Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Retrofitting beam-to-column joints for improved seismic performance microform Hoffschild, Thomas E. 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_hoffschild_thomas.pdf [ 10.3MB ]
JSON: 831-1.0050490.json
JSON-LD: 831-1.0050490-ld.json
RDF/XML (Pretty): 831-1.0050490-rdf.xml
RDF/JSON: 831-1.0050490-rdf.json
Turtle: 831-1.0050490-turtle.txt
N-Triples: 831-1.0050490-rdf-ntriples.txt
Original Record: 831-1.0050490-source.json
Full Text

Full Text

Retrofitting Beam-to-Column Jointsfor Improved Seismic PerformancebyThomas E. HoffschildB.A.Sc. (Civ. Eng.), University of British Columbia, Vancouver, BC, 1990A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREEE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF CIVIL ENGINEERINGWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIADecember 1992©Thomas Hoffschild, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ^Civil EngineeringThe University of British ColumbiaVancouver, CanadaDate^Deremhe r 31 , 1992DE-6 (2/88)ABSTRACTBefore the 1970's, most codes for the design of reinforced concrete structures did notinclude provisions for ductility during seismic events was not prevalent in most codes. Theguidelines that did exist were minimal, and often left a fair amount of room forinterpretation by the design engineer. Hence, many of the reinforced concrete structuresdesigned during that time are suspect under today's more stringent design guidelines.Moreover, even the present designs are often deficient and vary from building to buildingand from jurisdiction to jurisdiction.This report is a presentation of the findings of an experimental study to evaluate amethod of retrofit which addresses a particular weakness that is often found in reinforcedconcrete structures, namely the lack of sufficient reinforcement in and around beam-to-column joints. Many of these structures lack the required confining reinforcement withinthe joints and in adjoining beams and columns. The result is a reinforced concrete framethat is weak in the joint area and lacks sufficient ductility during a seismic event.The proposed retrofit method consists of encasing the reinforced concrete joint witha grouted steel jacket that provides confinement to the joint area, and imparts ductility tothe frame. In this study, two styles of retrofit jacket were tested: a circular steel tube anda rectangular casing. It was found that circular steel jackets have the advantage of providingiidirect concrete confinement and, as well, of furnishing a ductile force transfer mechanismthrough the jacket itself, but are also more difficult and expensive to fabricate thanrectangular jackets. Although rectangular jackets do not provide the same degree ofconcrete core confinement as circular jackets, the amount available seems sufficient toprevent damage in the joint area. The load transfer mechanism of the rectangular jacketswas found to be adequate in withstanding the loads and deflections typical for seismic events.In this thesis, the two jacket styles are evaluated for strength, stiffness and ductility,and their relative merits are discussed.iiiTABLE OF CONTENTSAbstract^ iiList of Tables viiiList of Figures^ ixAcknowledgement xv1 Background and Proposal^ 11.1 Introduction  ^11.2 Importance of Ductility ^  21.3 Retrofit of Seismically Inadequate Structures ^  32 Previous Research^ 82.1 Introduction - Reinforced Concrete Columns^  82.2 Increaseing Ductility of Reinforced Concrete Columns^ 82.3 Confinement of Reinforced Concrete Columns by Steel Jacket^ 102.4 Problems in Existing Structures ^  112.5 Advantages of the Steel Jacketing Method^  112.6 Behaviour of Beam-to-Column Joint  132.7 Description of Retrofit Study^  143 Test Specimen Design^ 16iv3.1 Design Philosophy and Codes^  163.2 Specimen Design Considerations  173.3 Unretrofitted Test Specimens^  183.4 Retrofit Design^  184 Experimental Apparatus and Procedure^ 234.1 Introduction^  234.2 Testing Apparatus  234.3 Measurements for Rotation ^  254.4 Strain Gauges^  274.5 Testing Procedure  305 Unretrofitted Results^ 335.1 Introduction  335.2 Observed Behaviour of Specimens During Test^  345.3 Failure Mechanisms Withing the Joint Area  375.4 Material Properties ^  405.5 Comparison With Modern Seismic Design^  415.6 Measured Data: RCBC1 and RCBC2  415.7 Problems With Measuring Devices ^  455.8 Theoretical Values^  465.9 Evaluation of Specimen Performance^  486 Retrofitted Test Results^ 496.1 Introduction  496.2 RETRO-SU: Undamaged Specimen With Square Jacket^ 506.3 RETRO-CU: Undamaged Specimen With Circular Jacket  606.4 RETRO-SD: Previously Damaged Specimen With Square Jacket^ 706.5 RETRO-CD: Previously Damaged Specimen With Circular Jacket ^ 796.6 Summary of Retrofit Testing^  897 Evaluation of Retrofit Schemes 917.1 Introduction ^  917.2 Reliability of Bond Between Concrete and Retrofit Steel ^ 927.3 Improvement in Ductility ^  937.4 Positioning the Gap for Plastic Hinge Development^ 947.5 Dimensions of the Retrofit Jacket^  967.6 Confinement Effects ^  977.7 Rating the Retrofit Schemes  988 Summary and Conclusions^ 1008.1 Summary^  1008.2 Conclusions  1019 Recommendations for Further Study^ 1039.1 Clarifications and Modifications to This Retrofit Study^ 1039.2 Beyond the Simple Beam-to-Column Joint ^  104References^ 106Appendix A: Structure and Specimen Design^ 109Appendix B: Material Properties^ 112viAppendix C: Data for Unretrofitted Specimens^ 117Appendix D: Data for Retrofitted Specimens 124viiLIST OF TABLES5-1 RCBC1 and RCBC2 Section Analyses ^  476-1 RETRO-SU Section Analyses ^  606-2 RETRO-CU Section Analyses  706-3 RETRO-SD Section Analyses ^  796-4 RETRO-CD Section Analyses  89B-1 Material Properties of Concrete and Steel ^  113viiiLIST OF FIGURES1-1 Testing Program^  51-2 Beam-to-Column Joint Cyclic Loading Test^  63-1 Specimen Design^  193-2 Retrofit Concepts  224-1 Testing Frame and Specimen^  244-2 Placement of LVDTs  264-3 Rotation Types Within the Joint^  284-4 Placement of Strain Gauges  294-5 Testing Sequence^  315-1 Hysteresis Loops: Unretrofitted Specimens^  355-2 Joint Damage in Unretrofitted Specimens  365-3 Failure of Unretrofitted Specimens^  385-4 Hysteresis Loop of a Specimen of Modern Design^  425-5 Rotation Components: Unretrofitted Specimens  44ix6-1 Retrofitted Specimen RETRO-SU^  516-2 Rotation: RETRO-SU^  546-3 Strains: RETRO-SU  576-4 Retrofitted Specimen RETRO-CU^  616-5 Rotation: RETRO-CU^  646-6 Strains: RETRO-CU  686-7 Retrofitted Specimen RETRO-SD^  716-8 Rotation: RETRO-SD^  776-9 Strains: RETRO-SD  776-10 Retrofitted Specimen RETRO-CD^  806-11 Rotation: RETRO-CD^  836-12 Strains: RETRO-CD  86A-1 Frame Design ^  110A-2 Specimen Design  111B-1 Reinforcement Bar #1^  114B-2 Reinforcement Bar #2  114B-3 Reinforcement Bar #3^  115B-4 Reinforcement Bar #4  115B-5 Retrofit Jacket #1 ^  116B-6 Retrofit Jacket #2  116xC-1 RCBC1: Total Rotation ^  118C-2 RCBC1: Rotation Components  118C-3 RCBC1: Cantilever Rotation ^  119C-4 RCBC1: Rigid Rotation  119C-5 RCBC1: Shear Rotation ^  120C-6 RCBC1: Shear and Rigid Rotation ^  120C-7 RCBC2: Total Rotation ^  121C-8 RCBC2: Rotation Components  121C-9 RCBC2: Cantilever Rotation ^  122C-10 RCBC2: Rigid Rotation  122C-11 RCBC2: Shear Rotation ^  123C-12 RCBC2: Shear and Rigid Rotation ^  123D-1 RETRO-SU: Total Outer Rotation ^  125D-2 RETRO-SU: Total Inner Rotation  125D-3 RETRO-SU: Hinge Rotation at Gap^  126D-4 RETRO-SU: Rotation Components  126D-5 RETRO-SU: Cantilever Rotation ^  127D-6 RETRO-SU: Rigid Rotation  127D-7 RETRO-SU: Shear Rotation ^  128D-8 RETRO-SU: Shear and Rigid Rotation ^  128D-9 RETRO-SU: Front Column Longitudinal (FCL)^  129xiD-10 RETRO-SU: Front Column Transverse (FCT)^  129D-11 RETRO-SU: Rear Column Longitudinal (RCL)  130D-12 RETRO-SU: Rear Column Transverse (RCT)^  130D-13 RETRO-SU: Rear Joint Longitudinal (RJL)  131D-14 RETRO-SU: Rear Joint Transverse (RJT)^  131D-15 RETRO-SU: Side Joint Longitudinal (SJL)  132D-16 RETRO-SU: Side Joint Transverse (SJT) ^  132D-17 RETRO-SU: Top Beam Longitudinal (TBL)  133D-18 RETRO-SU: Top Beam Transverse (TBT)^  133D-19 RETRO-SU: Top Extension Transverse (TET)  134D-20 RETRO-CU: Total Outer Rotation ^  135D-21 RETRO-CU: Total Inner Rotation  135D-22 RETRO-CU: Hinge Rotation at Gap^  136D-23 RETRO-CU: Rotation Components  136D-24 RETRO-CU: Cantilever Rotation ^  137D-25 RETRO-CU: Rigid Rotation  137D-26 RETRO-CU: Shear Rotation ^  138D-27 RETRO-CU: Shear and Rigid Rotation ^  138D-28 RETRO-CU: Front Column Longitudinal (FCL)^  139D-29 RETRO-CU: Front Column Transverse (FCT)  139D-30 RETRO-CU: Rear Column Longitudinal (RCL)^  140D-31 RETRO-CU: Rear Column Transverse (RCT)  140xiiD-32 RETRO-CU: Rear Joint Longitudinal (RJL)^  141D-33 RETRO-CU: Rear Joint Transverse (RJT)  141D-34 RETRO-CU: Side Joint Longitudinal (SJL) ^  142D-35 RETRO-CU: Side Joint Transverse (SJT)  142D-36 RETRO-CU: Top Beam Longitudinal (TBL)^  143D-37 RETRO-CU: Top Beam Transverse (TBT)  143D-38 RETRO-CU: Top Extension Longitudinal (TEL)^  144D-39 RETRO-CU: Top Extension Transverse (TET)  144D-40 RETRO-SD: Total Outer Rotation ^  145D-41 RETRO-SD: Total Inner Rotation  145D-42 RETRO-SD: Hinge Rotation at Gap^  146D-43 RETRO-SD: Rotation Components  146D-44 RETRO-SD: Cantilever Rotation ^  147D-45 RETRO-SD: Rigid Rotation  147D-46 RETRO-SD: Shear Rotation ^  148D-47 RETRO-SD: Shear and Rigid Rotation ^  148D-48 RETRO-SD: Front Column Transverse (FCT)  149D-49 RETRO-SD: Rear Joint Longitudinal (RJL)^  149D-50 RETRO-SD: Side Joint Longitudinal (SJL)  150D-51 RETRO-SD: Side Joint Transverse (SJT) ^  150D-52 RETRO-SD: Top Beam Longitudinal (TBL)  151D-53 RETRO-SD: Top Beam Transverse (TBT)^  151D-54 RETRO-SD: Top Extension Transverse (TET)^  152D-55 RETRO-CD: Total Outer Rotation ^  153D-56 RETRO-CD: Total Inner Rotation  153D-57 RETRO-CD: Hinge Rotation at Gap^  154D-58 RETRO-CD: Rotation Components  154D-59 RETRO-CD: Cantilever Rotation ^  155D-60 RETRO-CD: Rigid Rotation  155D-61 RETRO-CD: Shear Rotation ^  156D-62 RETRO-CD: Shear and Rigid Rotation ^  156D-63 RETRO-CD: Front Column Longitudinal (FCL)^  157D-64 RETRO-CD: Front Column Transverse (FCT)  157D-65 RETRO-CD: Rear Joint Longitudinal (RJL)^  158D-66 RETRO-CD: Side Joint Longitudinal (SJL)  158D-67 RETRO-CD: Side Joint Transverse (SJT) ^  159D-68 RETRO-CD: Top Beam Longitudinal (TBL)  159D-69 RETRO-CD: Top Beam Transverse (TBT)^  160D-70 RETRO-CD: Top Extension Transverse (TET)  160xivACKNOWLEDGEMENTThe author is very grateful to his supervisors, Professor Helmut Prion and ProfessorSheldon Cherry for their guidance, suggestions and encouragement throughout this research.The author also wishes to express his gratitude to Professor Carlos Ventura for reviewingthe manuscript.Appreciation is extended to Mr. Paul Symons, Mr. Ron Dolling, Mr. Guy Kirsch, Mr.Bernie Merkli, Mr. Richard Postgate, Mr. Harald Kullmann and Mr. Andrew Zyp for theirhelpful participation and assistance during the experimental investigation.Financial support through an NSERC operating grant is also gratefully acknowledged;and the author wishes to thank his friends and family for their support throughout hisuniversity career.xvCHAPTER 1BACKGROUND AND PROJECT PROPOSAL1.1 INTRODUCTIONIn the past, earthquakes have occurred in many different areas of the world, andCanada has not been immune to this natural disaster. In fact, parts of Canada face anextreme risk of a very large earthquake sometime in the future. In such hazard regions,considerable attention must be paid to the seismic design and earthquake construction ofbuildings and other structures, as well as to the seismic capacity of existing structures. Olderreinforced concrete buildings are of special concern. Many of these do not meet thereinforcing and detailing requirements of modern building codes and therefore are extremelysusceptible to the effects of cyclic loads, because of the brittle nature of the material.Although possessing superior strength in compression, concrete has poor materialcharacteristics in tension and relies primarily on reinforcing steel for strength and ductility.Large structures are routinely constructed of reinforced concrete, the design of which isgoverned by appropriate design codes, such as the ACI Standard 318-1983 in the UnitedStates, and the CAN3-A23.3-M84 standard in Canada. These codes are compiled andregularly updated by technical experts using current research information. Over the years,1as more and more information regarding the behaviour of concrete structures under seismicaction becomes available, engineers gain a better understanding about the design ofstructures for seismic effects, and design codes are continually modified to reflect thisincrease in knowledge.In the 1950's and 1960's, design and construction practices were different than theyare today, especially where seismic effects are concerned. In particular, the relatively smallamount of transverse reinforcement prevalent in the pre-1970's designs has resulted in theexistence of many moment-resisting frames that today are considered to have a high risk ofdeveloping shear failures in the columns and of exhibiting a low level of ductility and energydissipation ability. If such structures do not include other structural elements that maycontribute to the overall ductility and provide other sources for dissipating energy, they mayprove to be seriously inadequate when compared to current design standards.A major turning point for seismic design, at least in North America, was the 1971 SanFernando Valley earthquake in California. Many reinforced concrete structures that werebuilt to the design and construction standards at that time behaved poorly during theearthquake, prompting many modifications to the then existing ACI and Canadian codes.1.2 IMPORTANCE OF DUCTILITYMost code modifications referred to above, and implemented after the San Fernandoearthquake addressed the ductility requirements of reinforced concrete members and, inparticular, much attention was focused on the joint region between beams and columns. Thejoint region is subject to large shear forces during lateral seismic loading, particularly when2beam moments on opposite faces of a column have the same direction. The longitudinalreinforcement is stressed in the same direction in this situation, and the bond between thereinforcement and the concrete is heavily relied upon to provide the required transfer offorces through the joint. Under severe seismic loading, plastic hinges are expected to format the ends of the beams adjacent to the joint, and transverse reinforcement in both thebeam and the column are required to provide confinement to the concrete in the coreregion, thereby safeguarding the ductility of the joint. All of the aformentioned points wereexamined following the 1971 San Fernando Valley earthquake and the design codes weremodified accordingly, in line with the recommendations of ACI-ASCE Committee 352 forthe design of connections in reinforced concrete frames with ductile moment-resistingcapacity (ACI, 1991).While updated design codes address the construction of new structures, structures thatwere built according to earlier design codes may not meet today's seismic standards. Manyare inadequate, and may pose a severe risk to society. What can be done about them? Oneavailable option is to retrofit such structures. Retrofit is the process of modifying an existingstructure so that it meets the current code design provisions.13 OBJECT AND SCOPE OF INVESTIGATIONThe purpose of this study was to investigate several methods of retrofit that can beapplied to strengthen the joint region of a beam-to-column connection in a reinforcedconcrete frame built in accordance with earlier design codes. One of the more commoncharacteristics of structures designed in accordance with outdated codes is that they possess3insufficient transverse reinforcement in the joint core, resulting in inadequate ductility in theconnection area. The method of retrofit that was studied here is to encase the deficient (ordamaged) reinforced concrete members and joint with a steel jacket, and to fill the voidbetween the jacket and the joint members with concrete grout. This is an elegant andsimple solution; the steel provides shear reinforcement and adds ductility to the joint throughits action of confining the core concrete. This approach can be applied to both existing,undamaged but deficient structures and, when appropriate, to existing structures which havebeen damaged by an earthquake.The effectiveness of introducing a steel jacket around the beam-to-column connectionwas examined by performing six cyclic loading tests on a total of four specimens builtaccording to typical 1960's design specificatons (fig. 1-1). The testing program included twotests on unretrofitted specimens, causing some initial damage and intended as a likelyscenario after an earthquake. The last four tests were performed on the two damaged andthe two undamaged specimens after they were retrofitted with a confining steel jacket (fig.1-2). Two jacket types were tested, and the results compared. Two of the specimens wereretrofitted with a circular jacket, the other two with a rectangular jacket. Each jacket typewas tested with one damaged and one undamaged specimen, so as to facilitate thecomparison of the effect of the initial damage on the behaviour of the retrofit. However,due to changes in the design details of the retrofits which followed the outcome of some ofthe initial tests, a comparison between previously damaged and undamaged specimens wasnot possible.It is well known that circular jackets provide superior confinement and hence better4Figure 1-1 Testing Program5Figure 1-2 Beam-to-Column Joint Cyclic Loading Test6ductility to the core concrete than square or rectangular jackets (Mander et al, 1988).However, in the joint region, where rectangular members meet, it might be more practicalto use a rectangular casing. The relative effectiveness of the jacket geometry on thebehaviour of the joint region is compared with existing studies on their effectiveness on thebehaviour of plastic hinge regions of columns alone.The specifics of the steel jacketing used in each retrofitting case conformed tomodern Canadian Code design (CAN3-A23.3-M84). Some of the parameters consideredincluded: the overall size dimensions of the steel section, the diameter to thickness ratio (orwidth to thickness ratio) of the steel section, the use of grout or concrete, the appropriatelength of reinforcement (hinge area vs. entire length of section), and the need for steelsection continuity through a joint.The tests that were performed as a part of this study reflect only the simplest ofconnections, that is to say a simple external beam-to-column connection. As far as standarddesign and construction procedures are concerned, this is a dramatic oversimplification.Under normal circumstances, a reinforced concrete structure will have joints that consist ofmore than just one beam framing into a column. As an example, a typical interior joint willconsist of four beams as well as a slab framing into the column. Retrofitting a joint of thistype with a steel jacket would be a much more complicated proposition. However, thisresearch was intended as a preliminary study to determine the retrofit behaviour ofincorporating steel jackets at a connection. For this purpose, a simple beam-columnconnection should be sufficient to determine whether or not the retrofit method proposedis worthy of further research in more realistic situations.7CHAPTER 2PREVIOUS RESEARCH2.1 INTRODUCTION - REINFORCED CONCRETE COLUMNSIt is well known that the design of reinforced concrete columns for seismic loadsrequires attention to detailing in the plastic hinge zone so that adequate strength andductility can be achieved. Sufficient transverse reinforcement must be provided for thispurpose. Closely spaced transverse reinforcement confines the concrete core of the column,and not only substantially increases the compressive strength of the column, but alsoenhances the ductility. Design codes require a certain level of ductility in columns to ensurethat brittle failures do not occur, and to provide a certain level of post-yielding strength. Itis important to note that high axial loads adversely affect the strength and ductility ofcolumns. In all cases, the codes require that the shear strength of columns be sufficient toensure that the flexural strength is reached in advance of a brittle shear failure, which isstrictly forbidden.2.2 INCREASING DUCTILITY OF REINFORCED CONCRETE COLUMNSThe usual method of providing confinement in reinforced concrete columns is to8provide transverse reinforcing steel in the form of hoops or spirals, depending on thegeometry of the column. To increase the effectiveness of the confinement, and henceincrease the strength and ductility levels, a number of options can be considered when usinghoops or spirals as transverse reinforcement. An increase in the transverse steel ratio, eitherby increasing the size of the transverse steel, or by decreasing the spacing or pitch of thereinforcement, was found to increase the compressive strength and ductility of columns(Mander et al, 1988).It has also been shown that circular hoops are more effective for confinement thansquare or rectangular hoops, as the confinement stress is equally distributed along the entireperimeter of the spiral, and not heavily concentrated at corners, as is the case in non-circularlateral reinforcement (Mander et al, 1988). The problem with circular structural members,and hence circular reinforcement patterns, is that they are much more difficult to construct,particularly in joint regions. Economic considerations usually dictate the use of rectangularmembers in construction above circular members.Although an increase of the steel ratio is generally used to enhance the strength andductility of columns, other methods have been either tested or adopted to achieve similarresults. Increasing the yield strength of the steel also increases the strength and ductility ofthe columns (Muguruma, 1984). The longitudinal reinforcement geometry can also beadjusted to assist in the confinement of the concrete core (Sheikh and Uzumeri, 1982).Spreading the steel around the perimeter of the core by using a larger number of smallerbars is more effective in providing confinement than having only a few large bars. Otherattempts at either providing or improving strength and ductility have included the use of9steel fiber reinforced concrete (Ganesan and Murthy, 1990), welded-wire fabric (Razvi andSaatcioglu, 1989) and prestressed bolts (Cheong and Perry, 1991). Limited success has beenachieved in improving strength and ductility by increasing the concrete strength (Shin et al,1989).2.3 CONFINEMENT OF REINFORCED CONCRETE COLUMNS BY STEEL JACKETAn alternative method of providing confinement for concrete in columns is to encasethe concrete core within a hollow steel member. In new construction this is typicallyachieved by filling hollow steel sections with concrete. When reinforced concrete columnsneed to be retrofitted, they can be enclosed in a steel jacket, and the space between the twoelements filled with grout. In both cases, the external steel completely encloses the entireconcrete core, and effectively confines all the concrete, inclusive of the cover concrete, thusassisting in reducing bond failures in columns. Confining the concrete core with a steeljacket also prevents the steel from buckling, and improves the behaviour of the column tothe extent that the strength of the confined column is greater than the combined separatestrengths of the steel and concrete (Tidy, 1988). It is important that the steel jacketingprovide confinement only, and should not actually carry any of the axial load. Applying anaxial load to the steel jacket as well as to the concrete core, and designing the retrofit witha continuous steel jacket, would alter the characteristics of the structure, resulting in lowerductility ratios and higher moment capacities at the critical sections (Priestley et al, 1990 ;Priestley and Park, 1985). It is absolutely essential to maintain a shear capacity that ishigher than the corresponding moment capacity to ensure ductile failure modes.102.4 PROBLEMS IN EXISTING STRUCTURESBefore the early 1970's, the seismic design provisions were significantly different thanthey are today. The requirements for seismic design were less stringent, and reflected theamount of knowledge available at that time regarding the behaviour of reinforced concreteunder cyclic loading. In terms of present standards, earlier designs of reinforced concretecolumns commonly exhibited poor detailing of transverse reinforcement, either in terms ofthe use of ties, anchorage of ties and hoops, or simply excessive spacing of transversereinforcement (Mitchell, 1991). Other problems included inadequate lapping of longitudinalreinforcement in hinge areas (Priestley and Park, 1984) or, in the case of short columns,designs in which the flexural strength exceeded the shear strength, often resulting in brittlefailure modes (Chai et al, 1991). As a result of experience gained from recent earthquakes,and in particular the San Fernando earthquake of 1971, an extensive retrofit program wasundertaken in California to upgrade existing structures to modern earthquake coderequirements. Although all new structures are, of course, built to modern standards, thelargest percentage of existing structures in the different seismic regions of the U.S. andelsewhere (including British Columbia), do not meet modern specifications, and may needto be upgraded.2.5 ADVANTAGES OF THE STEEL JACKETING METHODA favoured method of retrofitting reinforced concrete columns is to either partiallyor fully encase the member in a steel jacket. In the case of short columns, the entire columnlength is typically encased, as shorter reinforced concrete columns are particularly susceptible11to brittle shear failure (Chai et al, 1991). In columns with higher slenderness ratios, whichare the ones usually found in structures, only the plastic hinge region need be confined ina steel jacket. The steel jacket is expected to provide confinement only, and enhances theductility of the column, so as to reliably provide a load carrying mechanism after generalyielding of the reinforcement. Columns which have to support large axial loads needjacketing along the entire length to enhance the compressive strength as well. In this case,the steel jacketing is not expected to actually carry any of the axial load, a function that isleft to the added concrete alone.Numerous studies dealing with the effectiveness of different methods of confinementof reinforced concrete columns have either been completed or are presently ongoing. Theinvestigation presented here focuses on the retrofitting of reinforced concrete members atframe connections. In particular, attention was paid to the behaviour of connectionsbetween beams and columns that have been retrofitted after suffering some initial damage.The object of this study was to assess the strength and ductility improvement in thebehaviour of a retrofitted beam-to-column joint.As is the case with hoops and spirals, in reinforced concrete members and columnsretrofitted with steel jackets it is expected that a beam-to-column joint that has beenretrofitted with a circular jacket will exhibit more desirable characteristics than a jointretrofitted with rectangular or octagonal jackets. Columns retrofitted with a circular jacketwere found to exhibit greater ductility due to confinement, a larger overall strength, andgreater bond strength between the concrete and both the reinforcement steel and the steeljacket (Morishita et al, 1988).122.6 BEHAVIOUR OF BEAM-TO-COLUMN JOINTDuring extreme cyclic loading, which is common during a significant seismic event,the forces within a beam-to-column connection are relatively large and often approach orexceed the load-carrying capacity of the connection. The connection is considered as partof the column, and ideally should behave elastically, without the development of yield hingeswithin the joint, in order to avoid failure of the shear panel, and subsequent anchoragefailure of the reinforcement (Otani, 1991). Under such conditions, the energy absorptioncapacity of adjacent plastic hinges is thus maintained without the shear or anchorage failureof the joint core (Kaku and Asakusa, 1991). To avoid collapse of the structure, plastichinges are designed to occur in the beams following the rules of standard practice: strongcolumn/weak beam, strong joint/weak element, strong shear/weak moment (Bolong andYuzhou, 1991). Modern design codes address these concepts by specifying a ratio ofstrengths between beam and column. For instance, the ACI-ASCE Committee 352 on thedesign of reinforced concrete connections specifies that the ratio of flexural strengthsbetween column and beam should be at least 1.4, so as to avoid the development of aflexural hinge in the column.Over the last 20 years, a fair amount of research has been done on the behaviour ofbeam-to-column connections, which is reflected in the recommendations by the ACI-ASCECommittee 352 for the design of beam-to-column connections. Before theserecommendations, the joint was considered to behave much like a deep beam, and designedaccordingly (Ehsani and Wight, 1990). Today we know that this is not the case, that thejoint behaves in a manner that is in some ways different and in other ways similar to a deep13beam. During seismic loading, the force transfer between the beam and the column withina joint occurs in the form of a truss action, consisting of a concrete strut and a tension tieformed by the transverse reinforcement steel within the joint, which is activated by bondstress and anchorage. Recent experimental results suggest that the concrete strut transfersmost of the force, and hence only enough transverse steel to provide confinement is requiredwithin the joint (Leon, 1990).Joint performance appears to be a function of the joint shear stress and confinementlevel (Alameddine and Ehsani, 1991). These authors showed that high joint shear stressesreduce the energy absorption capacity and cause a rapid loss of load-carrying capacity of thejoint. The primary role of transverse joint reinforcement is one of confinement (Ehsani andWight, 1990); joint deterioration due to high joint shear stress can be prevented by providingsufficient anchorage of the longitudinal steel (Leon, 1990). Other recommendations for theimprovement of joint behaviour are the avoidance of large plastic deformations within thejoint, the limitation of concentration of damage to prescribed sections, and the avoidanceof brittle failures.2.7 DESCRIPTION OF RETROFIT STUDYThe object of the beam-to-column connection retrofit/repair program under study wasto improve the behaviour of the connection region during cyclic lateral loading. Theproblems associated with joint shear failure (yield hinges within the joint, and insufficientconfinement and anchorage) were to be solved by the use of steel jackets. The intention wasto develop plastic hinges within the beam adjacent to the connection, and to increase the14ductility of the system at the same time. This would result in an improvement of the seismicbehaviour of the overall frame. The one condition that had to be avoided when retrofittingby means of a steel jacket was the potential development of a plastic hinge in the beamoutside of the retrofitted region. A plastic hinge outside of the retrofitted region wouldmean that the original problem of insufficient ductility had been moved elsewhere, and noreal advantage would have been gained. A plastic hinge that has been moved along thebeam away from the joint would also mean higher shear stresses in the unretrofitted sectionof beam, resulting in a higher risk of brittle shear failure.As noted earlier, it is a well-established fact that circular and elliptical steel jacketsprovide superior performance in retrofitted columns compared to jackets of square,rectangular or other geometrical configurations. However, the expected advantages ofincorporating such jackets in the retrofit of beam-to-column connections have not beenadequately studied. The forces that act within a joint are somewhat more complicated thanthose developed in the isolated column retrofit case. It was the intention to investigate therelative merits of both the circular and rectangular jackets surrounding a beam-to-columnconnection. Whereas the circular jacket was expected to behave in a superior fashion, therectangular jacket was simpler to apply and might provide sufficient improvement in thebehaviour of the connection to favour its use in practice.15CHAPTER 3TEST SPECIMEN DESIGN3.1 DESIGN PHILOSOPHY AND CODESThe aim of this research project was to observe the expected joint damage in an oldercode-designed reinforced concrete frame building under seismic load conditions and toinvestigate means of retrofitting the joint to improve its seismic behaviour. In order that thestudy be viable, practically, a realistic building was designed for the investigation and a jointsub-assembly of this structure was considered for the detailed study.A two-storey office structure situated in Vancouver, BC was first designed inaccordance with all of the requirements of the 1970 National Building Code of Canada, andprimarily satisfied the CAN3-A23.3-M66 (1966) code for reinforced concrete structures. Thereinforced concrete beam-to-column joint specimens were then taken from this officestructure, and their designs were modified slightly.The design of the test specimens was controlled by the available materials and thetesting equipment limitations. The existence of a set of reusable forms from a previousproject dictated the size of the specimens (Katzensteiner, 1992). The specimens wereaccordingly designed as part of a frame whose overall dimensions reflected the forms; this16resulted in an approximately half-scale beam-to-column joint model.3.2 SPECIMEN DESIGN CONSIDERATIONSIn deciding on the lengths of the beam and columns in the specimens it should benoted that the intention of the experiment was to observe the behaviour of the joint regionunder seismic loading by applying a cyclic load at the end of the beam. The columns of thetest specimens were to be cut off at the presumed inflection points, and in the testingapparatus would be replaced with hinges, resulting in zero moment at the ends of thecolumns. However, because of the existence of the testing apparatus from a previousexperiment (Kuan, 1992), it was decided to match the length of the columns to the availablespace in the test apparatus. Consequently, both the upper and the lower columns weremade to be of equal length, thus simplifying the testing process somewhat.A decision was made to alter the design of the joint from that specified in the 1966reinforced concrete code. All transverse reinforcement was omitted from the joint region.The main reason for taking this step was that it conformed to common practice due toambiguities in the interpretation of that edition of the code. Whereas the joint region shouldbe considered as part of the column, with specific transverse reinforcement requirements,this was never explicitly stated in that issue of the code, as opposed to the actual moderncode. Designers at that time typically did not extend the transverse column reinforcementinto the joint for a variety of reasons. Foremost among them, is ease of construction, sincetransverse reinforcement within a beam-to-column joint causes considerable difficulty inreinforcement placement; it also increases the possibility of the formation of honeycomb17voids due to the congestion of steel within the joint area. Joints with three or more framingbeams were considered "confined", due to the action of the beams, and joint reinforcementwas deemed unnecessary. The resulting design of the test specimens is outlined in AppendixA.3.3 UNRETROFM'ED TEST SPECIMENSOne of the specific aims of this research project was addressing the deficiencies inbeam-to-column joints caused by neglect, construction mistakes or misinterpretation of thecodes. The four identical test specimens (fig. 3-1) were fabricated to meet the programgoals. The positive moment capacity of the beam was designed for 16.27 kNm; to resist thismoment, four 10mm bars were needed at the ends of the beam. Overall, the steelrequirements were: two 10mm bars for both positive and negative flexure elsewhere in thebeams, and three 10mm bars for flexural reinforcement on either side of the columns;transverse reinforcement consisted of 10mm ties at 70mm spacing in both the beams and thecolumns.3.4 RETROFIT DESIGNIn the first phase of the experimental program, two specimens were loaded cyclicallysuch that the expected damage to the joint region occurred. One damaged specimen andone undamaged specimen were subsequently retrofitted with a circular steel jacket, while theother pair of specimens were retrofitted with square steel jackets. The dimensions of thejackets were similar in both cases.18Figure 3-1 Specimen Design19The steel jackets were intended to provide confinement and additional ductility to thejoints, without increasing the moment capacity of the specimens. Modern design codes fornew construction require an increased amount of transverse reinforcement near the jointregions to increase ductility. The steel jackets were to provide the desired ductility bysubstituting the reinforcement steel that was missing in the original design. Under moderndesign codes, the spacing of the transverse reinforcement should have been 35mm in boththe beams and the columns, provided the same 10mm bars were used in these elements.The same transverse reinforcement spacing would have been required in the joints, sincethey were not framed on all four sides by beams. To make up for the missing steel, thethickness of steel required for the joint governed, as there was no existing transverse steelwithin the joint. The result was a steel jacket with a thickness of 2.86 mm (0.110r). Thelengths of the retrofit were defined as being equal to the member depth (d) along thecolumn and twice the member depth (2*d) along the beam, measured from the beam-to-column interface. A gap was left halfway along the length of the beam jacket, to create aflexural hinge at that point.Preliminary tests of retrofitted beams (Ross, et al, 1992) show that the flexuralstrength of retrofitted beams can be increased approximately eight-fold. A retrofit thatconsists of an unbroken tube length of 2*d along the beam would force the flexural hingeto the end of the retrofit, and no gain in ductility would result. This inadequate retrofitwould merely increase the overall strength of the beam, without increasing its ductility;furthermore, it would also increase the risk of shear failure. The gap in the retrofit beamjacket was therefore introduced to assure that a flexural hinge would occur within the20retrofitted area. The joint area would still be confined by the steel tube, and ductility wouldbe added to the joint region. There would still be an increase in the overall momentresistance; in the present case this was expected to be about one-third.The size of the steel jackets was kept to a minimum to reduce the amount ofdisruption to the structure during the retrofit process, and also to limit the increase in thecapacity of the frame members beyond their original design values. An increase in themoment capacity of a section in one portion of the structure could force a failure elsewherein the structure by increasing the loading at that point beyond design values. Thedimensions for each of the retrofit strategies used in this study are summarized schematicallyin fig. 3-2a,b.An original aim of the retrofit portion of this study was to compare the retrofitbehaviour of both undamaged and previously damaged specimens. However, as the testingprogressed, modifications in the program were deemed necessary, which resulted indifferences in the retrofit details between the two sets of retrofitted specimens. Forexample, because of the state of the reinforced concrete of the original specimen, it wasimpractical to compare the results of the two specimens retrofitted with a circular jacket.With the square jacket retrofit scheme, premature failure of the beam outside of the retrofitrange forced an additional length of jacket to be added to the second beam specimen. Asimilar premature failure in the column of the circular retrofit scheme led to the additionof extra gaps in the beam jacket, in order to assure a reduction in the moment capacity ofthe retrofitted joint. These changes are detailed in the chapters dealing with the retrofittedspecimens.21Figure 3-2 Retrofit Concepts22CHAPTER 4EXPERIMENTAL APPARATUS AND PROCEDURE4.1 INTRODUCTIONThe cyclic loading of both the retrofitted and unretrofitted specimens was carried outin a similar manner. The unretrofitted specimens, however, were expected to behave poorlyand, as a result, to have a much lower ductility and to fail after only a few cycles. A largetesting frame in the Civil Engineering Structures Laboratory at UBC was used to apply thecyclic loading. The frame incorporated two separate actuators: one to supply the columnaxial load, and the other to provide the load at the end of the beam for cyclic loading. Dataacquisition from load cells, linear variable differential transformers (LVDT's) fordisplacement measurements, and numerous strain gauges was handled by an Optilog model200 data collection system.4.2 TESTING APPARATUSThe testing frame loading apparatus is shown in fig. 4-1. The cyclic load at the endof the beam was applied by a double-acting 100 kip (446 kN), 24" (610mm) strokeTemposonics Actuator, and the constant axial load for the column was applied by a 100 kip23Figure 4-1 Testing Frame and Specimen24(446 kN), 6" (150mm) stroke MTS Actuator. Both actuators were controlled by an MTSservo-control system. The column actuator was set to load control, so that it provided aconstant load during the testing sequence. The Temposonics actuator was operated underdisplacement control. It was much easier to control the cyclic load and displacement in thismanner.The applied loads were measured by 100 kip (446 kN) load cells that were locatedin line with the actuators. Deflections were detected with a set of LVDTs, and by thedisplacement transducer that is built into the Temposonics actuator. Strains on the jacketsof the retrofitted specimens were measured with the use of 350-ohm CEA strain gauges.All of the loads, displacements and strains were collected and stored into a computer by theOptilog data acquisition system.4.3 MEASUREMENTS FOR ROTATIONThe LVDTs were arranged as illustrated in fig. 4-2a,b. The actual locations at whichthe LVDTs were placed in the unretrofitted and retrofitted tests varied, due to the differingsizes of the members (fig. 4-2). The individual measured rotations were used to separateand identify the various types of movement that make up the total rotation of the beam.The LVDTs on the beam (#4 in fig. 4-2a, #4 and #5 in fig. 4-2b) measured the totalrotation of the beam. This rotation consists of the rotation caused by elastic deformationof the column, the shear deformation of the joint, and the rotation of the beam assuminga fixed joint (cantilever action). Each type of rotation was identified and analyzedseparately, using the data supplied by the remaining LVDTs. The various types of rotation25CYCLIC LOADLVDT #1LVDT #2 ED^<^LVDT #3 I LVDT #4CYCLIC LOADLVDT #1LVDT #2 ^LVDT #3 En^LVDT #5a) Unretrofitted Specimensb) Retrofitted SpecimensFigure 4-2 Placement of LVDTs26recorded are illustrated in Figure 4-3. The rotation caused by the shear deformation of thejoint was measured by LVDT #1, and it can be seen that the rigid joint rotation, which isa result of the elasic deformation of the column, is determined by subtracting the sheardeformation rotation from the rotation as measured by LVDTs #2 and #3. The cantileverrotation of the beam is then evaluated by subtracting the shear deformation and rigidrotation effects from the total deflection measured by LVDT #4.4.4 STRAIN GAUGESStrain gauges were used in the retrofitted tests to determine the behaviour of thesteel jackets during cyclic loading, and their layout is shown in fig. 4-4. This figure illustratesthe maximum number of gauges used corresponding to the tests of the retrofitted,undamaged specimens RETRO-SU and RETRO-CU. After reviewing the data from thesetwo tests, it was decided to omit some of the gauges when testing the retrofitted, previouslydamaged specimens RETRO-SD and RETRO-CD; these gauges are identified in fig. 4-4.The strain gauges were concentrated within the joint area and were intended to provide aninsight into the complicated stress pattern of the steel jacket in that area. Of particularinterest was the amount of longitudinal, circumferential and transverse stresses developedat the same time within the steel jacket.The intention of the steel jacket was solely to provide confinement for the concretecore. Ideally the steel jacket should only undergo circumferential or transverse stress.However, due to the complicated nature of the stresses within the joint due to confinementand concrete to steel bonding, it was anticipated that longitudinal stresses would be27a) Joint Shear Deformation^b) Rigid Joint Action...............................................c) Beam Cantilever Action d) Combined ActionFigure 4-3 Rotation Types Within the Joint28* Used in both RETRO-CU and RETRO-CD* Used in both RETRO -SU and RETRO -SDa) Specimens with Circular Retrofitb) Specimens with Square RetrofitFigure 4-4 Placement of Strain Gauges29developed, resulting in an increase in the moment capacity of the specimen. The placementof a jacket gap within the retrofit section of the beam was intended to mitigate some ofthese undesired effects.4.5 TESTING PROCEDUREThe test procedure for the unretrofitted tests was straightforward. The axial load wasfirst applied to the column, and this load was maintained throughout the duration of the test.A moment was then applied to the joint by loading the end of the beam, in an upward(positive) direction. The load was steadily increased until the observed yield moment wasreached. Due to the non-linear response of the specimen, the definition of the yieldmoment depended somewhat on judgment. An analog plot of the Temposonic actuator loadversus displacement was used to decide when the yield moment was reached. A ductilityfactor of 8 =1 was associated with this moment. This same moment was then applied in theopposite direction (negative), and then the full cycle was repeated a second time. Two morecomplete cycles were then applied, but at double the displacement. The cyclic loadingpattern of the unretrofitted joints is illustrated in fig. 4-5a. For the unretrofitted tests, fourcycles of loading caused enough distress to the joint that it could be considered damaged,but repairable. The unretrofitted tests were not intended to completely destroy the joints,since one of the original aims of this study was to compare the behaviour of damaged andundamaged joints when retrofitted with steel jackets.The loading pattern of the retrofitted joints was similar to that followed with theunretrofitted joints, as shown in fig. 4-5b,c. The retrofitted specimen RETRO-SU followed300-6CC08 0=2cE 0-1°-0 0-1P.- 0-20DO 0=40-60-80=40=80=6Q 0=40I-0 0-21 0 —1021- 0-1r-- 0=20D0 0=40-60-8a) Unretrofitted Specimens RCBC1, RCBC20-60=8b) Retrofitted Specimen RETRO-SU0=8 1c) Retrofitted Specimens RETRO-CU, RETRO-SD, RETRO-CDFigure 4-5 Testing Sequence31the pattern established for unretrofitted specimens RCBC1 and RCBC2, save that each cyclebegan with the load applied in the negative direction. For the retrofitted specimensRETRO-CU, RETRO-SD and RETRO-CD, after 0 =1, each subsequent cycle wasincremented by half of one full ductility factor. After 0 =4, each cycle was incremented byone full ductility factor. The joints were loaded until failure, or to the stroke limit of theactuator.The unretrofitted specimens RCBC1 and RCBC2 were tested first, and once theretrofit had been completed, the retrofitted specimens RETRO-SU, RETRO-CU, RETRO-SD and RETRO-CD were tested.32CHAPTER 5UNRETRO} I 1 1ED TEST RESULTS5.1 INTRODUCTIONThe tests on the unretrofitted reinforced concrete specimens yielded results consistentwith previous studies, i.e. lack of shear strength and ductility in the joint area. Thespecimens performed inadequately by today's standards, as they did not attain an acceptableductility ratio, nor did they even reach their theoretical elastic limits under certain conditions,as will be shown. Both RCBC1 and RCBC2 were tested under cyclic loading. Each loadingcycle began with the beam being pulled upwards in positive loading, such that the bottomlayer of flexural steel was stressed in tension. Each cycle ended with negative loadingproducing a deflection which was equal and opposite to the deflection imposed during thepositive loading, so that the top layer of steel was in tension. The deflections in the first andsecond cycles corresponded approximately to the yield deformation of the lower layer ofsteel. The deflection of each successive pair of cycles was increased to correspond to thenext whole number multiple of this initial yield deformation, until failure took place. Allrecorded data for the unretrofitted specimens are presented in Appendix C.335.2 OBSERVED BEHAVIOUR OF SPECIMENS DURING TESTThe hysteresis loops obtained for both RCBC1 and RCBC2 (fig. 5-1a,b) show thata rapid degradation in stiffness and strength occurred during the first four cycles. Thedisplacement during the fourth cycle corresponded to twice the yield displacement. Afterthe fourth cycle the damage within the joint was so extensive that only a fraction of themaximum strength and stiffness still remained, and failure of the specimen had obviouslytaken place. The strength loss during the final cycle was of the order of 60%. Thehysteresis loop exhibited the characteristic pinching effect that is prevalent in reinforcedconcrete members tested under cyclic loading.There was absolutely no indication of a failure of any type having occurred withineither the beam or the column. Flexure cracks formed within the column and the beam atregular intervals; these cracks appeared as expected according to the design. There wasplenty of evidence, however, to suggest that all the major damage had occurred within thejoint region. As shown in the photographs (fig. 5-2a,b), both specimens developed a largeamount of shear cracking within the joint region. These cracks opened and closed insynchronization with the cyclic load that was applied to the end of the beam. After thedisplacement passed the yield point during negative loading, the shear cracks within the jointno longer closed completely. A large amount of the concrete cover spalled off the rear faceof the column within the joint area. Some concrete crushing also took place within the hingearea at the interface of the column and the beam.At best, both specimens were only able to develop a rotational ductility of 0=2. Anoticeable reduction in both strength and stiffness took place on the negative loading portion34Figure 5-1 Hysteresis Loops: Unretrofitted Specimens35U BCJUN 3SPECIMEN RCBC SYIELD X 2 NEG 2a) RCBC1 b) RCBC2Figure 5-2 Joint Damage in Unretrofitted Specimens36of the 3rd cycle (the first at 0=2), and this trend continued into the 4th cycle, where asignificant strength and stiffness loss took place. Failure of both specimens was declared inthe fourth cycle at a rotational ductility of 0=2. This indicated that at this point the jointcapacity had fallen to less than 80% of its original strength.5.3 FAILURE MECHANISMS WITHIN THE JOINT AREAThe failure mechanism within the joint area was complex. At the point when failureof the specimen was considered to have taken place, the size of the cracks in the joint werevisibly large (fig. 5-2), particularly during the negative load cycle. The top layer oflongitudinal reinforcement consisted of 4 bars in the part of the beam adjoining the column,as opposed to only 2 bars in the bottom layer. The larger amount of steel in the top layerprevented hinging during the positive cycle, which in turn pushed the damage into the jointregion. However, the reinforcement in the bottom layer of steel actually attained the yieldstress, and a plastic hinge formed in the beam near the interface of the beam and thecolumn (fig. 5-3a). During the negative portion of the cycle, a different type of failureoccurred, as the stress in the top layer of steel never reached yield stress. The attainmentof maximum moment was limited by the failure of the bond between the longitudinalreinforcement bars and the surrounding concrete in the joint area, as well as by the largeamount of straining in the cracked concrete within the joint area (fig. 5-3b). Evidence ofbond failure was indicated by the spalling of the cover concrete, particularly on the rear faceof the joint along the column (fig. 5-3c). This suggests that the hooked ends of thelongitudinal reinforcement moved outwards to split the cover and lose its bond with the37a) Hinge in Jointb) Shear Cracking in Joint AreaFigure 5-3 Failure of Unretrofitted Specimens38c) Spalling on Rear Face of ColumnFigure 5-3 Failure of Unretrofitted Specimens39interior concrete.On removal of the concrete rubble after the tests, a large amount of loose concretewas found in zones adjacent to the longitudinal bars, particularly those of the top layer ofsteel. After the bond between the top layer of steel and the core concrete was lost, the toplayer of longitudinal steel in the beam was free to move back and forth over the concretecore, resulting in further damage. The beam effectively pivoted about a point near the lowerlayer of steel, which had not lost its bond. It is interesting to note that the concrete core ofthe interior of the joint was virtually free of visible cracks.5.4 MATERIAL PROPERTIESThe steel used in the testing program was obtained from the same batch ofreinforcing bars. A number of tensile coupon tests were carried out and indicated that thetensile yield strength of the steel was approximately 566 MPa (appendix B); the ultimatestrength of the steel was approximately 800 MPa. The stress-strain curves did not featurea marked yield plateau and the standard 0.2% offset strain was therefore used to define thetensile yield stress of the reinforcement steel. It should be noted, however, that thisapproach only provides a nominal yield value for use in design; it does not define a yieldpoint in the true sense.Due to the lack of a defineable yield plateau, it was difficult to pinpoint exactly theonset of yielding during the cyclic loading tests. As a result, a certain amount of judgmentwas used during the tests to decide when this "yield point" was reached. Clearly, this led toopportunities for discrepencies in determining the commencement of yield for RCBC1 and40RCBC2, although every attempt was made to be consistent in assigning their respective yieldpoints.The compressive strength of the concrete was determined from uniaxial tests ofstandard 30 cm. cylinders and was found to be 26.3 MPa (Appendix B). All the test materialstrengths lie within the commonly used range of values.5.5 COMPARISON WITH MODERN SEISMIC DESIGNWhen comparing the hysteresis loops of the two unretrofitted specimens, RCBC1 andRCBC2, with loops obtained from specimens that are designed according to currentstandards, the shortcomings of the two tested specimens become evident. Figs. 5-1a,b showthe hysteresis curves for RCBC1 and RCBC2, whereas fig. 5-4 shows a loop obtained froma specimen of similar configuration whose reinforcement design conformed to currentseismic standards (Filiatrault, 1992). The specimen used to produce the hysteresis loopshown in fig. 5-4 was full size, and designed according to the most recent Canadian Standard,which included proper detailing for ductility within the beam-to-column joint area. NeitherRCBC1 nor RCBC2 attained a rotational ductility greater than 0=2, whereas the specimentested by Filiatrault attained a ductility ratio of about 0=3 without loss of strength.5.6 MEASURED DATA: RCBC1 AND RCBC2Instrumentation was positioned to capture the major expected deformations of thespecimens. As noted in section 4.3, LVDT's were placed to determine the contributions ofeach of the deformation types to the total rotation of the joint within the specimen.41Figure 5-4 Hysteresis Loop of a Specimen of Modern Design42As a general rule, the deformation of the column, which produced the "rigid" rotationof the joint area, made the smallest contribution to the total rotation (figs. 5-5a,b), and wasprimarily dependent on the bending stiffness of the columns. The columns were designedto withstand much higher axial loads and moments than those applied. The axial loadsapplied in the tests represented the dead load of the proposed structure and a certainproportion of the live load, and were relatively low. The design of the columns was alsogoverned by the prescribed minimum steel ratio, and these members were thus able towithstand a moment much higher than the beam yield moment (figs. 5-5a,b). The sheardeformation of the joint area initially contributed only a small percentage of the totaldeformation, but once cracking was initiated within the joint the contribution of the sheardeformation increased significantly. During negative loading, when the top layer of the beamlongitudinal steel was in tension, the shear deformation component of the total rotation wasthe greatest. During positive loading, the shear contribution was small, as only a few tensioncracks developed within the joint. The greatest deformation component during positiveloading was due to the elastic cantilever deflection of the beam, and eventually the rotationof the plastic hinge that developed at the interface of the beam and column.During negative loading, no flexural plastic hinge formed either within the beam orthe joint. The force that was applied to the free end of the beam, and thus the moment atthe beam-to-column interface, never attained the theoretical capacity of the reinforcementin the top layer of the beam. However, there was still a large amount of deflection that wasnot caused by shear deformation of the concrete, elastic deformation of the column, or byhinge rotation at the interface of the beam with the column. This seems to indicate that a43150100CANTILEVER0.10.050-0.05-0.10^ 50SCAN NUMBERTOTAL^SHEAR   RIGIDCOMPARISON OF ROTATION COMPONENTSRCBC150^ 100^ 150SCAN NUMBERTOTAL^SHEARa) RCBC1COMPARISON OF ROTATION COMPONENTSRCBC2^ RIGID^CANTILEVERb) RCBC2Figure 5-5 Rotation Components: Unretrofitted Specimens44bond failure had taken place between the longitudinal steel of the beam that was anchoredwithin the joint and the concrete core. Since the top layer of steel was stressed in tension,once bond failure occurred these bars were prevented from being completely pulled outfrom the concrete by the resistance provided by the bearing of the reinforcing bar anchorson the core concrete. The post-testing evidence suggests, however, that the bars werecrushing the concrete at the bearing areas, as well as sliding back and forth, thus damagingthese bearing areas further, and allowing more movement to take place. Due to thearrangements of the transducers, this deformation shows up in the data partly as sheardeformation.The respective strengths and stiffnesses of RCBC1 and RCBC2 were comparable, andboth specimens behaved in a similar manner, although it appears that RCBC2 wasmarginally weaker than RCBC1, most noticeably during negative loading.5.7 LIMITATIONS OF THE MEASURING DEVICESThe placement of the LVDT's, and the methods used to support these measuringdevices, proved not to be entirely reliable for securing exact measurements of the variousdeformation modes. The LVDT that measured the shear deformation was attached to thelower column by a bracing system located immediately below the joint area. Two problemswith this system tended to skew the measured shear deflection of the joint area accordingto the direction of loading. Firstly, the area in the column immediately below the joint,where this brace was situated, was a highly disturbed area, in that this region was greatlyinfluenced by what occurred in the joint during cyclic testing. During negative loading, when45a large number of cracks appeared in the joint area, some of the stress was transferred tothe bordering areas in the column and the beam. The resulting deformation caused thetimber brace holding the LVDT to crush and rotate somewhat, resulting in inflated sheardeflection readings. The second problem involved the spalling of the cover concrete at therear face of the column. As the cover concrete spalled off the column, the large crack thatoccurred between the core concrete region and the cover concrete caused the sheardeflection LVDT to record a deflection that was larger than the actual value during positiveloading and smaller than the actual value during negative loading. This same spallingproblem affected the pair of LVDT's that measured the combination of shear deformationand column rotation; however, the error in the individual instruments of the pair was notquite the same, as the magnitude of the error in each was dependent upon the pattern ofspalling and cracking of the cover concrete.These difficulties in measurement also affected the determination of each of the othertypes of rotation. The total measured rotation was a relatively reliable measurement,however, as it was measured from a point that was well removed from the disturbed regions.It is important to note that the individual contributions by shear deformation, columndeflection and beam deflection are only reasonably reliable up to the approximate yieldpoint of the joint. After yielding, these deformations only provide an indication of the jointbehaviour to the failure mode, and are thus useful in determining the mechanism of failure.5.8 THEORETICAL VALUESThe strength values experimentally obtained for the unretrofitted specimens RCBC146and RCBC2 conformed very well with those that were predicted based on standard sectionanalysis of a reinforced concrete specimen (Table 5.1). Joint strength during negativeloading was reduced because of the disturbed state of the joint region after attaining yieldduring positive loading. The bending moment (M b) required to cause yielding at the beam-to-column interface was based on section analysis. The equivalent theoretical yieldingmoment at the center of the beam-to-column joint (My) and the actual experimental momentapplied (Mapp) are all moments measured at the center of the joint. The contributions tothe total rotation by the elastic deformation of the column and the hinge rotation at thebeam-to-column interface are reasonable. The only components of the total rotation thatare unaccounted for by the theory are the rotations caused by shear deformation within thejoint and also by the slippage of the longitudinal bars, which was greatly influenced by thatsame shear deformation.Table 5-1: RCBC1 and RCBC2 Section AnalysesSpecimen Mb (kNm) My (kNm) Mapp (kNm)RCBC1Positive Loading 16.8 18.8 20.0Negative Loading 32.4 36.7 24.8RCBC2Positive Loading 16.8 18.8 20.2Negative Loading 32.4 36.7 22.047It is not necessarily important to know how much of the rotation was actually causedby the shear and by damage due to loss of bond within the joint; it is sufficient to know thatsuch effects did occur. The important concern is to find an effective method of eliminating,or at the very least, reducing, the outcome of these occurrences.5.9 EVALUATION OF SPECIMEN PERFORMANCEConsidering the real shortcomings in the design of the older buildings under study,and hence in their representative test specimens, the outcome of the tests was as expected.The lack of detailing for ductility in and around the beam-to-column joint resulted in poorbehaviour during cyclic loading, which was characterized by the development of a lowrotational ductility of 0 =2. In addition, the lack of confining ties within the joint arearesulted in poor shear behaviour. The deficiency in shear strength within the joint alsoresulted in the inability of the specimen to develop the full elastic tensile capacity of the toplayer of steel. This could have serious consequences during a seismic event, particularly inlight of the poor ductility of the structure. Ideally, the best possible behaviour for thisspecimen would have occurred if the beam would have had the ability to form a plastic hingeduring both negative and positive loading; this undoubtedly would have increased theintegrity of the joint.48CHAPTER 6RETROFITTED TEST RESULTS6.1 INTRODUCTIONA total of four retrofitted specimens were tested. Two of the specimens wereretrofitted with a circular steel jacket, and two were retrofitted with a square steel jacket.In each pair, one of the specimens was previously damaged, and the other was undamaged.The specimens were labelled as follows (in order of testing):RETRO-SU: Square, undamaged specimenRETRO-CU: Circular, undamaged specimenRETRO-SD: Square, previously damaged specimen (RCBC2)RETRO-CD: Circular, previously damaged specimen (RCBC1)The original objective was to have identical designs for each of the two retrofit shapes, inorder to compare the different behaviour of the undamaged and previously damagedspecimens. As testing proceeded, however, it became apparent that some undesirableeffects changed the outcome of the tests, and some modifications were subsequently made.49The modifications are discussed in the appropriate sub-section. With these modificationsto the jackets, it was no longer possible to make direct comparisons based on the previouscondition of the original specimen; however, other comparisons can be readily made basedon geometry, spacing of gaps within the retrofit, and modes of failure.The steel jackets had a yield strength of 267 MPa (Appendix B), despite the fact that400 MPa strength steel had been specified. The characteristic compressive strength of theoriginal concrete was found to be f c'= 30.3 MPa, and the strength of the grout used for theretrofit was f:= 31.3 Mpa. All recorded data for each of the retrofitted specimens arepresented in Appendix D.6.2 RETRO-SU: UNDAMAGED SPECIMEN WITH SQUARE JACKETRETRO-SU (fig. 6-la) developed two modes of failure, depending on the directionof the loading. With positive (upward) loading, a flexural hinge formed at the gap betweenthe two sections of steel jacket in the beam (fig. 6-1b). With negative (downward) loading,a flexural hinge formed at the end of the retrofit area of the beam (fig. 6-1c). Withrepeated cycling this flexural hinge progressed into a shear failure of the specimen at thispoint (fig. 6-1d). The flexural hinge that had been developing on the bottom face of the gaparea stopped expanding once the shear cracks on the upper face became very large. Thefailure of the specimen outside of the retrofit area was considered to be undesirable in theextreme. A subsequent section analysis of the specimen showed that the moment capacityduring negative loading at the gap was large enough to deflect failure to the end of theretrofit.50a) Before Test^b) Hinge Formation During Positive LoadingFigure 6-1 Retrofitted Specimen RETRO-SU51c) Hinge Formation During Negative Loading^ d) FailureFigure 6-1 Retrofitted Specimen RETRO-SU52Not much strength loss was recorded during the progression of the experiment (fig.6-2a,b), although there was some reduction in stiffness. Some pinching occurred during bothpositive and negative loading, but it was much less than that observed in the unretrofittedspecimens.There was not much shear deformation within the joint area (fig. 6-2d). About 5%of the total rotation of the joint resulted from shear deformation within the joint region,primarily during upward loading when the top layer of steel was in compression. The rigidrotation of the joint that was caused by column deformation tended to be the highestcomponent of the total joint rotation, due to the increased stiffness of the joint area. Thecantilever deformation of the beam area tended to be significant, but was limited oncehinging began taking place either in the gap or outside the beam retrofit region. Both therigid rotation and the cantilever rotation within the joint were elastic. The hysteresis loopsof the gap (fig. 6-2c) show the progressive development of the plastic hinge at the top of thegap during positive loading. Since hinging during negative loading was outside the gap, thetop layer of steel in the gap never reached yield stress, and hence little or no plastic actionwas observed for this portion of the loading cycle.During the test of this specimen, a number of "pops" were heard as the bond betweenportions of the concrete and the steel jacket was lost and the jacket buckled outwards. Thiswas not unexpected, since square tubing has a tendency to buckle under compression. Thequestion of load transfer through bond became quite prevalent in this study. If no bondexisted between the concrete and the steel jacket, flexural failure of the specimen shouldhave occured in the joint area, as in the unretrofitted specimens, since this was the location53MOMENT VS. TOTAL OUTER ROTATIONFigure 6-2 Rotation: RETRO-SU546040200-20-40-60-0.1 -0.05 0HINGE ROTATION (RAD)0.05 0.1MOMENT VS. RtITIRIALGE ROTATIONc) Hinge at Retrofit GapCOMPARISON OF ROTATION COMPONENTSFigure 6-2 Rotation: RETRO-SU55with highest applied moment. However, the flexural failures did not occur near the joint;in both cases they were located at points where no retrofit steel was present, indicating thatthe steel jacket significantly affects the moment capacity of the beam, and likely the columnas well. Other load transfer mechanisms may also have played an important role. It isnoted that the confinement properties of square (or rectangular) steel tubing are not nearlyas effective as those of circular steel tubing. This is due to the fact that the faces of thesquare steel tube under compression tend to lose their bond with the concrete interior andhence confinement by physical enclosure is lost.At no point during the test were plastic deformations of the jacket evident (fig. 6-3)since all recorded strains in the jacket remained within the elastic limit (less than e y 1335microstrain). The only visible signs of plastic deformation only occurred long after failurehad already taken place in the specimen, when some bending of the jacket surrounding theretrofit gap was observed. The strains that were recorded by the gauges did not produceany major surprises; the strain distribution reasonably reflected expected patterns. Therelatively low strains observed in some areas was due to a loss of bond between the concreteand steel. Loss of bond was generally indicated by an abrupt change in the strain readings(fig. 6-3a). This was accompanied by an audible "pop" when the bond was lost.Most of the strain in the steel casing took place in the beam retrofit region borderingthe retrofit gap, or at the end of the beam retrofit. The column and joint area strain gaugesgenerally behaved elastically, suggesting no apparent damage of the internal concrete inthese locations. There were also many indications of bond failure in these areas. The straingauges on the beam jackets showed the existence of a net circumferential tensile strain in56z*.W2020W3aa4Figure 6-3 Strains: RETRO-SU57400 600604020E 00a -20-40MOMENT VS. STRAINRETRO-SU: TOP BEAM TRANSVERSE (#008)-60-300^-200^-100^0^100STRAIN (MICROSTRAIN)200 300c) Top Beam Transverse (TBT)MOMENT VS. STRAINRETRO-SU: TOP EXTENSION TRANSVERSE (#010)6040.g 20zhe02Oa.a. -20-40d) Top Extension Transverse (TET)Figure 6-3 Strains: RETRO-SU58-60-600^-400^-200^0^200STRAIN (MICROSTRAIN)the steel on both sides of the gap (fig. 6-3b,d), and a net longitudinal compressive strain inthe steel jacket between the gap and the joint (fig. 6-3c). The tensile strain in the radialdirection was due to the lateral expansion of the concrete under compression; thecompressive longitudinal strain was due both to this same lateral expansion and to thedevelopment of a friction bond between the concrete and steel. There was also someevidence of a permanent tensile strain in the jacket at the side of the joint, parallel to thecolumn and perpendicular to the beam.At one point during the test the entire specimen was lifted off the lower hinge. Thiswas due to the fact that the downward load provided by the column actuator under loadcontrol was insufficient to hold the specimen in place. This load was subsequently increasedto avoid a repetition of this condition.Under positive loading, yielding occurred in the retrofit gap at an applied moment(MaPP ) of 27.4 kNm. Under negative loading, the yielding occurred outside the retrofit areaat the end of the beam, at an applied moment of 42.7 kNm. As indicated in Table 6.1, thelocation of failure and the value of the yield moment (My) can be reasonably accuratelypredicted by analysis. The bending moment (M b) required to cause yielding at a section wasbased on a section analysis. The equivalent theoretical yielding moment at the center of thebeam-to-column joint (My) and the actual experimental moment applied (Mapp) are allmoments measured at the center of the joint. The Mapp values are the yield moments takenfrom the hysteresis loop.59Table 6-1: RETRO-SU Section AnalysesFailure Location Mb (kNm) My (kNm) Map p (kNm)Positive LoadingBeam - End of Retrofit 16.8 46.5 ----Beam - Gap Region 18.7 29.9 27.4Column 22.9 59.0 ----Negative LoadingBeam - End of Retrofit 16.8 46.5 42.7Beam - Gap Region 35.8 57.3 ----Column 22.9 59.0 ----63 RETRO-CU: UNDAMAGED SPECIMEN WITH CIRCULAR JACKETSpecimen RETRO-CU (fig. 6-4a) also developed two modes of failure, depending onthe direction of loading. With positive loading, a flexural hinge developed at the gapbetween the two sections of the steel jacket surrounding the beam (fig. 6-4b). With negativeloading, a flexural hinge once again formed outside the retrofitted region, but in this caseits location was at the end of the retrofit on the upper portion of the column (fig. 6-4c). Inorder to avoid a catastrophic buckling failure of the column, negative loading past a ductilityfactor of 0=1.5 was avoided, since it was deemed more desirable to observe the effects ofthe development of the plastic hinge within the retrofitted zone. Had the experimentcontinued according to the originally planned cycle of loading in both the positive and60a) Before Test^b) Hinge Formation During Positive LoadingFigure 6-4 Retrofitted Specimen RETRO-CU61d) Failurec) Column Hinge Formation During Negative LoadingFigure 6-4 Retrofitted Specimen RETRO-CU62negative directions, it is quite possible that final failure of the specimen would have resultedfrom the formation of a hinge within the column. As it happened, the failure of thespecimen was caused by the tensile rupture of the lower layer of beam reinforcing steelunder positive loading (fig. 6-4d). It is apparent that the yield moment was reached in thefirst cycle of testing, resulting in a ductility factor of about 0=2 (fig. 6-5). The failure of thisspecimen occurred during the 4th cycle, corresponding to a ductility factor of about 0 =5.The flexural hinge was concentrated over a very short distance, due to the confinement ofthe core concrete. The column load was increased on two occasions during the testing, toavoid the specimen lifting off of the lower support.The hinge that developed in the column was cause for concern. While a failureoutside the retrofit area is undesirable, it would be potentially catastrophic to have thatfailure develop in the column rather than the beam. The somewhat longer sleeve of thecircular jacket, compared to that of the square jacket, prevented a flexural hinge fromforming at the end of the beam retrofit. The added moment capacity of the section, bothdue to the placement of the steel jacket and the added concrete area, caused the failure tooccur in the column rather than the beam.During positive loading, when the flexural hinge developed within the retrofit gap,there was no loss of strength of the specimen (fig. 6-5a,b). However, a reduction in thestrength of the specimen occurred during negative loading due to P-6 effects. Fig. 6-5indicates that some reduction in the stiffness of the section and some hysteresis looppinching took place during the test, most noticeably during the negative loading, when failureoccurred within the unretrofitted section of the column.63Figure 6-5 Rotation: RETRO-CU64AIIMIL M. IfFigure 6-5 Rotation: RETRO-CU65There was not much deformation within the joint area. Shear deformation again onlyaccounted for a small portion (5%) of the total rotation of the joint (fig. 6-5d).Inconsistencies were detected between the measured and observed behaviour of the rigidand cantilever rotations of the joint area. Although visual observations indicated quiteclearly that a substantial rotation of the entire joint occurred during both negative andpositive loading, this is not evident from the LVDT data. The discrepency stems from theamount of deflection of the entire column, and hence the lateral movement of the joint area,during testing. The LVDTs, which are limited to an effective range to about 1" in eitherdirection, were not able to reliably record the data once the columns deflected more than1" laterally. Beyond this limit the LVDT measuring the lateral deflection of the top of thejoint lost contact with the specimen. Thus, these readings are highly skewed for rigidrotation in the positive direction and cantilever rotation in the negative direction, and arenot useful past the point of initial yielding.The hinge hysteresis loop (fig. 6-5c) shows the progression of the plastic hinge withinthe bottom layer of steel during positive loading. The top layer of steel apparently neverreached the yield stress under negative loading, since the flexural hinge under negativeloading formed within the column.There was no audible indication and very little visual evidence of any bond failurebetween the concrete and the steel jacket. Only a small portion of concrete around the gaplost its bond with the steel jacket. This was due to the propagation of the flexural crack inthe region where failure eventually took place. There was no reason to believe that ageneral bond failure took place over a substantial portion of the retrofit area. Some spalling66of the surface concrete was observed in the gap area.No portion of the jacket was strained beyond yield (1335 microstrain) during the test(fig. 6-6). Most of the measured strains within the joint area were in the 400 microstrainrange. Again, the strains (and hence the stress patterns) recorded by the strain gauges didnot indicate anything unusual. Observed tensile transverse strains in the retrofit area wereas expected and tended to be of a higher magnitude than those observed during testing ofRETRO-SU. Tensile transverse strains were monitored in RCT, RJT, SJL (fig. 6-6b) andTET (fig. 6-6d) during negative loading, and RJT, SJT, FCT (fig. 6-6a), TBT (fig. 6-6c) andTET (fig. 6-6d) during positive loading. Permanent strains occurred primarily in the beamextension retrofit (the retrofit beyond the gap), and to a lesser extent in the retrofit areabetween the joint and the gap. This was likely due to the damage of the concrete core,because of its proximity to the flexural hinge in the gap area.Under positive loading, yielding was initiated in the gap region in the beam at anapplied moment of 39.6 kNm. This value is somewhat higher than might be expected forthis section, because both the top and bottom layers of reinforcing steel were in the tensionzone. It was found that the flexural capacity of this specimen under negative loading wasgoverned by the flexural capacity of the column (56.4 kNm) at the end of the retrofittedarea. As indicated in table 6.2, the location of failure and the value of the yield moment canbe predicted reasonably accurately by analysis.67MOMENT VS. STRAINRETRO-CU: FRONT COLUMN TRANSVERSE (#007)6040200-20-40600400-60-600^-400^-200^0^200STRAIN (MICROSTRAIN)a) Front Column Transverse (FCT)Figure 6-6 Strains: RETRO-CU68c) Top Beam Transverse (TBT)Figure 6-6 Strains: RETRO-CU69Table 6-2: RETRO-CU Section AnalysesFailure Location Mb (kNm) My (kNm) Mapp (kNm)Positive LoadingBeam - End of Retrofit 16.8 75.6 ----Beam - Gap Region 21.4 37.6 39.6Column 22.9 63.2 ----Negative LoadingBeam - End of Retrofit 16.8 75.6 ----Beam - Gap Region 38.9 68.3 ----Column 22.9 63.2 56.46.4 RETRO -SD: PREVIOUSLY DAMAGED SPECIMEN WITH SQUARE JACKETBased on the results from RETRO-SU, it was decided to add some length to the steeljacket along the beam. To achieve this, the extension section of the sleeve was removedfrom specimen RETRO-SU, and welded onto the extension section of RETRO-SD. Theentire length of the beam of specimen RETRO-SD was thus effectively wrapped in the steeljacket (fig. 6-7a). The critical section for both loading cases was now expected to be withinthe gap region. This specimen actually developed the desired mode of failure under loadingin both directions. That is, the flexural hinge formed within the gap region under bothpositive and negative loading (fig. 6-7b,c).Failure of the specimen was caused by flexural yielding and tensile rupture of the70UBCSEP 1SPECIMEN RETRO CUCYCLEX YIELD 1UB CSEP 1SPECIMEN RETRO SDCYCLE 8X YIELD Sa) Before Test b) Hinge Formation During Positive Loading Figure 6-7 Retrofitted Specimen RETRO-SD71c) Hinge Formation During Negative LoadingFigure 6-7 Retrofitted Specimen RETRO-SD72d) Failurebottom layer of steel (fig. 6-7d), but only after a very large number of cycles. A maximumductility factor of 0=6 was reached during the ninth cycle. Failure took place on thepositive loading portion of the 10th cycle; the steel failed before a ratio of 0 =6 could bereattained in this last cycle (fig. 6-8a,b). Only a marginal loss of strength was recordedduring cycles four to seven, but that strength was regained during cycles eight and nine,probably due to strain hardening in the reinforcement steel. In this specimen, hystereticloop pinching only occurred during negative loading, and it was clearly smaller than thepinching noted in both RETRO-SU and RETRO-CU. This can be explained by the factthat the bending stiffness under negative loading was mainly due to the reinforcing steel(tension and compression), whereas the concrete in compression contributed significantly tothe stiffness under positive loading. As the flexural cracks closed, the stiffness duringpositive loading did not increase until the cracked concrete surfaces came into contact again.The shear deformation within the joint was approximately constant and of the orderof 7% of the total joint rotation (fig. 6-8d). The rigid rotation caused by the columndeformation, and the cantilever beam rotation also remained relatively constant over theduration of the test. There were some fluctuations in the rotation of the joint, but these areattributable to the loss of stiffness in the section within the gap area, where yielding wastaking place. Because of the stiffening effect of the jacket, the rigid rotation of the jointarea represented the largest component of total rotation. The cyclic motions within the jointarea took place about the origin and remained elastic during the testing sequence.The hysteresis loops (fig. 6-8a,c) that illustrate the development of the plastic hingein the gap region indicate a stable increase in the ductility ratio. All the plastic motion73-0.1^-0.05^0^0.05^01TOTAL OUTER ROTATION (RAD)a) Outer Beam Locationb) Inner Beam LocationFigure 6-8 Rotation: RETRO-SD74Figure 6-8 Rotation: RETRO-SD75occurred within the gap region of the retrofit.As was the case with specimen RETRO-SU, a number of "pops" could be heardduring the testing cycle. Most of these noises occurred during the first two cycles, indicatinga loss of bond between the concrete and steel jacket over certain regions of the retrofit.There was some visual evidence of plastic deformation of the steel jacket in the areabordering the gap of the retrofit steel on the beam. Strain gauges were not located closeenough to the ends of the retrofit to record these plastic deformations. The deformationswere caused by the compressive crushing and subsequent lateral expansion of the concretein this area, and were most noticeable along the lower surface of the jacket duringcompression under negative loading. The deformations were characterized by an outwardbulging of the middle of the jacket; the jacket corners tended to remain square.Stress and strain patterns tended to follow the usual pattern, commensurate with thedirection of loading. Good examples of indicators pointing to the loss of the concrete-steelbond can be seen in the moment versus strain graphs of FCT, SJL and RJL (fig. 6-9a).These indicators are the sudden change in strain with a small change in applied moment.Permanent deformations are visible in the beam and beam extension plots (fig. 6-9b,c,d):tensile in the transverse direction, and compressive in the longitudinal direction. Thesepermanent deformations were located in sections bordering the area where extensivecracking occurred. Some strain gauges did not indicate any significant strains, which is nottoo surprising considering the remoteness of some of these areas from the plastic hinges.Under positive loading, the flexural hinge began to form at the gap in the retrofit atan applied moment of 29.0 kNm. This yield moment was predicted using section analysis.76MOMENT VS. STRAIN6040• 20z2 002o.▪ -20-40-60RETRO-SD: TOP BEAM LONGITUDINAL (#004)604052. 20z2 000aa.•-20-40-60MOMENT VS. STRAINRETRO-SD: REAR JOINT LONGITUDINAL (#000)-300^-200^-100^0^100^200^300STRAIN (MICROSTRAIN)a) Rear Joint Longitudinal (RJL)-1500^-1000^-500^0^500^1000^1500STRAIN (MICROSTRAIN1b) Top Beam Longitudinal (TBL)Figure 6-9 Strains: RETRO-SD77Ma •• /M.^i■, EL Mi^MEI ANL^IE. M••••■• Mt • M. •6040S. 20zE0020W3aa -20ct-40-60MOMENT VS. STRAINRETRO-SD: TOP EXTENSION TRANSVERSE (#003)-600^-400^-200^0^200^400^600STRAIN (MICROSTRAIN)d) Top Extension Transverse (TET)Figure 6-9 Strains: RETRO-SD78Under negative loading, the flexural hinge began to form in the gap at an applied momentof 44.3 kNm, which is substantially below the value of 57.3 kNm predicted by sectionanalysis. This lack of correlation can be attributed to the inability of the short, longitudinalreinforcement bars in the top layer to develop their yield strength due to an insufficientdevelopment length. Table 6.3 shows that the locations of failure were as expected, eventhough the exact values of the yield moments were not accurately predicted in both cases.Table 6-3: RETRO-SD Section AnalysesFailure Location Mb (kNm) My (kNm) Mapp (kNm)Positive LoadingBeam - End of Retrofit 16.8 151.2 ----Beam - Gap Region 18.7 29.9 29.0Column 20.7 59.0 ----Negative LoadingBeam - End of Retrofit 16.8 151.2 ----Beam - Gap Region 35.8 57.3 44.3Column 20.7 59.0 ----6.5 RETRO-CD: PREVIOUSLY DAMAGED SPECIMEN WITH CIRCULAR JACKETThe testing of specimen RETRO-CU showed that a modification was necessary toavoid development of a flexural hinge in the column. To rectify the situation, an extra pair79a) Before Testb) Hinge Formation During Positive LoadingFigure 6-10 Retrofitted Specimen RETRO-CD80c) Hinge Formation During Negative Loadingd) FailureFigure 6-10 Retrofitted Specimen RETRO-CD81of equidistant gaps were cut out of the retrofit steel on the beam between the originallydesigned gap and the joint (fig. 6-10a). The object of this alteration was to reduce theapplied moment necessary to cause failure in the beam, and hence move the critical sectionfor negative loading from the column into the beam area. Original concerns about theability of the retrofitted joint area to withstand excessive cracking and deformation duringtesting were shown to be unfounded during the earlier tests. This specimen's failure modewas exactly as had been expected and designed for - the development of a plastic hingewithin the gap region of the retrofit under both positive and negative loading (fig. 6-10b,c).Failure of the specimen was caused by flexural yielding and subsequent tensilerupture of the bottom layer of reinforcing steel during positive loading (fig. 6-10d). Theextra gaps helped to spread out the yielding length of the reinforcing bars, while stillproviding enough confinement for the concrete that tensile failure of the bars could beassured. The bottom layer of steel failed in tension during the positive loading portion ofthe 11th cycle (fig. 6-11). A ductility ratio of 0=7 was attained during the 10th cycle, andthe 11th cycle was undertaken to see whether a ductility ratio of 0=8 could be secured; thiswas not achieved. The yielding moment was overshot in the first cycle, and was adjustedfrom that point onward (fig. 6-11a). A small amount of strength loss was observed duringpositive loading, but virtually none during negative loading. A small gain in the strengthoccurred in the two cycles before failure, due to strain hardening of the reinforcement bars.No significant stiffness loss took place, and the pinching behaviour exhibited by thisspecimen was the most favourable observed in all the specimens. The hysteresis loops wereof a rounded shape, without the characteristic pinching effects for positive loading. For82MOMENT VS. TOTI ALIDNNER ROTATION6040-40-60-0.1^-0.05^0TOTAL INNER ROTATION (RAD)0.05 01b) Inner Beam LocationFigure 6-11 Rotation: RETRO-CD836040•• 201-z2 0020a- 20e_ --40-60-0.1 -0.05 0HINGE ROTATION (RAD)0.05 0.1MOMENT VS. HINGE ROTATIONRETRO-CDc) Hinge at Outermost Retrofit GapFigure 6-11 Rotation: RETRO-CD84negative loading there was some evidence of pinching, which could be attributed to the lossof bond in the shorter reinforcement bars of the top layer. Most of the bending strain afteryielding was concentrated in the first gap, with some deformation in the other two gaps.The shear deformation contribution to the total rotation was virtually unnoticeable,amounting to less than 1% (fig. 6-11d). The rigid rotation caused by column deformationwas constant as expected, since this was due to the elastic deflection of the column. Owingto the addition of the two extra gaps, the total rotation of the joint area included the primaryarea of the plastic hinge, and hence the plastic deformation of the first and second gaps.As a result, the part of the total rotation that was caused by cantilever action was extremelyhigh. Virtually all of this deformation took place at the gap, rather than in the joint areaitself. With such a short distance between the column and the first gap, the actualcontribution of elastic cantilever action should have been small. The graph of momentversus total inner rotation (fig. 6-11b) includes the hinging which took place in the first twogaps, whereas the graph of moment versus hinge rotation (fig. 6-11c) shows the deformationcaused by the flexural hinge in the 3rd gap. Once again, the curves of shear deformationrotation and rigid rotation cycle about the origin in a characteristically elastic fashion.There was neither any visual or audible evidence of loss of bond between theconcrete and the steel jacket over the retrofit length of the beam adjacent to the column.Localized slip occurred around the edges of the jacket at the gaps. No plastic deformationwas detected by the strain gauges; none of the recorded strains reached the 1335 microstrainnecessary to indicate yielding (fig. 6-12). The strain patterns recorded by the strain gaugesdid not indicate any unexpected deformations in either phase of the loading cycle. The85b) Side Joint Transverse (SJT)Figure 6-12 Strains: RETRO-CD86MOMENT VS. STRAINRETRO-CD: TOP BEAM LONGITUDINAL (#006)-60-600^-400^-200^0^200STRAIN (MICROSTRAIN)600400c) Top Beam Longitudinal (TBL)MOMENT VS. STRAIN60405: 20zY%mOmaWna.a -204 RETRO-CD: TOP BEAM TRANSVERSE (#005)-60-1000^-500^0^500^1000STRAIN (MICROSTRAIN)d) Top Beam Transverse (TBT)Figure 6-12 Strains: RETRO-CD87permanent strain effects in the beam extension that had been observed in the three previousspecimens were not as prevalent here. Rather, some of those effects were found to appearin the beam section adjacent to the joint, as well as within the joint area itself, as evidencedby the strain values obtained from FCT, RJL, SJL (fig. 6-12a), SJT (fig. 6-12b), TBL (fig.6-12c) and TBT (fig. 6-12d). There had been some concern that the proximity of the gapto the joint area might unduly affect the integrity of the joint, but the measured levels anddistribution of strain in this area suggest that this concern was unwarranted. The joint areawas not confined by the steel jacket on all sides and the bond and confinement effects couldnot be relied upon in this complicated region. Except for the greater strains in the jointretrofit steel, and the greater capacity of the specimen, RETRO-CD behaved similarly toRETRO-CU.Under positive loading, the applied moment capacity of this specimen was found tobe 29.8 kNm, which was very close to the predicted value of 29.5 kNm from a sectionanalysis. Under negative loading, the applied moment capacity was found to be 44.3 kNm,which was significantly smaller than the theoretical capacity of 51.9 kNm. Again, this canbe attributed to the inability of the entire top layer of steel to attain its yield stress. FromTable 6.4 it may be seen that while the positions of the flexural hinge were accuratelypredicted, the actual values of the applied yield moment were not.88Table 6-4: RETRO-CD Section AnalysesFailure Location Mb (kNm) My (kNm) Map p (kNm)Positive LoadingBeam - End of Retrofit 16.8 75.6 ----Beam - Gap Region 21.4 28.5 29.8Column 22.9 63.2 ----Negative LoadingBeam - End of Retrofit 16.8 75.6 ----Beam - Gap Region 38.9 51.9 44.3Column 22.9 63.2 ----6.6 SUMMARY OF RETROFIT TESTINGThe testing program of the four retrofitted specimens exposed a few flaws in theoriginal designs of the retrofit schemes. Because flexural failure occurred at the end of theretrofit area of the specimen RETRO-SU it was desirable to add extra sleeve length to theextension of specimen RETRO-SD. Another option would have been to cut an extra setof gaps into the retrofit steel in the beam adjacent to the joint. However, the formerapproach was thought to be the safest in terms of the integrity of the joint area.Flexural failure in the column outside the retrofitted area of specimen RETRO-CUalso necessitated a change in the retrofit design of specimen RETRO-CD. This wasaccomplished by cutting the two extra gaps into the jacket of the beam, adjacent to the joint.89All of these design changes resulted in a set of unique specimens. This presented difficultiesin comparing specimens having similar geometric jackets, since comparisons based upon theinitial state of the original specimens were not possible.Because of excessive column deflection, which resulted from the development of theflexural hinge in the column, the LVDT's were found to be inadequate for measuring jointdeformations in specimen RETRO-CU.It was generally observed that while bond failure occurred in the two squarespecimens, no significant bond failure occurred in the two circular specimens. All thespecimens developed tensile strains in their jackets during testing. However, only thecircular jackets were effective in causing significant confinement stress on the core concrete,leading to a generally superior behaviour with regard to pinching.90CHAPTER 7EVALUATION OF RETROFIT SCHEMES7.1 INTRODUCTIONBefore considering the advantages and disadvantages of the different retrofit schemesunder study, it may be useful to summarize what was actually achieved in the retrofit processunder evaluation. The detailed discussions that follow may help to avoid pitfalls in thefuture planning and designing of retrofit schemes.First and foremost, the retrofit strengthened the deficient or damaged reinforcedconcrete joint area to such an extent that hinging or failure was deflected to adjacent areas:to the beam or column outside the retrofit, or to the beam within the intentionally weakened"hinge section". Any comparisons of "before" and "after" are thus somewhat misleadingbecause the "problem" was shifted from the most critical link to the next one in line. It isthus useful to consider the retrofit process in its entirety, which involves not only assessingthe benefits of introducing strengthening measures, but also of evaluating the behaviour ofthe post retrofit critical failure zone and of assuring that sufficient ductility is provided tomaintain favourable structure response.917.2 RELIABILITY OF BOND BETWEEN CONCRETE AND RETROFIT STEELIn the design of the retrofit jacket, the possibility of developing considerable bondbetween the concrete and the retrofit steel was considered. To break the continuity of thetubes in the longitudinal direction, circumferential gaps were provided in the casing. Ideally,the steel casing would act only as confinement for the concrete, and would not increase themoment capacity of the specimen. In this ideal situation, no longitudinal bond or interactionwould exist between the concrete and the retrofit steel. It is, however, impossible to expectlateral confinement of the concrete, resulting from a biaxial stress state in the jacket, withoutexperiencing some amount of mechanical interaction (friction) in the longitudinal direction.This interaction, of course, would increase the moment capacity, perhaps significantly,particularly where the beam jacket is physically connected to the column jacket.The chemical bond between the concrete and the steel did not appear to be veryconsistent, as was evidenced in certain areas of the square jacket retrofits, where bondfailure occurred early in the tests. A more likely scenario would be the development of amore reliable mechanism through friction or mechanical bond between the concrete and theretrofit steel. However, in the case of the retrofit tests, there was no conclusive proof of theexistence of any mechanical bond that could have resulted in an increase in the momentcapacity. Although, the moment capacities of all of the specimens did experience anincrease, it remains unclear which mechanisms were involved. No longitudinal tensile strainswere evident in three of the four specimens in the retrofit area between the gap and thejoint (RETRO-CU being the single exception). If sufficient bond of any kind had existed,it would have increased the moment capacity of the specimens and would have been92detected under negative loading as a tensile strain by the longitudinal strain gauge situatedat the top of the beam. In fact, during the tests this strain gauge registered compressivestrains under negative loading in all cases except for RETRO-CU, indicating a complete lossof bond in the longitudinal direction. The only explanation for this situation could be theinfluence of the transverse Poisson effect from the circumferential confinement strains.Without any contribution to the moment capacity from the steel jacket through bond,flexural yielding should have taken place at the beam-to-column interface.The actual situation appears to be one that is intermediate between the two extremes.The moment capacity of the section increased, but not because of a strongly positivemechanical or chemical bond between the concrete and steel. Rather, the bond appears tohave been fairly weak, but sufficient to provide a certain amount of friction between the twosurfaces, and enough to move the critical section along the length of the beam to the retrofitgap. This would imply a fair amount of slippage between the concrete and steel surfaces,so that all longitudinal tensile strains would be relieved. Because of cracking of the concretein the tension region, it cannot be concluded whether slippage did indeed occur in theretrofit section between the gap and the joint for the square specimens (as large amountsof concrete spalling took place in this area during the rotation of the flexural hinge thatformed at this point). A fair amount of slippage was, however, evident in the RETRO-CDspecimen.73 IMPROVEMENT IN DUCTILITYThe overall improvement in ductility of the specimens was substantial regardless of93the jacket shape used in the retrofit. The original unretrofitted specimens RCBC1 andRCBC2 were unable to withstand even two cycles of loading at a ductility level of 0 =2. Incomparison, even the square retrofits (RETRO-SU and RETRO-SD) provided a substantialimprovement of ductility, provided that failure occurred within the retrofitted zone. ForRETRO-SU, there was still substantial strength and stiffness at 0=3, before failure actuallyoccurred outside of the retrofit zone. For RETRO-SD, flexural failure occurred after 0=6had been successfully attained, showing a very significant improvement.The circular retrofit specimen RETRO-CU experienced a rotational ductility of about0 =5 before a flexural failure took place in the gap region. This has to be qualified becausethe onset of hinging in the column prompted a change in the intended loading cycle to forcea failure in the retrofit area. If testing had proceeded with a full cyclic load (in both positiveand negative directions) as originally planned, failure would certainly have occurred in thecolumn outside the retrofit area, before this level of ductility was reached. RETRO-CDshowed a remarkable improvement in ductility by successfully attaining a rotational ductilityof 0 =7 before flexural failure occurred.It should be noted that the true extent of the improvement in ductility for RETRO-SU and RETRO-CU cannot be fully assessed, since in both cases failure mechanismsdeveloped outside the retrofit zones, which affected the results of the intended experiments.7.4 POSITIONING THE GAP FOR PLASTIC HINGE DEVELOPMENTThe gap in the beam retrofit was initially incorporated in the design concept inresponse to the anticipated increase in the moment capacity from chemical and/or94mechanical bond between the retrofit steel and the concrete. The purpose of the gap wasto contain failure within the retrofitted zone of the specimen, and also to limit the resultingincrease in moment capacity. The gap was placed away from the joint area, due to theuncertainty of whether the joint area would be able to withstand extensive cracking withoutloss of bond and shear strength, even when retrofitted. It should be kept in mind that thejoint area was not entirely surrounded by the steel jacket, and placing a gap very close tothe beam-to-column interface would remove some of the confinement protection providedby the jacket. Once cracking occurred within the gap area during the formation of a plastichinge, it was necessary to prevent this network of cracks from extending into the joint area,which was shown to behave poorly during the unretrofitted tests. This was of greaterconcern for the square jacket retrofits which, because of their geometry, were not expectedto benefit from the confinement to the same extent as the circular jacket retrofits.Two major considerations thus governed the positioning of gaps: reduction of themoment capacity, which would be most effective with a gap close to the joint, andpreservation of the integrity of the joint zone, which called for an uninterrupted casing closeto the joint. A third consideration entered at a later stage, namely the placement of multiplegaps to provide an extended hinge zone. The higher moment capacities of retrofittedsections resulted in failure mechanisms outside of the retrofit areas, which promptedmodification of the retrofit design for subsequent tests: (a) the retrofit sleeve of RETRO-SDwas lengthened in response to an earlier failure mechanism developing outside the retrofitzone of RETRO-SU, and (b) extra gaps closer to the joint were placed in the beam retrofitof RETRO-CD in response to the column failure mechanism and relatively low ductility95levels developed in RETRO-CU. The desired behaviour was subsequently observed forRETRO-SD and RETRO-CD. No detrimental behaviour occurred in the joint area ofRETRO-CD despite the close proximity of the retrofit gaps to the joint area.7.5 DIMENSIONS OF THE RETROFIT JACKETThe thickness of the steel jacket was found to be sufficient for the purposes of thisparticular study. The basis of design was to replace the missing transverse steel in the jointarea, on a volume basis, with an equivalent amount of steel in the form of a steel jacket.A major consideration dictating the use of a steel jacket was ease of fabrication. Weldabilityof the casing required a minimum thickness to avoid excessive distortions and burn-throughwhen joining the component parts of the steel jacket. The weld needed to be able towithstand the high stress concentrations that would occur in the beam-to-column joint area;this could be a severe requirement, especially under repeated yield cycles.The fact that the retrofit steel had a yield strength of 267 MPa rather than thespecified 400 MPa did not alter the outcome of the experiment. Ideally, the steel should notadd moment capacity to the specimen, so a lower yield strength actually proved to be anadvantage in this particular case. In the design, the thickness of the jacket was chosen onthe basis of replacing 400 MPa transverse steel stirrups with an equivalent amount of 400MPa steel jacketing. Shear failure was not considered a problem with these tests, and thepurpose of the jacket was entirely to confine the concrete core. The 267 MPa steel fulfilledthat purpose without yielding. The 400 MPa steel would have behaved in much the samemanner since the elastic properties of steel do not vary noticeably with grade.96The length of the jacket sleeves and the placement of the gaps proved to be thecritical dimensions of the steel casing to avoid failure mechanisms occurring outside theretrofit zone. Two retrofit possibilities arose: extend the jackets sufficiently to force a failurein the joint itself, or provide a weakened hinge zone in a non-critical area. In both squareand circular jacket cases, if the retrofit gaps could be moved closer to the joint, thedimensions of the jacket as originally designed would be acceptable. In the case of RETRO-SD, this was viewed as an unnecessary risk, and the sleeve extention was opted for. Theretrofit lengths were based on the dimensions of the specimen after the retrofit; thereforethe circular and square retrofit sleeves had different lengths.7.6 CONFINEMENT EFFECTSThere was no evidence of an increase in compressive strength of the concrete due totriaxial confinement by the jacket. On the other hand, all failures of the retrofittedspecimens took place at sections that were not completely confined. At the retrofit gap,confinement could not be relied upon, because the jacket did not fully enclose the concrete.The measured moments at yield were generally consistent with predictions based on theconventional unconfined strength of the concrete. Had the concrete been able to developa larger compressive strength, the yield moment would have been even greater. Concreteneeds to be in triaxial compression to develop this extra compressive strength. In the beamsection of the specimens, this could only occur within the retrofit jacket. Even there, onlya small portion of the cross-sectional area would actually be in compression. The only wayto test this would be to force a failure in the beam portion of the specimen across a section97that is entirely encased in the retrofit steel. In general it may be stated that the increase inconcrete compressive strength due to confinement is only an issue in heavily loaded columns.In beams, the increase in bond and shear strength are of greater importance.The most obvious confinement effect provided by the steel jacket was thecontainment of the core, preventing the concrete from spalling and falling away. Theconcrete was kept in place throughout the testing sequence; even in the gap area the archingeffect was sufficient to contain the concrete. With the aid of mechanical bond (friction), thefailure zone was concentrated in the gap zone. A general bond failure between the beamreinforcement steel and the column concrete in the joint area was thus avoided.7.7 RATING THE RETROFIT SCHEMESConsidering the efficiency of the retrofit schemes, all the specimens behavedsatisfactorily, with a slightly better performance being exhibited by the circular retrofits. Thesquare retrofits did, however, behave well enough to merit consideration as a reasonableretrofit scheme. Evidently, the increase in the moment capacity provided by the concrete-to-steel bond, and the containment provided by the jacket, regardless of its geometric shape,were the overriding factors in the improvement of ductility. Placing the gap closer to thejoint helped the effectiveness of the jacket as designed.RETRO-CD provided the greatest retrofit improvement, with RETRO-SD a closerunner-up. Failure was observed to be entirely within the retrofit region of these twospecimens. RETRO-SU and RETRO-CU were affected by failure modes outside theretrofit region, and their performances should thus not be compared to RETRO-CD and98RETRO-SD. RETRO-CU developed a larger ductility than RETRO-SU, but was hamperedby the failure in the column under negative loading. RETRO-SU developed reasonableductility under positive loading, until flexural and shear failure outside the retrofit zoneunder negative loading took place.99CHAPTER 8SUMMARY AND CONCLUSIONS8.1 SUMMARYThe pre-1971 Canadian design standards for reinforced concrete construction lackedspecific guidelines for seismic design; for example, details relating to ductility did not existin the CAN3-A23.3-M66 (1966) standard. As a result, structures that were designed andbuilt before changes in the 1970's are poorly detailed for ductility, which may have direconsequences in the event of a major earthquake. The primary focus of this study centeredon the improvement of ductility of such deficient reinforced concrete frame structures,especially the ductility of the beam-to-column joint area.The purpose of this study was to determine the suitability of a possible retrofitmethod to strengthen and improve the ductility of damaged or deficient beam-to-columnjoints. The retrofit method considered involved the use of a steel jacket to encase the joint;the gap between the steel jacket and concrete core was filled with cement grout. Twoshapes of retrofit jacket were tested in this study: a circular steel jacket and a square steeljacket.The test specimens was based on a structure of about half standard size designed100according to the Canadian building code (NBCC, 1970) and concrete standard (CSA, 1966).The specimens themselves consisted of a subassembly of two column halfs and half a beam.The two initial tests of the reinforced concrete beam-to-column joint specimens confirmedthat the detailing for ductility was inadequate. These specimens exhibited poor behaviourunder cyclic loading.The joint specimens were then retrofitted with a 3.0 mm thick steel grout filled jacket.Four retrofitted specimens were tested: two were the initial previously tested and damagedspecimens, the other two were fresh specimens. The two damaged specimens wereretrofitted with a square and circular jacket, and the same was done for the two undamagedspecimens.8.2 CONCLUSIONSIn all the retrofitted specimens the joint area itself was sufficiently strengthened todeflect failure to adjacent areas. When considering overall behaviour, each specimenexhibited an improvement in ductility, although only two of the specimens actually developeda failure mode entirely within the retrofit region. Even the two specimens which developedpartial premature failures outside the retrofit jacket exhibited improved ductility behaviourin the retrofit region where hinging took place within the retrofit gap. It was generallyobserved that a circular jacket retrofit was more beneficial than a square jacket retrofit. Inpractical terms, however, the difference between the two shapes was not significant.A side effect of this retrofit method was the increase in moment capacity of thesections. This was a result of a composite action between the concrete and steel jacket, as101well as of a probable increase of the compressive strength of the core concrete. Thisincrease in the moment capacity should be minimized to avoid undue distress in otherregions of the structure that have not been designed for such large moments. To limit theincrease in moment capacity of the joint area, it is recommended that retrofit gaps be placedclose to the beam-to-column joint area. A number of gaps may be advisable to avoidconcentrated yielding and thus reduced ductility of the reinforcement bars within the gapareas. For the dimensions of the specimens in this study three gaps seemed to be sufficient.When using square jackets, care should be taken to avoid placing the gaps too close to thejoint area. The joint area has been shown to be weak, and without the added benefit of theradial confinement provided by a circular retrofit jacket this joint area, which is effectivelywithin the column, can be significantly affected by a plastic hinge forming too close to it.On the other hand, placing the gaps closer to the joint area would ensure plastichinging in the retrofit area and would avoid the need for an increase of the retrofit lengthto avoid failures outside the retrofit region. If this cannot be accomplished, the lengths ofthe sleeves must be increased in order to avoid failure at the end of the retrofit jacket, ineither the beam or the column. A jacket thickness designed to replace the missingtransverse steel, on a volume basis, was found to be sufficient. For practical reasons duringconstruction (eg. welding) a minimum thickness would be advisable.Since none of the retrofitted specimens failed in the same manner as the unretrofittedones, no conclusions can be drawn pertaining to the effectiveness of the proposed retrofitmethod in the immediate joint area. Follow-up tests are necessary to determine the amountof strengthening and added ductility needed in that region.102CHAPTER 9RECOMMENDATIONS FOR FURTHER STUDY9.1 CLARIFICATIONS AND MODIFICATIONS TO THIS RETROFIT STUDYA few of the material properties that were important to this study are still unclear,and some work should be undertaken to quantify their effects on the results obtained. Mostimportantly is the strength of the composite action or bond between the retrofit steel andthe concrete under bending. Another important material property that should bedetermined is the increase of concrete compressive strength under bending action in thebeam as well as in the joint area.Some of these effects can be examined through a retrofit study which is similar to theone described in this thesis, but having more instrumentation. Additional placement ofstrain gauges in sensitive areas of the retrofit jacket, such as in the joint and gap areas,would go a long way toward determining the strain patterns within the steel jacket. A setof strain gauges placed on the longitudinal reinforcement bars would also help to determinewhen and where yielding is taking place within the concrete reinforcement steel.Scale factors could also make a difference. The sizes of the specimens in this studywere limited by physical constraints of equipment and space in the Structures Laboratory103at the University of British Columbia. Specimens of nearly full scale would be desirable, asthe scales of the specimens and the retrofit jackets would not necessarily increase uniformly.A set of specimens should be tested in which all failure modes take place within the retrofitarea. This would require that the specimens be designed with the gaps suitably located. Afinite element analysis of a model could be conducted to predict the behaviour of theretrofitted joint once the requisite material properties have been determined.9.2 BEYOND THE SIMPLE BEAM -TO-COLUMN JOINTThe retrofit scheme proposed seems promising, as it achieved the goal of improvingthe ductility capacity of beam-to-column joints which were originally poorly detailed forductility. However, the simplicity of the tested joint has only limited usefulness. Columnswith only a single framing beam are rarely found, and then usually only in bridges. Cycliclateral loading is the simplest of possible motions that may affect a joint of this type. In thecase of a bridge structure, the motions during an earthquake may be much more complex,and may include motions parallel to the bridge deck, which would apply torsional motion aswell as lateral motion at 90 degrees to that observed in this study. To determine theseeffects, a scale model of a single or double bent frame should be tested on a shake table.Beam-to-column joints in other structures tend to be of greater complexity. Exteriorjoints also have transverse beams framing into the column, and interior joints have fourbeams joining at the column. Modifications to this study to include effects of transversebeams are straigtforward enough, but require a larger laboratory set-up. Of much greaterimportance is the presence and effect of slabs in a reinforced concrete structure. Most104buildings using this type of construction incorporate columns, longitudinal and transversebeams, as well as slabs. The presence of slabs significantly increases the complexity of theretrofit problem. Confinement of the column in the joint region can still be achieved withthe presence of a slab, but the slab interferes with the retrofit of the beam, and a reasonableplan must take into account the presence of the slab in the joint area.The results of this study indicate promising possibilities in the use of steel jacketretrofits for the improvement of ductility of reinforced concrete beam-to-column joints. Theabove recommendations can be used to extend this study, and to determine the usefulnessof this retrofit method for various kinds of reinforced concrete construction.105REFERENCESAmerican Concrete Institute (ACI),  (1991), "Recommendations for the Design of Beam-Column Joints in Monolithic Reinforced Concrete Structures.", ACI Manual of ConcretePractice 1991, Part 3, Detroit MI, pp 352R.1-352R.18Alameddine, Fadel.; Ehsani, Mohammed R., (1991), "High Strength RC ConnectionsSubjected to Inelastic Cyclic Loading.", Journal of Structural Engineering, ASCE Vol 117,No 3, pp 829-850Bolong, Z.; Yuzhou, C., (1991), "Behaviour of Exterior Reinforced Concrete Beam-ColumnJoints Subjected to Bi-directional Cyclic Loading.", Design of Beam-Column Joints forSeismic Resistance, James 0. Jirsa ed., ACI, Detroit MI, pp 69-96Canadian Standards Association (CSA), (1966), "Code for the Design of Plain or ReinforcedConcrete Structures.", CSA Standard A23.3-1966, Canadian Standards Association, OttawaONChai, Y.H.; Priestley, M.J.N.; Seible, F., (1991), "Seismic Retrofit of Circular Bridge Columnsfor Enhanced Flexural Performance ", ACI Structural Journal Vol 88, No 5, pp 572-584Cheong, H.K.; Perry, S.H.,  (1991), "Concrete Columns with Lateral Prestressing.", Journalof Engineering Mechanics, ASCE Vol 117, No 1, pp 70-87Ehsani, Mohammed R.; Wight, J.K.,  (1990), "Confinement Steel Requirements forConnections in Ductile Frames.", Journal of Structural Engineering, ASCE Vol 116, No 3,pp 751-767Filiatrault, Andre., (1992), "Reinforced Concrete Beam-to-Column Tests" PrivateCommunication, Ecole Polytechnique, Montreal PQGanesan, N.; Murthy, J.V. Ramana.,  (1990), "Strength and Behaviour of Confined SteelFiber-Reinforced Concrete Columns.", ACI Materials Journal Vol 87, No 1, pp 221-227Joh, O.; Goto, Y.; Shibata, T.,  (1991), "Influence of Transverse Joint and BeamReinforcement and Relocation of Plastic Hinge Region on Beam-Column Joint StiffnessDeterioration.", Design of Beam-Column Joints for Seismic Resistance, James 0. Jirsa ed.,ACI, Detroit MI, pp 187-223106Kaku, T.; Asakusa, H., (1991), "Ductility Estimation of Exterior Beam-ColumnSubassemblages in Reinforced Concrete Frames.", Design of Beam-Column Joints forSeismic Resistance, James 0. Jirsa ed., ACI, Detroit MI, pp 167-185Katzensteiner, Bryan 0., (1993), "Use of Steel Fibre-Reinforced Concrete in SeismicDesign.", MASc Thesis, University of British Columbia, To be publishedKuan, Steven Y. W., (1993), "Ductility Demands of Filtered Earthquakes on ReinforcedConcrete Frames." PhD Thesis, University of British Columbia, To be publishedLeon, Roberto T., (1990), "Shear Strength and Hysteretic Behavior of Interior Beam-ColumnJoints.", ACI Structural Journal, Vol 87, No 1, pp 3-11Mander, J.B.; Priestley, M.J.N.; Park, R.,  (1988), "Observed Stress-Strain Behaviour ofConfined Concrete.", Journal of Structural Engineering, ASCE Vol 114, No 8, pp 1827-1849Mitchell, Denis., (1991), "Detailing for Ductility in Bridge Columns.", Notes, Bridge DesignSeminar Vancouver, BCMorishita, Y.; Tomii, M.; Yoshimura, K.,  (1988), "Experimental Studies on Bond Strengthin Concrete Filled Circular Steel Tubular Columns Subjected to Axial Loads.", Investigationson Transversely Super Reinforced Concrete Structures and Concrete Filled Steel TubeStructures, M. Tomii ed., Kyushu University, Japan, pp 83-90Morishita, Y.; Tomii, M.; Yoshimura, K.,  (1988), "Experimental Studies on Bond Strengthin Concrete Filled Square and Octagonal Steel Tubular Columns Subjected to Axial Loads.",Investigations on Transversely Super Reinforced Concrete Structures and Concrete FilledSteel Tube Structures, M. Tomii ed., Kyushu University, Japan, pp 91-98Muguruma, H., (1984), "Ductility Improvement of Concrete Structural Members by UsingLaterally Confined Concrete with High Strength Hoop Reinforcement.", Proceedings, 8thWorld Conference on Earthquake Engineering San Francisco CA, Vol VI, pp 461-468National Building Code of Canada (NBCC),  (1970), National Research Council, Ottawa ONOtani, S., (1991), "The All Proposal of Ultimate Strength Design Requirments for RCBuildings with Emphasis on Beam-Column Joints.", Design of Beam-Column Joints forSeismic Resistance, James 0. Jirsa ed., ACI, Detroit MI, pp 125-144Priestley, M.J.N.; Park, R., (1984), "Strength and Ductility of Bridge Substructures.", RoadResearch Unit Bulletin 71, New Zealand National Roads Board Wellington, NZ107Priestley, M.J.N.; Park, R., (1985), "Concrete Filled Steel Tubular Piles Under SeismicLoading.", International Specialty Conference on Concrete Filled Steel Tubular StructuresHarbin, China Proceedings, pp 96-103Priestley, M.J.N.; Seible, F.; Chai, Y.H.; Sun, Z.L.,  (1990), "Steel Jacketing of BridgeColumns for Enhanced Flexural Performance.", Second Workshop on Bridge EngineeringResearch in Progress, National Science Foundation and Civil Engineering Department,University of Nevada, RenoRazvi, Salim.; Saatcioglu, Murat., (1989), "Confinement of Reinforced Concrete Columnswith Welded Wire Fabric.", ACI Structural Journal Vol 86, No 5, pp 615-623Ross, Bob.; Pianalto, Leonard.; Rylandsholm, Sven.; Sherkat, Reza.; Tong, Kevin.; Taheri,Omid.; Singha, Sonny.; Ng, George.,  (1992), "Cyclic Testing of Retrofitted ReinforcedConcrete Beams", Civil 321 Report, University of British ColumbiaSheikh, Shamim.; Uzumeri, S.M.,  (1982), "Analytical Model for Concrete Confinement inTied Columns.", Journal of the Structural Division, ASCE Vol 108, No ST12, pp 2703-2722Shin, Sung-Woo.; Ghosh, Satyendra K.; Moreno, Jaime  , (1989), "Flexural Ductility of Ultra-High Strength Concrete Members.", ACI Structural Journal Vol 86, No 4, pp 394-400Tidy, Michael S., (1988), "Hollow Circular Steel Tube Columns Filled with High-StrengthConcrete.", BASc Thesis, University of Toronto108APPENDIX ASTRUCTURE AND SPECIMEN DESIGNThe structure was designed according to the 1970 National Building Code of Canadaand the CAN3-A23.3-M66 Standard for Reinforced Concrete Design (1966).This appendix presents a summary of the design bending moments, axial loads andshear loads of the structure (fig. A-1), as well as the design specifications for the specimenused in the tests (fig. A-2).10925901Beams: 200 x 165Columns: 190 x 1901370137022.23.864.065.612.8 12.8 12.88.0 8.0 12.812. 4.826.640.35.3 7.2BENDING MOMENT (kNm)6.8-9.5a) Dimensions^b) Design Moments22.1■SHEAR LOADS (kN) AXIAL LOADS (kN)c) Design Shear Loads^d) Design Axial LoadsFigure A-1 Frame Design110A-A B-B C-CColumn: 190 x 190aaaaaA10mm 0 hoopsSpacing 70mmA10mm 0 stirrupsSpacing 70mmBeam: 165 x 200-•.- C2550495 --s-1200C1175Figure A-2 Specimen Design111APPENDIX BMATERIAL PROPERTIESTable B-1 presents the compressive strengths of the concrete and grout as well as thetensile strengths of the reinforcing steel and the retrofit steel used in the tests. Figures B-1to B-6 show the tensile coupon test results obtained from the reinforcing steel and jacketsteel. For the reinforcing bar specimens, the standard 0.2% offset strain method was usedto determine a value for the yield strain used in the calculations.112Table B-1: Material properties of concrete and steel.Concrete Strength: Unretrofitted specimensConcrete Strength: Retrofitted specimensTest #1Test #2AverageTest #1Test #2Average25.1 MPa27.5 MPa26.3 MPa29.3 MPa31.3 MPa30.3 MPaGrout Strength Test #1Test #2Test #3Test #4Average29.1 MPa28.3 MPa38.1 MPa29.8 MPa31.3 MPaReinforcing Steel Yield Strength Test #1Test #2Test #3Test #4Average535 MPa513 MPa551 MPa664 MPa566 MPaRetrofit Steel Yield Strength Test #1Test #2Average267 MPa267 MPa267 MPa1131141000800.1- 600a.aEE2000400STRESS VS. STRAINREINFORCEMENT BAR SAMPLE #3Figure B-3 Reinforcement Bar #3STRESS VS. STRAINREINFORCEMENT BAR SAMPLE #40^0.01^0.02^0.03^0.04^0.05^0.06^0.07^0.08STRAINFigure B-4 Reinforcement Bar #411540030010000.20 0.05 0.1STRAIN0.15Figure B-5 Retrofit Jacket #1STRESS VS. STRAINRETROFIT JACKET SAMPLE #2Figure B-6 Retrofit Jacket #2116APPENDIX CDATA FOR UNRETROFITTED SPECIMENSAppendix C illustrates all the data recorded for specimens RCBC1 and RCBC2, thetwo tests of the normal unretrofitted beam-to-column joint. The data are organized for eachof the two data sets in the following order: total rotation, comparison of rotationcomponents, cantilever rotation, rigid rotation, shear rotation and shear & rigid rotation.1170.10.050-0.05-0.1Figure C-1 RCBC1: Total RotationCOMPARISON OF ROTATION COMPONENTSRCBC10^ 50^ 100^ 150SCAN NUMBERTOTAL^SHEAR ^ RIGID ^ CANTILEVERFigure C-2 RCBC1: Rotation Components1183020100-10-20-30-0 1^-0.05^0CANTILEVER ROTATION (RAD)0.05 0.1MOMENT VS. CANTILEVER ROTATIONRCBC1Figure C-3 RCBC1: Cantilever RotationFigure C-4 RCBCP Rigid Rotation1193020.g. 10zYaiu11aa. -104-20-30-0 1^-0.05^0SHEAR & RIGID ROTATION (RAD)0.05 01Figure C-5 RCBC1: Shear RotationMOMENT VS. SHEAR & RIGID ROTATIONRCBC1Figure C-6 RCBC1: Shear and Rigid Rotation1200.1-0.10.050-0.05Figure C-7 RCBC2: Total RotationCOMPARISON OF ROTATION COMPONENTSRCBC20^ 50^ 100^ 150SCAN NUMBERTOTAL^SHEAR ^ RIGID   CANTILEVERFigure C-8 RCBC2: Rotation Components121Figure C-9 RCBC2: Cantilever Rotation111%711.117,11P1IINWIN (n.ftsr)Figure C-10 RCBC2: Rigid Rotation122ormoul n V IA IFigure C-11 RCBC2: Shear RotationFigure C-12 RCBC2: Shear and Rigid Rotation123APPENDIX DDATA FOR RETIROF111ED SPECIMENSAll data recorded for specimens RETRO-SU, RETRO-CU, RETRO-SD andRETRO-CD are presented in this Appendix. The data are organized by specimen, and foreach specimen the data are in the following order: total outer rotation, total inner rotation,hinge rotation, rotation components, cantilever rotation, rigid rotation, shear rotation, shear& rigid rotation and all of the strain gauge data.1240-1^i 0-2^1,3-3MOMENT VS. TOTRAL,SUTER ROTATION6040202^(;)0O:3e: -2044001Figure D-1 RETRO-SU• Total Outer RotationFigure D-2 RETRO-SU• Total Inner Rotation-60-0.1^-0.05^0^0.05TOTAL OUTER ROTATION (RAD)12550^100^150SCAN NUMBER200 250MOMENT VS. HINGE ROTATIONRETRO-SU6040200-20-40-60-0.1^-0.05^0^0.05^01HINGE ROTATION (RAD)Figure D-3 RETRO-SU: Hinge Rotation at GapCOMPARISON OF ROTATION COMPONENTSTOTAL^SHEAR ^ RIGID ^ CANTILEVERFigure D-4 RETRO-SU: Rotation Components12660-6040200-20-40Figure D-5 RETRO-SU• Cantilever RotationMOMENT VS. RIGID ROTATIONRETRO-SU-0.1 -0.05^ 0^ 0.05RIGID ROTATION (RAD)Figure D-6 RETRO-SU: Rigid Rotation011276040200-20-40-60-0.1^-0.05^0^0.05SHEAR & RIGID ROTATION (RAD)01Figure D-7 RETRO-SU• Shear RotationMOMENT VS. SHEAR & RIGID ROTATIONRETRO-SUFigure D-8 RETRO-SU: Shear and Rigid Rotation128Figure D-9 RETRO-SU: Front Column Longitudinal (FCL)MOMENT VS. STRAINRETRO-SU: FRONT COLUMN TRANSVERSE (#007)1Y_FZw2020w3a.a -20a-60 ^-300 -200^-100^0^100^200^300STRAIN (MICROSTRAIN)Figure D-10 RETRO-SU: Front Column Transverse (FCT)129Figure D-11 RETRO-SU• Rear Column Longitudinal (RCL)130Figure D-13 RETRO-SU• Rear Joint Longitudinal (RJL)Figure D-14 RETRO-SU: Rear Joint Transverse (RJT)1316040200-20-40-60-300^-200^-100^0^100STRAIN (MICROSTRAIN)200 300MOMENT VS. STRAINRETRO-SU: SIDE JOINT LONGITUDINAL (#005)Figure D-15 RETRO-SU: Side Joint Longitudinal (SJL)Figure D-16 RETRO-SU: Side Joint Transverse (SIT)1322202-200^-100^0^100STRAIN (MICROSTRAIN)200Figure D-17 RETRO-SU• Top Beam Longitudinal (TBL)MOMENT VS. STRAINRETRO-SU: TOP BEAM TRANSVERSE (#008)Figure D-18 RETRO-SU: Top Beam Transverse (TBT)133Figure D-19 RETRO-SU• Top Extension Transverse (TET)134Figure D-20 RETRO-CU: Total Outer RotationFigure D-21 RETRO-CU: Total Inner Rotation135Figure D-22 RETRO-CU: Hinge Rotation at GapFigure D-23 RETRO-CU• Rotation Components136Figure D-24 RETRO-CU• Cantilever RotationFigure D-25 RETRO-CU• Rigid Rotation137Figure D-26 RETRO-CU• Shear RotationFigure D-27 RETRO-CU: Shear and Rigid Rotation138Figure D-28 RETRO-CU: Front Column Longitudinal (FCL)Figure D-29 RETRO-CU: Front Column Transverse (FCT)139-40-60-400^-300^-200^-100^0^100STRAIN (MICROSTRAIN)-204020600200 300 400MOMENT VS. STRAIN6040g• 20z5CW2 0020W:3I- -20a-40-60RETRO-CU: REAR COLUMN TRANSVERSE (#001) MOMENT VS. STRAINRETRO-CU: REAR COLUMN LONGITUDINAL (#000)Figure D-30 RETRO-CU: Rear Column Longitudinal (RCL)-400^-300^-200^-100^0^100^200^300^400STRAIN (MICROSTRAIN)Figure D-31 RETRO-CU: Rear Column Transverse (RCT)140Figure D-32 RETRO-CU: Rear Joint Longitudinal (RJL)141Figure D-34 RETRO-CU• Side Joint Longitudinal (SJL)Figure D-35 RETRO-CU:. Side Joint Transverse (SJT)1426040Ui20-40-60-400^-300^-200^-100^0^100STRAIN (MICROSTRAIN)200 300 400Figure D-36 RETRO-CU• Top Beam Longitudinal (TBL)MOMENT VS. STRAINRETRO-CU: TOP BEAM TRANSVERSE (OOS)Figure D-37 RETRO-CU: Top Beam Transverse (TBT)143-60-400^-300^-200^-100^0^100STRAIN (MICROSTRAIN)-20-404020600200 300 400MOMENT VS. STRAINRETRO-CU: TOP EXTENSION LONGITUDINAL (#020)Figure D-38 RETRO-CU• Top Extension Longitudinal (TEL)6040g1 201-zW2 0020W=Iaa -20't-40MOMENT VS. STRAINRETRO-CU: TOP EXTENSION TRANSVERSE (#010) -60-1000^-500^•^0^500^1000STRAIN (MICROSTRAIN)Figure D-39 RETRO-CU• Top Extension Transverse (TET)144Figure D-40 RETRO-SD: Total Outer RotationFigure D-41 RETRO-SD: Total Inner Rotation145Figure D-42 RETRO-SD: Hinge Rotation at GapFigure D-43 RETRO-SD: Rotation Components146.^.Figure D-44 RETRO-SD: Cantilever RotationFigure D-45 RETRO-SD: Rigid Rotation147Figure D-46 RETRO-SD: Shear RotationFigure D-47 RETRO-SD: Shear and Rigid Rotation148Figure D-48 RETRO-SD: Front Column Transverse (FCT)Figure D-49 RETRO-SD: Rear Joint Longitudinal (RJL)149Figure D-51 RETRO-SD: Side Joint Transverse (SIT)150.^,Figure D-52 RETRO-SD: Top Beam Longitudinal (TBL)Figure D-53 RETRO-SD: Top Beam Trans-verse (TBT)151Figure D-54 RETRO-SD: Top Extension Transverse (TET)152Figure D-56 RETRO-CD: Total Inner Rotation1536040-20200-40-60010.05-0.1 -0.05MOMENT VS. H NGE ROTATIONRETRO-CD0HINGE ROTATION (RAD)Figure D-57 RETRO-CD: Hinge Rotation at GapFigure D-58 RETRO-CD: Rotation Components154Figure D-59 RETRO-CD: Cantilever RotationFigure D-60 RETRO-CD: Rigid Rotation155Figure D-61 RETRO-CD: Shear RotationiFigure D-62 RETRO-CD: Shear and Rigid Rotation156Figure D-63 RETRO-CD: Front Column Longitudinal (FCL)Figure D-64 RETRO-CD: Front Column Transverse (FCT)1571 20YOa. -20-40-60-300^-200^-100^0^100STRAIN (MICROSTRAIN)4060200 300MOMENT VS. STRAINRETRO-CD: SIDE JOINT LONGITUDINAL (#002)MOMENT VS. STRAINRETRO-CD: REAR JOINT LONGITUDINAL (#000)Figure D-65 RETRO-CD: Rear Joint Longitudinal (RJL)-300^-200^-100^0^100STRAIN (MICROSTRAIN)Figure D-66 RETRO-CD: Side Joint Longitudinal (SJL)200^300158RETRO-CD: SIDE JOINT TRANSVERSE (#001)60400W:7,aa -20a-406040F. 20zYW202aW..1ELa -20a-40-60-600^-400^-200^0^200STRAIN (MICROSTRAIN)600400MOMENT VS. STRAIN-60-300^-200^-100^0^100^200^300STRAIN (MICROSTRAIN)Figure D-67 RETRO-CD: Side Joint Transverse (SJT)MOMENT VS. STRAINRETRO-CD: TOP BEAM LONGITUDINAL (#006)Figure D-68 RETRO-CD• Top Beam Longitudinal (TBL)1596040202F2 00aa -20-40-60MOMENT VS. STRAINRETRO-CD: TOP BEAM TRANSVERSE (#005)0STRAIN (MICROSTRAIN)Figure D-69 RETRO-CD: Top Beam Transverse (TBT)-1000^-500 500^1000Figure D-70 RETRO-CD: Top Extension Trans-verse (TET)160


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items