pile to Bent Cap Connection - Interior Specimen Figure 5.16 Compressive Stress in Pile at Pile to Bent Cap Connection - Interior Specimen 3,000 2,500 2,000 1,500 1, Displacement (in.) Figure 5.17 Yield Moment - Interior Specimen 43

53 Moment (k-in.) Exterior Specimen Similar to the results of the interior specimen, for the exterior specimen strain data was collected for the pile at the pile to bent cap connection. In this specimen, as shown in Figure 5.18 cracking is expected to occur at 1,000 kip-in. The stresses in the pile at the bent cap connection are displayed below in Figures 5.19 and As shown in Figure 5.21, the side of the pile in compression at the pile to bent cap connection is expected to experience crushing at 1,200 k-in. Figures 5.22 and 5.23 show the stress distribution in the pile at the pile to bent cap connection. A moment displacement curve was created for the exterior pile showing the displacement of the pile at a point 156 in. from the pile to bent cap connection. This is shown in Figure The progression of stresses in the pile at the critical points of the analysis is shown in Figure ,000 2,500 2,000 1,500 1, E E E E E E E-04 Strain (in./in.) Figure 5.18 Pile Cracking Moment - Exterior Specimen Figure 5.19 Tensile Stress in Pile at Pile to Bent Cap Connection - Exterior Specimen 44

54 Moment (k-in.) Figure 5.20 Tensile Stress in Pile at Pile to Bent Cap Connection - Exterior Specimen 3,000 2,500 2,000 1,500 1,200 1, E E E E E E E-05 Strain (in./in.) Figure 5.21 Pile Crushing Moment - Exterior Specimen 45

55 Moment (k-in.) Figure 5.22 Compressive Stress in Pile at Pile to Bent Cap Connection - Exterior Specimen Figure 5.23 Compressive Stress in Pile at Pile to Bent Cap Connection - Exterior Specimen 3,000 2,500 2,000 1,500 1,400 1, Displacement (in.) Figure 5.24 Yield Moment - Exterior Specimen 46

56 Figure 5.25 Stress Progression in Pile - Exterior Specimen Bent Cap Results In Figure 5.26, the stress distribution for the exterior specimen due to the force applied on the pile is shown. Results from both the interior and exterior models indicate that localized crushing in the connection region is not predicted to occur prior to reaching the ultimate capacity of the pile. Figure 5.26 Localized Stress - Exterior Cap 47

57 Compressive Strength, f c (psi) Chapter 6 Results 6.1 Introduction This chapter presents the results of the full-scale laboratory testing. Material properties are presented first. Following this section the results of the test specimens are presented. 6.2 Material Performance As previously mentioned both caps and piles were cast at Florence Concrete Products in Sumter, South Carolina. Caps were cast June 6, Piles were cast the morning of the same day. Prestressed strands were released 24 hours following completion of casting. At the time strands were cut, compressive strength testing showed an average achieved strength of 6,000 psi. The connections between the caps and piles were cast on October 8, 2010 and November 12, 2010 for the interior and exterior specimens, respectively. Compression testing was performed at 7, 14, 28 and 56 days with 4 in. by 8 in. cylinders created for casting of the caps, piles, and connections. Following testing, core samples were also taken to measure strength. The results of these tests are summarized below in Table 6.1. The Young s modulus was also determined at 28 days from the time of the respective casts. The results from these tests are shown in Table 6.2. It can be seen from Table 6.1 that each of the elements tested for strength achieve the desired design strength on or before 28 days. Table 6.1 Test Specimen Strength Data Cylinder Piles Int. Cap Ext. Cap Interior Exterior Connection Connection Casting Date 6/6/2010 6/6/2010 6/6/ /8/ /12/ Day AVG Day AVG Day AVG Day AVG Day AVG Core Sample

58 Modulus Table 6.2 Test Specimen Modulus Data Int. Exterior Cylinder Int. Cap Ext. Cap Connection Connection Cast Date 6/6/2010 6/6/ /8/ /12/2010 Ec (psi) * * * * * * * * * * * * * * * *10 6 AVG. 5.43* * * *10 6 Estimated 4.59* * * *10 6 Piles used in these tests were each fit with a vibrating wire strain gage placed into the eventual embedment region of the pile. The gages were placed such that their orientation was parallel to the plane of displacements that the piles would later be subjected to. These gages were used to calculate the effective confining stress induced onto the pile by the connection. The data acquired from these gages is shown in Table 6.3. Table 6.3 Pile Vibrating Wire Strain Gage Measurements Time u-ε Hz u-s C Interior, Post Connection Cast Interior, Prior to Test Difference Exterior, Post Connection Cast Exterior, Prior to Test Difference Interior Specimen Performance General Information The interior specimen was tested over a period of two days beginning December 7 th 2010, 60 days following the casting of the connection region of the specimen. The first three subtests corresponding to displacement cycles of ±1.4 in. were completed on the first day of testing. The remaining three subtests were completed the following day. The embedment length of the pile in this specimen was 26 in. Moment Capacity Figures 6.4 and 6.5 show the lateral force vs. displacement and moment vs. rotation behavior of the test specimen. Figure 6.6 shows the moment vs. displacement plot of this specimen. The interior specimen reached a yield displacement of 0.5 in. in both the positive and negative displacement directions. The yield displacements correspond to yield moment values of 1,800 k-in. in the positive direction and 2,000 k-in. in the negative direction. Ultimate moment capacity occurred at displacement levels of ±3.5 in. corresponding to +3,100 k-in. in the positive direction and 3,010 k-in. in the negative direction. As recommended by the Seismic Design Specifications for Highway bridges a moment-curvature analysis was conducted using XTRACT. The results of this analysis are shown in Figure 6.6. The model created considering strand slipping is shown with a dashed line in the figure while the model which did not consider slipping of the strands is shown as a solid black line. 49

59 As can be seen from the figure the experimental data meets or exceeds the predicted behavior up to ductility capacities of ±8.0 in. From this point through larger displacements the experimental data exceeds that of the predicted behavior. Moment Curvature Model Moment-curvature models were created using the XTRACT program. The material properties considered in the model were taken from SCDOT SDS (SCDOT, 2008). Mander models were used to describe the behavior of unconfined and confined concrete. Two stresses were used to model the stress of the prestressing strands: 1. Ultimate strand capacity. 2. The slipping stress based on an equation to calculate the development length of prestressing strands in confined sections. Figure 6.1 Unconfined Concrete Stress (ksi) Strain Curve Figure 6.2 Confined Concrete Stress (ksi) Strain Curve 50

60 Lateral Force (kips) Figure 6.3 Pile Cross-Section and Mesh Displacement (in.) Figure 6.4 Lateral Force vs. Displacement - Interior Specimen 51