GLULAM CONNECTIONSUSING EPOXY GLUED-IN REBARSbyRobert WiktorMgr rnz. (M.Eng.), Warsaw Technical University, Warsaw, Poland, 1990A THESIS SUBMITFED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Civil EngineeringWe accept this thesis as conformingto the required standardTHE UNIVERSiTY OF BRITISH COLUMBIAOctober 1994© Robert Wiktor, 1994In 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 C/ V/4 J/?/A/The University of British ColumbiaVancouver, CanadaDate ocn,&c, / k’DE-6 (2/88)ABSTRACTThis thesis is a part of the research aimed at developing a reliable moment resisting connectionmethod for timber structures based on gluing the reinforcing bars to the wood.The aims of this thesis are:- to investigate the influence of atmospheric factors (such as moisture and temperature) on theperformance of the glued-in rebar joints,- to examine the possibility of increasing the bearing resistance of the glulam (compression perpendicular to the grain) by gluing the rebars perpendicularly to the grain,- to implement the glued-in rebar technique in creating a moment resisting joint of a staticallyindeterminate glulam frame for a multi-storey building,- to provide the designers of timber buildings with information on how to design the glued-inrebar joints,- to verify the possible savings derived from the use ofthe moment resistingjoints in glulam frames.The findings of this research are:- the glued-in rebar connection may be considered reliable in the temperature and moistureconditions which can occur in the building; the connection was found to be influenced by thoseconditions to a lesser degree than the glulam structure itself,- the gluing of the rebars under the bearing plates can increase the compressive resistance of theglulam in the direction perpendicular to the grain by 100%,- the rebars glued perpendicularly to the grain have an additional effect of increasing the shearcapacity of the glulam members,- full size beam-to-columnjoints using the glued-in rebar idea were tested; the connection provedUto have bending and shear resistance equal or greater than the resistance of the glulam memberswhich were joined,- a ductile behaviour of the beam-to-column connections was observed prior to the failures,- a set ofguidelines was developed to facilitate the design ofglued-in rebar joints in the multi-storeytimber frames,- by using moment resisting glued-in rebar connections it is possible to save 15% of the glulamvolume versus the traditional hinged frame design.IllTABLE OF CONTENTSAbstract iiTable of Contents ivList of Tables viiList of Figures ixAcknowledgementsChapter 1 INTRODUCTION AND LITERATURE OVERVIEW 11.1 Introduction. 11.2 Literature Overview. 3Chapter 2 MOISTURE CONTENT AND TEMPERATURE TESTS 152.1 Test Objectives. 152.2 Development of a Test Specimen. 172.3 Moisture Content Tests. 282.4 Temperature Tests. 422.5 Conclusions. 52Chapter 3 COMPRESSION PERPENDICULAR TO THE GRAIN 543.1 Test Objective. 543.2 Materials, Test Setups, and Test Procedures. 553.3 Flush Rebar Tests. 593.4 Protruding Rebar Tests. 723.5 Guidelines for Bearing Reinforcement with #15 Rebars. 793.6 Conclusions. 80lvChapter 4 BEAM-TO-COLUMN JOINTS 814.1 Test Objectives. 814.2 Preliminary Analysis of the Frame for a Multi-storey Building. 824.3 Test Setup. 874.4 Materials. 924.5 Joint Design. 944.6 TestT-1. 994.7 Test T-2. 1104.8 TestT-3. 1244.9 Conclusions. 135Chapter 5 DESIGN GUIDELINES FOR GLUED-IN REBAR JOINTS 1365.1 Objective. 1365.2 Design Philosophy. 1375.3 Design Assumptions. 1385.4 Moment Resistance of the Joint. 1405.5 Bearing Reinforcement of the Column Plates. 1415.6 Axial Resistance of the Joint. 1435.7 Combined Axial and Bending Resistance of the Joint. 1435.8 Shear Resistance of the Joint. 1445.9 Resistance of the Bolts, Plates, and Welds. 1455.10 Serviceability Check. 146Chapter 6 COMPARISON WITH STATICALLY DETERMINATE FRAME 1476.1 Objective. 1476.2 Design of the Pinned Frame. 1476.3 Design of the Members. 1486.4 Cost Comparison. 154VChapter 7 CONCLUSIONS 156Bibliography 158Appendix A THEORETICALDEFORMATIONSAND STRESSES IN AGLUED-IN 161REBAR SPECIMEN DURING FREEZING/HEATING CYCLE1. Assumptions. 1612. Freezing Cycle: from +20°C to -30°C. 1633. Heating Cycle: from +20°C to +50°C. 1684. Summary. 172Appendix B THE ANALYSIS OF THE RIGID FRAME FOR A MULTI-STOREY 173OFFICE BUILDING1. General Design Assumptions. 1732. Loads. 1753. Load Combinations. 1814. Frame Analysis. 1825. Results of the Analysis. 1836. Design of the Members. 201vLIST OF TABLESTable 2.1 Selected thermal properties of steel and wood. 15Table 2.2 Results of tension tests performed on #15 rebars. 19Table 2.3 Results of the control sample testing. 24Table 2.4 Results of the M.C. sample testing. 40Table 2.5 Results of the temperature sample testing. 50Table 3.1 Results of the preliminary tests with bearing plates. 60Table 3.2 Results of the “beam specimen” tests with bearing plates. 63Table 3.3 Results of the tests with two bearing plates. 68Table 3.4 Results of the multi-rebar tests. 70Table 3.5 Results of the preliminary tests with protruding rebars. 73Table 3.6 Results of the “beam specimen” tests with protruding rebars. 76Table 4.1 Internal forces governing the design of the elements. 85Table 4.2 Calculated resistance and size of the elements. 85Table 4.3 Results of tension tests performed on #20 rebars. 92Table 4.4 Capability of glued-in rebars to transfer tension loads. 96Table 4.5 Components of the total deflection in test T-1. 107Table 4.6 Components of the total deflection in test T-2. 120Table 4.7 Comparison of the LVDT 3 readings in tests T-1 and T-2. 121Table 4.8 Comparison of the LVDT 4 readings in tests T-1 and T-2. 121Table 4.9 Components of the total deflection in test T-3. 131Table 4.10 Comparison of the LVDT 3 readings in tests T-1, T-2, and T-3. 132Table 4.11 Comparison of the LVDT 4 readings in tests T-1, T-2, and T-3. 133Table 5.1 Minimum embedment length for various size of the rebars. 139vuTable 6.1 Maximum internal forces in the hinged frame. 148Table 6.2 Volume of the glulam used for different types of frame. 154Table 6.3 Cost of a glued-in rebar joint. 154Table 6.4 Cost comparison between frames. 155Table B.1 Load cases used in the numerical analysis of the frame. 182Table B.2 Internal forces due to the load combination “a”. 183Table B.3 Internal forces due to the load combination “b”. 186Table B.4 Internal forces due to the load combination “c”. 188Table B.5 Internal forces due to the load combination “d”. 191Table B.6 Internal forces due to the load combination “e”. 193Table B.7 Internal forces due to the load combination “f’. 196Table B.8 Internal forces due to the load combination “g”. 198Table B.9 Internal forces governing the design of the elements. 201yinLIST OF FIGURESFig. 1.1 Typical column base connection as developed by H.Riberholt. 4Fig. 1.2 Purlin joint (Riberholt, 1986). 4Fig. 1.3 Moment resisting knee joint (Riberholt, 1986). 4Fig. 1.4 Beam-column connection used in timber classrooms (McIntosh, 1989). 5Fig. 1.5 Beam-column connection used in the Te Awamutu College Gymnasium 5(McIntosh, 1989).Fig. 1.6 Portal frames at Jellie Park (Buchanan and Fletcher, 1989). 6Fig. 1.7 Portal apex connection (Buchanan and Fletcher, 1989). 6Fig. 1.8 Beam splice (Townsend, 1990). 7Fig. 1.9 Portal knee connection using structural bracket. 7Fig. 1.10 Beam-column connections (Fairweather, 1992). 8Fig. 1.11 Bearing connection using glued-in rods. 9Fig. 1.12 Shear reinforcement of the glulam beam. 9Fig. 1.13 Moment resisting splice connection of the glulam beam. 10Fig. 1.14 Knee joint and column to foundation joint (Turkovskij, 1991). 10Fig. 1.15 Pull-out test (Malczyk, 1993). 12Fig. 1.16 Beam splice test (Malczyk, 1993). 12Fig. 1.17 Racking test of the glulam column (Malczyk, 1993). 13Fig. 1.18 Knee joint test (Malczyk, 1993). 14Fig. 2.1 Specimen used for M.C. and temperature tests. 18Fig. 2.2 Typical L-D curve obtained from the rebar tension test. 20Fig. 2.3 Specimen during the test in the tension apparatus. 21Fig. 2.4 Typical L-D relationship for the control specimens. 23Fig. 2.5 Failure Mode I - rebar yielding. 25Fig. 2.6 Stress distribution along the rebar. 27IxFig. 2.7 Stress distribution at different load levels. 27Fig. 2.8 An example of the M.C. changes. 29Fig. 2.9 Drying of the specimens. 30Fig. 2.10 Load-deformation curves for the M.C. specimens. 32Fig. 2.11 Failure Mode U - pull-out. 34Fig. 2.12 Failure Mode ifi - glulam split. 35Fig. 2.13 Specimens Ki and K2 after failure. 36Fig. 2.14 Deformation range observed after pull-out failure. 37Fig. 2.15 Stresses, deformations, and M.C. in function of time. 39Fig. 2.16 Comparison of the control sample and the M.C. sample. 41Fig. 2.17 Temperature variations during the treatment. 42Fig. 2.18 Load-deformation curves for the temperature specimens. 44Fig. 2.19 Dimensional changes of the specimen as a function of time. 47Fig. 2.20 Internal stresses as a function of the temperature. 48Fig. 2.21 Internal stresses as a function of time. 49Fig. 2.22 Comparison of the control sample and the temperature sample. 51Fig. 3.1 Preliminary test setup. 55Fig. 3.2 Specimen in the testing machine during preliminary test. 56Fig. 3.3 Simple supported beam setup. 57Fig. 3.4 Specimen with top and bottom bearing plates. 58Fig. 3.5 Load-deformation curves for the specimens with 76x102mm plate. 65Fig. 3.6 Load-deformation curves for the specimens with 102x152mm plate. 65Fig. 3.7 Comparison of L-D curves for the specimens with 76x102mm plate. 66Fig. 3.8 Comparison of L-D curves for the specimens with 102x152mm plate. 67Fig. 3.9 Load-deformation curves for the specimens with two bearing plates. 69Fig. 3.10 Comparison of L-D curves for the multi-rebar specimens. 71Fig. 3.11 Specimens after the preliminary tests. 75Fig. 3.12 Specimens after the preliminary tests. 75xFig. 3.13 Typical L-D curves for the “beam” specimens with protruding rebars. 77Fig. 3.14 Failure loads versus the embedment length for the specimens with protrud- 78ing rebars.Fig. 4.1 Overall dimensions of the building. 83Fig. 4.2 Locations of the splice joints in the beams. 86Fig. 4.3 Overall dimensions of the reaction frame and the specimen. 87Fig. 4.4 Specimen in the reaction frame. 89Fig. 4.5 Load, reactions, and resulting moment and shear diagrams. 90Fig. 4.6 Locations of the gauges in the frame. 91Fig. 4.7 Locations of the gauges within the connection region. 91Fig. 4.8 Moment transfer from the column to the beam. 94Fig. 4.9 Forces created at the end of the column. 95Fig. 4.10 Column part of the specimen T-1. 99Fig. 4.11 Beam part of the specimen T-1. 100Fig. 4.12 Specimen T-1 before the test. 102Fig. 4.13 Locations of the gauges in the specimen T-1. 103Fig. 4.14 Joint in the specimen T-1 after failure. 104Fig. 4.15 Fracture of the rebar in the specimen T-1. 105Fig. 4.16 Deflections of the column during the test T-1. 106Fig. 4.17 Behaviour of the tension side of the connection. 108Fig. 4.18 Bending of the beam plate. 109Fig. 4.19 Column part of the specimen T-2. 111Fig. 4.20 Beam part of the specimen T-2. 112Fig. 4.21 Routing grooves. 113Fig. 4.22 Drilling perpendicular holes. 113Fig. 4.23 Gluing perpendicular rebars. 114Fig. 4.24 Drilling inclined holes. 114Fig. 4.25 Fitting inclined rebars and plates. 115Fig. 4.26 Loading procedures for the specimen T-2. 116xlFig. 4.27 Joint in the specimen T-2 during final loading. 117Fig. 4.28 Failure of the specimen T-2. 118Fig. 4.29 Deflections of the column during the test T-2. 119Fig. 4.30 Partial pull-out of the beam rebar in the specimen T-2. 122Fig. 4.31 Column part of the specimen T-3. 124Fig. 4.32 Beam part of the specimen T-3. 125Fig. 4.33 Loading procedures for the specimen T-3. 127Fig. 4.34 Specimen T-3 during final loading. 128Fig. 4.35 Joint in the specimen T-3 after failure. 129Fig. 4.36 Deflections of the column during the test T-3. 130Fig. 5.1 Forces in the connection. 138Fig. 5.2 Geometry of the beam to column joint. 140Fig. 5.3 Forces perpendicular to the grain in the column. 142Fig. 5.4 Bending of the beam plates. 146Fig. 6.1 Braced frame of the building. 147Fig. A.1 Internal stresses in the specimen due to temperature changes. 162Fig. A.2 Deformation of the specimen in phase 1. 164Fig. AS Deformation of the specimen in phase 2. 166Fig. A.4 Deformation of the specimen in phaseS. 168Fig. A.5 Deformation of the specimen in phase 6. 170Fig. B. 1 Overall dimensions of the building. 174Fig. B.2 Model of the frame used in the numerical analysis. 182xl’ACKNOWLEDGEMENTSI am very grateful to my supervisor Professor Borg Madsen for his guidance, valuablesuggestions, and encouragement throughout this research.Iwould like to express my gratitude to Professor R.O. Foschi and Professor R.G. Sexsmithfor reviewing the manuscript and constructive criticism.I would like to thank Mr. Paul Symons for his helpful participation in preparing andmaintaining instrumentation used in the experimental part of this research.My appreciation is extended to Mr. Dick Postgate and all technicians for their assistanceduring various phases of the experiments.Researchgrand from the Science Council ofBritish Columbia is gratefully acknowledged.The support from the following companies is much appreciated: Industrial Formulatorsof Canada, Surrey Laminated Products Ltd., Structurlam, Dickie Dee Ice Cream.Finally, Iwish to thank my familyand my friends for their support throughout mygraduatecareer.This thesis is joyfully dedicated to my loving wife Justyna.CHAPTER 1INTRODUCTION AND LITERATURE OVERVIEW.1.1 INTRODUCTION.Connections are essential parts of structures. In many cases, they are the most difficultparts to design and to accomplish. This is especially true for timber structures. It is relativelysimple to design a timber member but, usually, the challenge starts when it has to be connectedto the other members.Over the years, engineers invented many connection techniques but still the most commonly used are the traditional connectors: bolts, nails and pins. Unfortunately, these connectorscan’t be successfully used in statically indeterminate timber structures due to their lack ofstiffness. In those structures, the joints should be able to resist reversed bending moments dueto dynamic loads on the structure such as wind loads or earthquake loads. The traditionalconnectorsdo not provide enoughstiffness for thejoint, often due to oversized holes. Therefore,the need of a reliable, moment resisting connection method for glulam structures still exists.One of the most promising new techniques is the use of steel rods bonded to the timberby glues. This idea originated in Scandinavia and was used in the industry for over 30 years.Various modifications have been developed, some of them with very narrow applications.Recently, an increasing interest in this technique was observed. Several research programs inthis field are presently performed in Finland, Russia, Australia, New Zealand, and Canada.This thesis is part of a Canadian research program which started in 1992 at the Universityof British Columbia under supervision of Professor Borg Madsen.1.1.1 Scope of Thesis.This thesis is the second one dealing with glued-in rebar connections being part of theresearch program at UBC. The first one was presented by Robert Malczyk in 19931(R.Malczyk, “Glued-in Re-bar Connection”, 1993). The work presented here is a naturalcontinuation of previous research.The main aims of this thesis are:- to implement the glued-in rebar technique developed earlier, in moment resisting jointsfor multi-storey frame structures,- to provide the designers of timber buildings with design information on the behaviourof glued-in rebar connections,- to investigate the influence of atmospheric factors (such as moisture and temperature)on the performance of the glued-in rebar joints,- to develop a reinforcement technique for glulam members subjected to bearing (compression perpendicular to the grain).1.1.2 Outline of Thesis.The thesis consists of three parts: one theoretical part and two practical parts.An overview of recent experimental work in the use of steel rods glued into timbermembers is presented in Chapter 1.Chapter 2 deals with the general behaviour of the glued-in rebars when subjected tosevere changes of temperature or moisturecontent in the glulam. An impact ofthosechangeson the capacity of the connection is presented. The stresses created in the connection bythe atmospheric factors are discussed.Chapter 3 and 4 presents practical implementation of the glued-in rebar technique.Chapter 3 describes the reinforcing of glulam in compression perpendicular to the graincreating a very efficient bearing connection method. Chapter 4 describes the beam-tocolumn moment resistingconnection and the tests conducted at UBCon full scale specimens.Finally, chapters contains design guidelines for the glued-in rebar joints and chapter6 provides a comparison to the traditional design.21.2 LITERATURE OVERVIEW.Two different approaches ofthe glued-in rebar idea have been adopted by the researchers.In New Zealand and in Australia the steel rods are glued parallel to the grain of the glulam.This is continuation of an original idea developed in Scandinavia and improved by Danishresearchers in 1960’s. A different approach was developed in Russia and Finland. Theresearchers there are creating rigid connections by gluing rebars at an angle to the grain. Thislater idea is also followed (with some modifications) in the research program conducted at theUniversity of British Columbia.A comprehensive overview of the research related to the topic of this thesis is presentedbelow.1.2.1 Scandinavian Research.The first research on glued-in steel rods in laminated timber was performed in Swedenaround 1965.Special glued-in bolts were used in Denmark to provide connection for a glulam rotorin wind turbines (Riberholt, 1982).The major work on steel bars glued into the glulam was performed by Riberholt (1986,1988) in Denmark. The idea of the rods glued parallel to the grain was followed in thatresearch. Riberholt completed tension and shear tests on steel rods glued into the end grainofglulam beams (Riberholt, 1986) and recommended design procedures for the connection.The collected information was then applied to the splice joint in a beam. The beams weretested under both wet and dry, service conditions. The same report also describes a momentresisting column to foundation joint (see Fig.1.1). This type of joint became a standardconnection method for the glulam industry in Scandinavia.3GLULAM COLUMN,iTHREADED RODSCONCRETE INFILLINC____ _—CAVITY IN CONCRETE—FO UNDA TIQI — — — . — — — — — ——I — — — — — — — — —.REBARSFig.1.1 Typical column base connection as developed by FLRiberholLHis report (Riberholt, 1986) also describes tests on purlin joints (Fig.1.2) and momentresisting knee joints for portal frames (Fig.1.3). Further reports provide more test resultsand some empirical formulae for failure loads in this type of connection (Riberhold, 1988).-I I Wedge bonded toPURLIN the glulamI p =45—oc.,.--Fig.1.2 Purlin joint Fig.1.3 Moment resisting knee joint(Rthethoh, 1986). (Riberholt, 1986).41.2.2 New Zealand Research.An extensive program of glued-in dowels has been carried out in New Zealand sincelate 1980’s. The epoxied steel rods were used in beam-column connection for timberclassrooms (Fig.1.4) and in the Te Awamutu College Gymnasium (Fig.1.5).24mm threaded steelrod set In epoxy gluein top end of135 x 135 glulam pastFig.1.4 Beam-column connectionused in timber classrooms(McIntosh, 1989).Buchanan and Fletcher (1989) report on the design and construction of two indoorswimming pools which use epoxied steel dowels for attaching curved glulam portal frames(Fig. 1.6).The base connection was similar to that developed by Riberhold. The beam-column connections also used threaded rods glued into the end grains of the beams and pass throughthe column.The same researchers also reported on the use ofepoxy glued steel rods in portal apexconnection (Fig.1.7).6mm Washer underNut in recess• Thr.od.d st..I rods fixed irdo•nds of beamFig. 1.5 Beam-column connectionused in the TeAwamutuCollege Gymnasium(McIntosh, 1989).5r1 q i4 IIC,R1-0II o0oTownsend (1990) carried out an extensive testing program on epoxy glued steel dowelsusing New Zealand materials. The tests considered loading in both tension and shear, different types of epoxy and various rod geometries. Townsend (1990) also reports on somefull size beam splices that were tested in bending. High strength deformed bars glued parallelto the grain were used in that joint (see Fig.1.8). The splice connection performed very welland results showed they were generally stronger than the beams themselves.Several types ofknee joints were tested by Buchanan and Townsend (1990). The mostsuccessful one, using a structural steel bracket is shown on Fig.1.9.Fig.1.9 Portal knee connection using structural bracket.Fig.1.8 Beam splice (Townsend, 1990).7Fig.1.1O Beam-column connections (Faiiweathei; 1992).Fairweather (1992) reports the tests of beam-column connection under quasi-staticcyclic loading with a horizontal load applied at the top of the column. Four types ofjointswere tested (see Fig.1.1O):A- direct beam-column connection,B- connection using inside steel bracket,C- connection using steel brackets on each side of the continuous column,D- connection between continuous column and twin continuous beams using steelside brackets nailed to the beams and bolted to the rods epoxied into the column.Fairweather reports strong and stiff behaviour of those connections and excellent ductility.Some problems with brittle fracture of the wood have been encountered.Type A ConnectionType C Connection Type D Connection81.2.3 Russian Research.Connections using reinforcing bars glued into the glulam were introduced in Russiain mid 1970’s (Turkovskij, 1991). The idea originated with a problem where the bearingstresses in the beam were too high. Instead of providing a larger bearing plate, a steel rodwas glued perpendicularly to the grain close to the end of the beam so the concentrated loadwas distributed throughout the length of the rod (Fig.1. 11). This detail was found to workvery well.CLUED-INFig.1.1 1 Bearing connection usingglued-in rods.The concept was then applied to increase shear capacity of the beams (Fig.1.12) andwas later used as a repair method. The innovation here was to use the rods glued into theglulam at an angle of 45°.\\\\\\\\\\Fig. 1.12 Shear reinforcement of the glulam beam.— —9Turkovskij (1991) also reports the use of similar technique to create moment resistingjoints for glulam. The main difference with respect to the connections described in previoustwo sections is the use of inclined steel rods (the Russians use reinforcing bars). Commonlythe angle of inclination is 300. The other innovation is that the bars are glued separately intoboth members and then joined on site by an extra steel plate welded to the rebars. Fig.1.13shows how this concept is applied to the moment resisting splice connection of the beam.STEEL PLATE WITHI I SLOTTED HOLES ToMATCH THE STUBSS% % —— ——S. S. —— ——%S.%S.%___ -————\STSEL PLATE WELDEDTO STUBSFig. 1.13 Moment resisting splice connection ofthe glulam beam.Fig.1.14 Knee joint and column to foundation joint (Turkovskzj, 1991).CLUED-IN RODS WITHSTUB CLUED-IN RODS BEFORECICh10The rebars are glued with the stubs about 100mm sticking out of the beam. After the gluehas cured these stubs are bent down to lay flat on the beam’s edge. On the site a steel platewith slotted holes matching the position of the stubs is welded to the stubs. Some otherexamples of moment resisting joints which were tested and later implemented in practiceare shown on Fig.1.14Avery interesting aspect of the glued-in rod methods is the forcing of the failure intothesteel elements ofthejoint. After establishing the embedment length ofthe rebar requiredto yield it and exceeding this length in practice, the joints can be designed as the steel joints!In that way, the advantage of the small variability of the steel properties can be utilized inthe glulam structures. it is especially important in an earthquake design where a highlypredictable behaviour of the structure is desirable together with a high ductility level.1.2.4 Canadian Research.The idea of forcing the failure into the steel is followed in the research conducted atthe University of British Columbia.Malczyk (1993) reports series of pull-out tests conducted on the rebars glued at anangle to the grain (Fig.1.15). Three parameters were investigated: an angle of inclinationof the rebar, the embedment length, and the rebar’s diameter. Although the tests weresimilar to those conducted by the researchers described in previous section, someimprovements in the connection method was made:- welding close to the glulam surface was eliminated (so called ‘pre-welded” connectionwas developed),- overall appearance ofthe connectionwas improved bywelding the rebars to thebottomface of the steel plates,- bolted connection was preferred instead of welding on the building site.11reb ar______ _____Fig.1.15 Pull-out test (Malczylç, 1993).Some full scale tests were conducted (Malczyk, 1993) to verify the behaviour of theglued-in rebar connection in different joint configurations. The pre-welded technique andthe rebars glued at 300 angle were consistently used in the following research.First, a beam splice joint was tested (Fig.1. 16). Then, a column to foundation jointwas tested under lateral loading (Fig.1.17), and finally, some tests of the knee joint in theportal frame were conducted (Fig.1.18).#20 REBARS BOL TED 0 UTSIDE PLATEN——GL ULAM BEAM INSIDE STEEL PLATE(welded to rebars)Fig.1.16 Beam splice test (Malczylç 1993).)12GLULAM 170x370 mmSflflfl/flflflflZFig.1. 17 Racking test of the glulam column (Malczylc, 1993).The foundation and knee joint tests were conducted under reversed loading conditions, i.e.positive and negative moments were created in the joint.Steel failures were observed consistently. A steel-like ductile behaviour of the connectionswas reported. Some bearing problems under the plates wereencountered, however,they were later solved by increasing the bearing area of the plates.13Fig.1.18 Kneejoint test (Malczylc 1993).1.2.5 Summary.1. The idea of gluing steel rods into the glulam has been used in several countries for along time.2. The bars glued at an angle to the grain engage signfficant portion of the glulam cross-section, whereas, the bars glued parallel to the grain behave more like skin connectors.3. Inclined bars increase the shear capacity of the glulam cross-section.4. If the embedment length of the bar is larger than the length required to yield the bar,the joints may be designed as the steel joints.SS3100 mm14CHAPTER 2MOISTURE CONTENT AND TEMPERATURE TESTS.2.1 TEST OBJECTIVES.The primary objectives ofmoisturecontent (M.C.) and temperature tests were to establishwhether changes of the connection’s environment would have a deteriorating influence on thestrength and durability of the connection.The glulam is highly sensitive to changes in the moisture content and the steel is sensitiveto changes in the temperature. These two materials are combined in the joint and bonded bythe epoxy glue, however, their thermal expansion coefficients as well as their thermal conductivity differ significantly. It is commonly known that the steel expands more than the glulamwhen both materials are exposed to the same increase of temperature. That, however, is true,but only in the direction parallel to the grain. The coefficient of thermal expansion for woodperpendicularly to the grain is roughly 4 times larger than the same coefficient for steel (seeTable 2.1; note: the values for wood depend on M.C. and density; they are quoted below foroven-dry wood of density p =500 kg/rn3).Thermal property steel wood woodparallel perpendicularexpansion coefficient [xlO6per K] 11 4 40conductivity [W/mK] 58 0.30 0.16Table 2.1 Selected thermaiproperties ofsteel and wood.15The thermal conductivity of each material plays an important role as well. It can be seenfrom the Table 2.1 that the conductivity of the steel is approximately 200 times larger than theconductivity of the glulam parallel to the grain, and almost 400 times larger than the conductivityof the glulam perpendicular to the grain. That means, the same amount ofheat energywill flowthrough a material of one unit area and one unit thickness 400 times faster in the case of thesteel rebar than in the case of the glulam block in the direction perpendicular to the grain.The differences in thermal properties of the materials used in the glued-in rebar connection create internal stresses within both materials when the outside temperature changes.The magnitude of those stresses depends on the change in temperature, time, and the joint’sconfiguration. A particular example of this phenomenon (for tested specimens) is discussed insection 2.4.2. Here, a more general illustration will be presented.Let us consider the case of a joint (a rebar glued parallel to the grain) located under theroof of an industrial hail. Further let us assume that the temperature at that location can reach+40°C during the sunny winter day and rapidly drop below 0°C at night. The rebar will shrinkfaster and will shrink more than the glulam part of the joint. The steel will develop tensionstresses while the glulam will be in compression. Eventually, after several hours, an equilibriumwill be reached. The process will be reversed by the increasing temperature during the day. Thesituation may repeat over several days. At the end a heavy snowfall may occur creating additionalstresses. Will the connection, under these extreme conditions, behave as well as the one whichwas not subjected to the temperature changes at all? Will the bonc4 provided by epoxy, be reliabledespite the relatively high temperature?In another situation the glulam may expand due to the high humidity of the surroundingair while the rebar is not affected by this factor. The result is similar: internal stresses are createdin both materials and the process is time dependant. Thus, the question may be restated - canthe connection still be relyed upon after many changes in the glulam moisture content?Tests conducted at UBC from May to December 1993 were designed to clarify the abovequestions. Unfortunately, to answer them fully, much more time would be necessary - speciallyfor M.C. tests. In a real structure, it takes months for moisture to penetrate the glulam elementand even more time to dry it out. Additionally, this process repeats many times during the life16span of the structure. To get the answers quicker, a reasonable reduction of size of the testedspecimens was desirable. The number of cycles had to be limited, due to time constraints, aswell. Yet, the tests should give the results which could be extrapolated on the real structures.To compensate for the reduction of size and number of cycles the severity of the changes wasmade greater than what is normally found in a structure. The process of developing a suitabletest specimen is discussed in the next section.2.2 DEVELOPMENT OF A TEST SPECIMEN.It was necessary to develop a test specimen because it wasn’t practical to test a full sizeconnection. A control sample, consisting of6 test specimens, was chosen to give a better pictureof the variability of the test results. This sample was tested before the other samples (i.e. M.C.sample and temperature sample). The control specimens were tested in a constant temperatureof +20°C and their moisture contents were measured to be 11%.The test specimens had to meet some functional requirements to allow for the futureextrapolation of the results to the real structures. The requirements were as follows:- the control specimens should be loaded identically to the M.C. and temperature specimens,but the loading conditions didn’t need to reflect the real structure (although they need tobe realistic);- a tension test should be performed to comply with the idea of forcing the failure into thesteel (described in chapter 1);- the most disadvantageous orientation of the rebar with respect to the glulam fibres shouldbe chosen, but excessive stress concentrations should be avoided;- the specimens should have realistic dimensions - as close to the real structures as practical;- the specimens should be small enough to allow reasonably quick moisture penetration andheat transfer;- the manufacturing of the specimens should be as simple as possible;- the specimens should be easy to handle during the treatment (wetting, drying, heating, etc.)as well as during the strength testing;17Unfortunately, some of above requirements were in conflict. Ajudgement of the priorities wasmade and it was decided that the length (dimension parallel to the grain) of the specimens shouldbe reduced, but the other dimensions, as well as the size of a rebar, should be kept realistic.The moisture penetrates the glulam faster in the direction of the fibres than across the fibres.Thus, the time necessary for this process can be reduced by making the specimens shorter thanin the real structure. The length of the specimen was chosen to be 250mm (see a sketch on Fig.2.1). The rebar was glued perpendicularly to the grain, so the weakest direction of glulam wastesteLFig.2.1 Specimen usedfor MC. and temperature tests.+MEA SUREDDISPLACEMENTI2504182.2.1 Materials.GLULAM.Material used for all the tests described in this chapter was Douglas Fir glulam 24fmanufactured by Structurlam Ltd. in Penticton, B.C. The dimensions of the glulam blockwere as follows:thickness - 175mm;length - 250mm;height - varied because the blocks were prepared from glulam beams of different sizes- minimum was 450mm.A 19mm (3/41) hole was drilled in the center of the block which is 3mm larger than the roddiameter. The depth of the hole was 405mm.REBAR.Deformed weldable reinforcing bars (rebars) of grade 400 (yield stress = 400MPa)were used. All tests were conducted with #15 rebars. Their nominal diameter was 16mmand nominal area 200mm.The nominal dimensions are equivalent to those of a plain roundbar having the same mass per meter as the deformed bar. Table 2.2 summarizes the resultsof tension tests of these rebars and Fig. 2.2 shows typical load deformation curve obtainedduring the tests.yield E ultimate# load stress strain elastic load stress strain[kN] [MPa] [MPa] [kN] [MPa]1 88.6 443 0.0024 184600 129.3 647 0.1172 89.5 448 0.0023 194800 133.2 666 0.1483 89.0 445 0.0024 185400 129.9 650 0.1884 89.4 447 0.0023 194300 130.2 651 0.129Average 89.1 446 0.0024 189800 130.7 654 n/aTable 2.2 Results of tension testsperformed on #15 rebars.1910U)Cl)wIC’)Fig.2.2 Typical load-deformation curve obtainedfrom the rebar tension test.The total length of the rebar was 905mm. It was embedded in the glulam block for adistance of 405mm while protruding 500mm (see Fig. 2.1).GLUEThe epoxy glue “IFC-SP” was used for the connections. It is manufactured by IndustrialFormulators of Canada Ltd., Burnaby, B.C. It consists of two parts: raisin and hardener.The mixing ratio is 100 to 42 (by weight) respectively.The glue was poured into the hole just before the rebar was inserted. By inserting therebar slowly, the glue was pushed up, and in this way, the creation ofair pockets was avoided.The air pockets, as described in chapter 1, did cause a lot ofproblems for previous researchersbecause the hole was drilled horizontally.An average of 65cm3of glue per hole was used (+/_ 5%). The pot life of the glueSTRAIN20varied from ito 3 hours depending on the outside temperature and the volume of the glueused. The glued specimens were cured at least 7 days before testing in order to reach thefull strength of the glue.upper beam with jaws(moving part)rebar protruding from the blockfastened in the jawsmiddle reaction beamglulam block with embedded rebar2.2.2 Test Setup and Test ProceduresFig.2.3 Specimen during the test in the tension apparatus.21A Baldwin tension apparatus was available in the Structures Laboratory of UBC CivilEngineering Department. This machine has 3 different ranges of loading:low 0- 7lkN (l6000lb),medium 0- 356kN (80000lb),high 0- 1779kN (4000001b).The apparatus was set for medium range and used to perform all tests.The protruding end of the rebar was fastened in upper jaws of the apparatus while thetop surface of glulam was supported (via 38mm thick steel plate with a hole for the rebar)by the middle reaction beam of the apparatus. The applied load created tension in the rebarand compression in the glulam block. The specimen set for testing in the apparatus is shownon Fig. 2.3. This setup was chosen for its simplicity and it was not intended to reflect anyparticular situation in the real life structures.It was decided to measure the displacement of a point located on the rebar’s surface,as close to the glue line as possible (see Fig.2.1), relative to the top surface of the glulam (orplate). This displacement consisted of the elongation of the embedded part of the rebarplus the deformation of the wood fibres at the perimeter of the glue coating.The body of the displacement gauge (LVDT) was fixed to the steel plate while itsmoving tip was resting on a stud glued to the rebar at the measuring point. The load appliedto the rebar was measured by the load cell of the apparatus. Both measurements, i.e. loadand displacement were recorded by the data acquisition system every second.The test was considered to be finished when:- the rebar broke, or- a significant drop of the load was observed (accompanied by increase of the displacement).The testing procedure took approximately 10 minutes.22awa:zC’)C,)wa:I—(1)2.2.3 Description of a Typical Load-Deformation Curve.A typical load-deformation curve obtained from a tension test performed on thecontrol specimen is shown on Fig.2.4.zFig.2.4 Typical load-deformation relationshpfor the control specimens.The shape of the curve is quite similar to that obtained from the tension testing of therebar alone. Two basic parts of the curve can be identified: a) the elastic part, and b) theplastic part. The elastic part, which is linear in shape, extends up to the first yield plateauwhere the plastic part begins. The plateau is located slightly above the nominal yield level(80 kN, i.e. 400 MPa). The first yield level for a particular specimen was always higher thanthe nominal yield level and varied from 83 kN (415 MPa) to 88 kN (440 MPa). Beyondthat a strain hardening is observed. The load increases up to 130 kN (650 MPa), which isequivalent to 160% the nominal load for a #15 rebar. The slope of the curve changes fromsteep at the beginning of the strain hardening to almost flat at the end. Finally, a largeELASTIC DEFORMATION DISPLACEMENT [mm]23elongation of the rebar (over 20mm), at almost constant load (final yield plateau), leads tothe failure of the rebar. The failure is preceded by necking.This kind ofthe load-deformationrelationshipwas observed in all the control specimentests. The results are summarized in Table 2.3.Specimen Ultimate Ultimate stress Ductility Failureforce glulam rebar ratio mode[kN] [MPa] [MPa]Hi 126.9 2.90 635 13 rebar yieldingH2 129.3 2.97 647 15 rebar yieldingH3 130.2 2.98 651 13 rebar yieldingH4 127.1 2.92 636 14 rebar yieldingH5 128.9 2.93 645 17 rebar yieldingIi 130.2 2.98 651 17 rebar yieldingAverage 128.8 2.95 645 15 n/aTable 2.3 Results of the control sample testing.The results of all 6 tests are highly consistent. The ultimate force varied among thespecimens slightly (+/_ 2.6%) and yielding failure occurred in the rebar during each test.The total elongation of the rebar varied more than the force (from 14mm to 24mm). Theplastic deformations were fairly large and resulted in the average ductility ratio (totalelongation/elastic elongation) of 13. This level of ductility for the connection can be considered as very desirable (for example, the bolted connections in timber have ductility ratioof 2 to 6).242.2.4 Discussion of Failure Mode I - Rebar Yielding.The dimensions of the control specimen (glulam block, glue area, rebar’s size, andembedment length) were intentionally chosen to force failure in the steel as this will alsobe the case in the real connection. However, after the treatment ofthe M.C. and temperaturesamples, other failure modes could be expected (pull-out, for example). It was importantto recognize the mechanism of failure in the control sample. The performance of the bondbetween the glue, steel, and the glulam was essential to understand the process. After thetemperature treatment the bond between glue and steel could have been affected.After the M.C. treatment the glue-woodbond could have lost its strength. Therefore all the imperfections of the bond inthe control sample had to be detected toform the basis for later comparisons withthe specimens in other samples.After the test the blocks were splitopen so the interfaces between the materials could be examined. The idealizedpicture of the connection after the failureis shown on Fig.2.5.Yielding failure in the rebar wasnamed Failure Mode I.Fig.2.5 Failure Mode I - rebaryielding.STEEL-GLUE BOND:(A) soun4(B) damaged;WOOD-GLUE BOND:(C) uninterrupted, straight woodfibres.25The top part of steel-glue bond (B) was damaged - probably due to large deformationsin the rebar at the very end of the test when high loads were experienced. The length of thispart varied from 1/10 to 1/3 of the total embedment length. The rest of the bond was notaffected (A). Also the bond between the glue and the wood fibres was not affected duringthe test (C). This picture indicates that part A provided the effective embedment lengthfor the connection. This was a very important observation.2.2.5 Instrumented Rebar Tests.To establish the stress distribution along the rebar two instrumented rebar tests wereconducted. Five strain gauges were placed on one side of the rebar. Because of the axialtension loading of the specimen, the strain was predicted to be uniform across the cross-section of the rebar. The readings from the strain gauges were recorded during the test bythe data acquisition system. The usual measurements of the loads and the displacementswere recorded as well. The location of the gauges and the stress distribution along the rebarat an external force of 80 kN (400 MPa) is shown on Fig.2.6.Unfortunately, the stress distribution close to the edge of the glue line (top part)remains uncertain. However, the damage of the glue-steel bond described in the previoussection indicates a stress concentration (marked as a dashed line) in that region. When theload was increased beyond 80 kN, the overstressed region propagated deeper into thespecimen which resulted in the breaking of the gauge closest to the top. That happenedbecause of a large deformation in the rebar exceeding the range of the gauge. Before theload reached its ultimate value, 3 out of 5 gauges were broken. The readings of the gaugesat 6 different levels of the external force (5 kN, 40 kN, 80 kN, 86 kN, 130 kN) are shownon Fig.2.7.26Cl)Ci)Cl)Fig.2.7 Stress distribution at different load levels.I’500 -400 -300 -200 -100-0-100 I 90 80rn -i r405Fig.2.6 Stress distribution along the rebarITHE SPECIMEN’S EDGE ILOAD: F .1I.0CoCowI—CobJ j......+-I3OkN /A-8kN500 IX 86 / 1:AIVI ....-40kN ..-kkN ,‘:0 II 1111111 III IIIIIIIIIII•I I I 111111110 40 80 120 160 200 240 280 320 360DISTANCE FROM REBAR’S END [mm]I I I I I I I I I400 440 48027On the graph above the heavy dotted vertical line represents the edge of the specimen.The lines shown on the right side of the dotted line represent the stress in the protrudingpart of the rebar at different load levels. These stresses are calculated as load/rebar’s area.The curves on the left side of the dotted line represent the stress in the glued part of therebar. These stresses are calculated according to Hooke’s law (strain x modulus ofelasticity).The actual yielding plateau for this specimen started at 430 MPa so there are 4 curves in theelastic region (at 5 kN, 40 kN, 80 kN, and 86 kN load levels). The 89 kN level representsthe beginning of the strain hardening phase. The ultimate level is 130 kN.About 85-90% of the total load is transferred from the rebar to the glue (and later tothe glulam) within half the embedment length (i.e. within first 200mm counting from the topface of the glulam). The point located 120mm from the bottom of the rebar (second fromleft) experienced only 50 MPa at the 89 kN load, which is only 12% of the stress in theprotruding part of the rebar. The increase of the stress at that point was observed after thelarge elongations occurred, i.e. far beyond the design load for the connection. The stress atthe very bottom of the rebar remained small even at the ultimate load showing that extrasafety is available.2.3 MOISTURE CONTENT TESTS.In this section, the influence of the changes in moisture content in the glulam member onthe behaviour of the connection is described.2.3.1 Parameters Investigated.2.3.1.1 Moisture content range.The moisture content of the glulam blockwas varied between 10% and 30%. TheM.C. was measured in two ways:a) by weight (all specimens),28b) with M.C. meter (additionally in the selected specimens).The two measurements in the selected specimens were found to be highly correlatedduring the whole treatment because moisture could easily enter the wood through theend grain. Usually, the average M.C. (measured by weight) was higher than local(measured with the meter) when the specimen was dry. The reverse happened whenthe specimen was wet. The changes in M.C. of one specimen during the treatment areshown on Fig.2.8.TIME [weeks]Fig.2.8 Example ofthe moisture content changes.The wetting part of the cycle took place in the moist room ofMaterial Laboratory.The specimens were sprinlded withwater twice a day and covered with a plastic tarpaulinto maintain high humidity in the vicinity of the specimen. The relative humidity of theroom was about 90% which was not sufficient for these tests. Immersion ofthe specimensin water, although more effective, were not considered to be realistic and didn’t takeplace. The fastest increase of M.C. was observed in first two days of wetting. Usually,C) rt\15 ..i,.•II.,0 2 4 6 8 10 12 1429the final M.C. of the specimen in the wetting part increased slightly with the number ofcycles.Under the conditions described above the wetting phase of the treatment lastedfrom 7 to 9 days.The drying of the specimens was performed in a plywood tunnel (shown on Fig.2.9) built in the basement of the Structures Laboratory. The specimens were placed inFig.2.9 Drying of the specimens.30the tunnel in two layers. Each layer and each specimen were separated to allow properventilation. To accelerate the drying a fan was installed at the inlet. Heating of the air(and the specimens) was avoided. The weighting of the specimens (to calculate M.C.)took place every 4-5 days. After the measurements the location of the specimens in thetunnel was changed - the top layer became the bottom one and the specimens placed onthe left side were moved to the right side of the tunnel. In that manner more uniformdrying conditions were achieved. The drying phase of the cycle lasted from 22 to 12 days.The rate of the drying increased slightly with the number of cycles.Because of the accelerated drying process the cracks formed on the surface of thespecimens. Most of them appeared during the first drying cycle and continued to expandthroughout the additional cycles. The increase in the drying (or wetting) rate can beexplained by increased access area for the moisture.2.3.1.2 Number of cycles.Each cycle consisted ofone wetting phase and one drying phase. Afull cycle lastedaround 3 weeks. Five cycles were conducted. However, some specimens were testedbefore completion of the full treatment so the number of cycles a particular specimenexperienced varied from 1 to 5. In this way the correlation between the number of cyclesand the change in the ultimate strength could be detected.2.3.1.3 Testing condition.The tests were performed on wet specimens as well as on dry specimens. Fivespecimens were tested after one of their wetting phases was completed. The M.C. ofthose specimens during the test varied from 28% to 30% (specimens tested wet). Twelveother specimens were tested after the drying phase was completed. Their M.C. variedfrom 11% to 13% (specimens tested dry).312.3.1.4 AngIe to the grain.Out of the total number of 17 Specimens, 5 specimens were prepared with rebarsglued in an angle of 300 to the grain (30° sample). In the remaining specimens the rebarswere glued perpendicularly to the grain. Both arrangements of the rebars can be usedin the real structures. Generally gluing perpendicularly to the grain is considered to bemore critical. The 30° sample was tested to detect possible differences.2.3.2 Discussion of Typical Load-Deformation Curves.Fig.2. 10 Load-deformation curves for the specimens after MC. treatment and a controlspecimen.Ta.ICOFfrROL SPECIMEN - FAIWRE MODE I 5 CYCLES, tested WET - FAIWRE MODE Ii::0 3010 15 25DISPLACEMENT nrrq5 CYCLES, tested WET - FAILURE MODE IIDISPLACEMENT Irmi5 CYCLES, tested WET - FAILURE MODE IIIlbU— 1W1EH0 10 15 26DISPLACEMENT m432Generally speaking, the load-deformation curves obtained from the tests after M.C.treatmentwere similar to those from the control sample. The only differences were observedin the specimens tested wet during the final stage of the tests (i.e. when large plasticdeformations in the rebar occurred). Two more failure modes were observed: Failure ModeII - pull-out, and Failure Mode III - wood split. These will be explained in the followingsections. The comparison of load-deformation curves obtained after M.C. treatment versusthe control curve is presented on Fig.2.10.All the curves have the same characteristics as that discussed in section 2.2.3 exceptthe final stage (failure). The ranges of the elastic and plastic parts of the curves are identicalto the control curves. However, a drop of stiffness within the elastic range (15%-20%) canbe observed. This last observation is valid for all M.C. treated specimens regardless of testcondition or angle to the grain. The ultimate loads oscillate around 130 kN. The ductilitylevels are also similar to those observed in the control sample (11-18).The differences in the curves begin after the ultimate load has been reached. ForMode II as well as for Mode III the drop in the load is not as sudden as for Mode I. This isbecause some resistance still exists in those specimens - friction in Mode II and shear inMode III.The similar shapes, location of characteristic points (first yielding plateau, ultimateload), and magnitudes of load and deformation can lead to the conclusion that the behaviourof the described specimens is almost identical to the control specimens and would not affectthe design range.2.3.3 Discussion of Failure Mode II- Pull-out.The pull-out failures were observed in 3 tests out of 17. They took place at the samelevel of stresses and deformations as Mode I which indicates similar mechanism of failure.Because of the large deformations in the rebar the damage of steel-glue bond propagated deeper into the specimen (zone B on Fig.2.11). In effect, the remaining length (A)33dropped below the necessary embedmentlength and the rebar started to come out ofthe glulam puffing the bonded wood fibresup (E). The fibres located above moved aswell (D). Top fibres remained bonded tothe glue coating (C) because the rebar wasfree to move in that zone due to the failureof steel-glue bond.STEEL-GLUE BOND:(A) sound,(B) dcLmcLged,WOOD-GLUE BOND:(C) uninterrupted, straight wood fibres,(D) uninterrupted, curved wood fibres,(E) interrupted, curved wood fibres(shear failure).Fig.2.1 1 Failure Mode 11-pull-out.Some friction resistance was still present after the failure. In the lower part (A) thefriction occurred on the wood-glue line. In the upper part (B) the friction between glue andthe rebar took place.2.3.4 Discussion of Failure Mode III - Wood Split.Glulam failure occurred during 2 (out of 17) tests. This Failure Mode was observedexclusively in the specimens tested wet.34STEEL-GLUE BOND:(A) sound.(B) dctmaged,WOOD-GLUE BOND:(C) uninterrupted, straight wood fibres,(D) uninterrupted, curved wood fibres.Fig.2. 12 Failure Mode III - glulam sp&4The damaged part of steel-glue bond (B) was very short and only a small movementof neighboring fibres was observed (see Fig.2.12). Probably the local variation of the glulamproperties combined with high moisture content (30%) are responsible for this kind offailure.The splitting of the glulam block originated in the middle of the specimen (close tothe rebar’s end) and propagated up along the glue line. However, the crack did not reachedthe top surface of the block but deflected sideways. This indicates that the top part of rebar(B) was free to move at that moment.352.3.5 Deformation Range.After the failure some specimens were bisected along the grain to expose the interiorof the glulam and the glue coating. In the case of pull-out failure the deformations of thewood fibres due to the movement of the rebar were observed. The distance where thosedeformations could be seen (with an unaided eye) varied significantly depending on thedirection to the grain (see Fig.2.14).In the parallel direction the fibres were curved within 50-60mm from the glue coating(60-70mm from the center of the rebar). In the perpendicular direction, a couple millimetersFig.2.13 Photo ofthe specimens K2 (Failure Mode II) and K! (Failure Mode III). Notethat the top surface ofthe specimen Ki not split.36000000 00000 do000000000 d10000000000 0000oooo 900oooo000000 00000900000009000000 O00O0O9o000000000000L o00000000000 00000 000000Fig.2.14 Defonnation range observed afterpull-outfailure.away from the coating any deformation of the wood fibres could not be detected. This canbe a good indication for future spacing rules in multi-bar connections. The spacing alongthe grain has to be much larger then across.SIDE VIEWI— 120—140mmEND VIEWL1i.”i9’21—22372.3.6 Stresses Due to M.C. Changes.The change in M.C. of the glulam causes changes in the dimensions of the block. Withrising M.C. the glulam swells and with falling M.C. the glulam shrinks. In the transversedirection (i.e. perpendicular to the grain) these dimensional changes can be very large.Within the discussed range of M.C. (1O%-30%) the height of the glulam block can vary by3-4% (12-16mm). When the dimensional changes are restricted, stresses develop in theglulam. The relationship between stress and strain is viscoelastic in nature and cannot beexplained so easily as in case of temperature related stresses (see section 2.4.3).In the case of the investigated connection, the rebar restricts the deformation of theglulam block in the direction perpendicular to the grain. In effect, opposing stresses developboth in the glulam block and in the rebar. Their magnitude and direction depends on actualM.C., time, and M.C. during manufacturing of the specimens.The gluing of the rebar took place in the laboratory. The glulam blocks had beenstored there, before cutting and drilling, for about 2 months. That allowed them to reachequilibrium moisture content (EMC) of 11-12%. The M.C. treatment described in previoussections started with wetting of the specimen. Because swelling of the glulam was restrainedby the rebar compression stresses developed in the glulam and tension stresses were createdin the rebar. Some increase of the height of the specimen also occurred due to elasticelongation of the rebar (due to the inducted stresses). When the limit of the swelling wasapproached a relaxation ofthe swelling stresses occurred. After reaching the fibre saturationpoint (M.C.=24%) the swelling stopped and the relaxation lowered the stresses to a levelslightly above zero. At the end of the wetting the glulam was subjected to compression andthe rebar to tension stresses. When the drying part of the treatment started the situationreversed. The glulam shrank and released the tension in the rebar. But shrinkage continuedand, most probably (due to lack of relaxation), the stresses were reversed. The compressionstresses were created in the rebar and tension stresses in the glulam. The hypotheticalrelationship between stresses, deformation, and moisture content during the first cycle oftreatment is presented on Fig.2.15.38COMPRESSIONoIi— r—- —/II I IIII I IIII I II N/ H I IIL_ I—+——----H EXPANSIONO‘ONTRACINM. C. %TIMEFig.2. 15 Relationship between stresses, partially restrained deformation, and moisture content injunction oftime (based on Sasaki & Yamada 1972, and Perkitny & Kingston 1972).It is very difficult to evaluate the magnitude of those stresses. At the end of dryingthey couldn’t bevery high because tension perpendicular to the grain is theweakest propertyof glulam (O.89MPa according to the code). There was no evidence that the limit strengthof the glulam was reached. Similarly, there was no evidence that the glulam reached itscompression capacity during sweffing. However, the glulam split failures could be relatedto the level of the stress developed during the treatment.392.3.7 Results and Comments.The results of moisture content tests are given in Table 2.4 and illustrated on Fig.2.16.Spec. Angle Testing No. Ultimate Ultimate Ductility Failureto conditions of force stress ratio Modethe cycles in rebargrain [kN] [MPa]Ji 90° wet 5 132.0 660 18 IKi 90° wet 5 126.3 632 16 IIK2 90° wet 5 125.5 628 10 IIIK3 90° wet 4 124.7 624 13 IIK4 90° wet 3 117.4 589 6 IIILi 90° dry 3 i27.8 639 12 IL2 90° dry 3 i29.7 649 13 IL3 90° dry 5 128.9 645 12 IL4 90° dry 5 130.5 653 17 IL5 90° dry 1 123.9 620 11 IIL6 90° dry 2 129.3 647 18 IQi 90° dry 5 127.0 635 14 IMl 30° dry 5 127.8 639 15 IM2 30° dry 1 131.0 655 19 IM3 30° dry 3 126.6 633 14 IM4 30° dry 3 128.2 641 16 IM5 30° dry 5 128.9 645 18 ITable 2.4 Results ofthe tests conducted after MC. treatment ofthe specimens.40aCl)Cl)wI(1)The main conclusion from all M.C. tests is thatthere are no significant differences within the elastic range,when compared to control sample. The elastic range is the design range for future connections. The only variation which should be taken into consideration is the decrease instiffness of the connection.Other findings are:- the ultimate strength of the specimens apparently did not depend on number ofcycles (within the range tested);- no differences in ultimate strength between 900 sample and 300 sample wereobserved;- the ultimate strength didn’t vary significantly;Fig.2.16 Graphical comparison of the ultimate load between the control sample and theMC. samples. A horizontal line at 400MPa represents the design levelfor the connection(elastic defonnations).41- the deformations developed during the treatment did not deteriorate the gluebonding;- all specimens had similar level of ductility.Although three failure modes were established, all the specimens experienced steel yieldingand, in effect, large deformations of the rebars. Therefore, yielding of the rebar may beconsidered the primary cause of failure in tested glued-in connections.2.4 TEMPERATURE TESTS.In this section, the influence of temperature changes on behaviour of the connection ispresented.2.4.1 Parameters Investigated.2.4.1.1 Temperature range.0wccFig.2.17 Temperature variation outside and inside the specimens.21TIME [daysj42The temperature of the specimens during treatment varied from -30°C to +50°C- well beyond the range expected in normal building practice. The temperature wasmeasured outside the specimens and inside the specimens (small holes were drilled in 2specimens). Usually, both readings matched after 24 hours stay in constant ambienttemperature. The temperature variations are shown on Fig.2.17.The freezing part of the cycle took place in a walk-in freezer of Dickie Dee IceCream, Coquitlam, B.C. The specimens were placed there for 2-3 days. After that, theywere transported to the Structure Lab and left in the temperature about +20°C for theperiod of at least 24 hours.To heat the specimens the plywood tunnel with sealed outlet (same as describedin section 2.3.1.1) was used. At the inlet a heater with a fan was located. To reducedrying of the specimens, shallow open tanks with water were placed close to the heater.Usually, this phase lasted 2-3 days. After the heating the specimens were left in the lab’stemperature for another 24 hours before freezing.2.4.1.2 Number of cycles.Each cycle consisted of the freezing part and the heating part. A full cycle lasted6-10 days. Five cycles were conducted. However, some specimens were tested beforecompletion of the treatment, so the number of cycles a particular specimen experiencedvaried from 1 to 5. In this way the correlation between the number of cycles and thechange in the ultimate strength could be detected.2.4.1.3 Test condition.The specimens stayed in the lab’s temperature for at least 24 hours prior to thetesting. Usually this temperature was around +20°C. The changes of the test temperature within few degrees were considered to be insignificant when compared to thetemperature range during the treatment.4324.2 Discussion of Typical Load-Deformation Curve.T0aIFig.2.18 Typical load-deformation cuivefrom the temperature treated specimen comparedto the control one.The comparison between the load-deformation curves of the temperature treatedspecimen versus a control specimen is shown on Fig.2.18.The pull-out failure (Mode U) dominated in the temperature treated sample. But twoother failure modes were observed as well. Again, all the curves were similar to those fromthe control sample except close to failure. The differences were identical to those describedfor the M.C. sample in section 2.3.2. Similar drop of stiffness within elastic range (20%) wasobserved.2.4.3 Stresses Due to the Temperature Changes.Unlike in the case of M.C. changes, the internal stresses due to the temperaturechanges are easy to establish because of their elastic nature. However, the process ofdeveloping these stresses is quite complicated. It depends on the magnitude of the temCONTROL SPECIMEN . FAILURE MODE I 5 CYCLES - FAIWRE MODE IILi 10 15 20DISPLAGEMENT (n’rj25DISPLAGEMENT [rrrr44perature change and the location of the rebar with respect to the grain. It is also timedependent. Two thermal properties of the materials are important here:linear expansion coefficient - responsible for the magnitude of the deformations, andconductivity - responsible for the time after which the deformations develop.The values for both coefficients are given in Table 2.1 at the beginning of this chapter.In case of the investigated specimens the following phases of stress development canbe observed (see Fig.2.19, Fig.2.20, and Fig.2.21):phase 0 the specimen is manufactured at about +20°C; internal forces in thespecimen are in equilibrium with zero stresses in rebar and glulam;phase 1 the specimen is moved into the freezer; the surrounding temperature dropsto -30°C; however, due to the poor conductivity of glulam, only the rebaris affected by the changed temperature and starts to contract (the protruding part of the rebar creates a large surface through which the heat canbe transmitted); the glulam resists contraction of the rebar; tension stressdevelops in the rebar and compression stress develops in the glulam;eventually, an extreme of the stresses is reached (theoretically at 4OMPafor the rebar and 0.18MPa for the glulam - see Appendix A for the calculations); this phase lasts about 1 hour;phase 2 the glulam cools and shrinks; the tension in the rebar is released; becausethe wood contracts more then steel the specimen passes the neutral (zerostresses) point and continues to shrink; compression is introduced into therebar and tension into the glulam; after about 24 hours the equilibrium isreached and the specimen doesn’t contract any more; the stresses are:1O1MPa in the rebar and 0.46MPa in the glulam; total contraction of thespecimen is 0.44mm (based on 405mm total length); these conditions lastuntil the end of the freezing (2-3 days);45phase 3 the specimen is moved out of the freezer; the temperature jumps to about+20°C; again, only the rebar is affected by the change at the verybeginning; the rebar expands creating even more compression in it andmore tension in the glulam; the stresses can reach 142MPa for the rebarand 0.65MPa for the glulam; however, the glulam starts to expand as welland the stresses at the end of this phase are, probably, slightly lower; thisphase is also very short and lasts about an hour;phase 4 the glulam expands more and more, and during the next 24 hours, theoriginal state of equilibrium is restored; zero stresses both in the rebar andin the glulam;phase 5 the specimen is moved into the heating box; the surrounding temperaturereaches +50°C; the rebar elongates immediately; the compression iscreated in the rebar and tension in the glulam; the stresses reach 25MPain the rebar and 0.11MPa in the glulam;phase 6 the glulam starts to expand and reduces the stresses; the specimen passesthe neutral point and continues to expand; the tension is created in therebar and compression in the glulam; after about 24 hours the specimenreaches its maximum elongation ofabout 0.26mm; the stresses are: 61MPafor the rebar and 0.28MPa for the glulam; this equilibrium lasts until theend of the heating;phase 7 the specimen is moved out of the box; the temperature changes to about+20°C; the rebar reacts first and tends to contract creating more compression in the glulam and more tension in the steel; the stresses can reachabout 86MPa for the rebar and 0.39MPa for the glulam (practically theyare slightly lower due to some contraction of the glulam); this phase is quiteshort and lasts about an hour;46phase 8 the glulam shrinks and releases the stresses; after about 20 hours theoriginal state of equilibrium is restored; there are no stresses in the glulamas well as in the rebar.zIFig.2. 19 Dimensional changes ofthe specimen (perpendicular to the grain) as a function oftime. The dotted line shows theoretical calculations, the solid one represents more realisticbehaviour.The changes of internal stresses due to the varying temperature are presented graphicallyon Fig.2.20 as a function of temperature and on Fig.2.21 as a function of time. The deformations of the specimen (based on 405mm length) during one cycle of the treatment areshown on Fig.2.19.TiME [hours]47t0C0‘I,aE0C,IIa0)0)LU0)C00)CC00,0) ‘..,jSaS0C)CSa0)0)LUCoC00)CSTRESS IN REBARFig.2.20 Internal stresses as a function ofthe temperature.16&.429mm2044mm*0.09=::::::::::::::. 415mm.+0 26mn.frO.llmn—1o I I I I I-40 -30 -20 -10 0 10 20 30 40 50TEMPERARTURE [ C]STRESS IN GLULAMQ6as 4j +0.17m::415mmcia0mm: ::::-j.......&......-cia .-cia :::;;;.—..:429mm:w t I60o I I I I I-40 -30 -20 -10 0 10 20 30 40TEMPERARTURE [ C]48C 040co 03c20o Qj0Ui0-04CCSTRESS IN REBARFig.2.21 Internal stresses as a function oftime.14012Q•icxII-60I fl.. RESTING\ HEATINGN\1TIME [hours]STRESS IN GLULAMCl RRESTINGFREEZING/ RESTING:HEATING-07-nR I I I0 12 24 36 48 60 72 84TIME [hours]I I I I96 108 120 132 1.449Above graphs present the calculated stresses and deformations. The relaxation of stresseswas not taken into consideration. However, due to the short time, the relaxation couldn’tbe significant. The thin solid line on Fig.2.20 presents theoretical situation when thedimensional change of the glulam starts right after the expansion (or contraction) of therebar was completed. The heavy solid line presents more likely behaviour.2.4.4 Results and Comments.The results of the temperature tests are given in Table 2.5 and illustrated on Fig.2.22.Spec. Angle No. Ultimate Ultimate Ductility Failureto Of force stress ratio Modethe cycles in rebargrain [kN] [MPa]Ni 90° 5 112.0 560 6 ifiN2 90° 3 128.2 641 10 IIN3 90° 5 127.4 637 14 IIN4 90° 1 131.3 656 12 IN5 90° 5 127.0 635 22 IIN6 90° 3 117.2 586 14 IIQ2 90° 5 128.0 640 ii ITable 2.5 Results ofthe tests conducted after the temperature treatmentof the specimens.The main conclusion from the temperature tests is, again, thatthere are no significant differences within the elastic range,when compared to the control sample. The decrease in stiffness, however, was even moresignificant than in the M.C. samples.50Lwvr FAILURE MODE: I - rebar yIelding; II - pull-out; Ill - glulam spiltwa- 600Cl)Cl)wICl)800 --I !!-‘ 7!!/..Ji /../.‘‘ / / / / / ‘ — IT! .‘ / // /-, / / A’ / / —A’ / / r /, A’ / / A’ / /A’ / / F /F A’ / / A’ / /A’ / / F /p. , / / ,. / /A’ / / t /p., / / ., / /A’ A’ / F A’4OO—7 r744eA’ / / / / / / //‘ / / /pp / / / / / / //p.. A’ / // // //////// //.4 4 .4 2200 CONTROL 1 I_VAR1NG TEMPERATUREt11JGOdeg— /,IVAyArAr/JI/J VAr/li/i VA VA VA /II I INUMBEROFCYCLES: 1 3 3 5 5 5 5Fig.2.22 Graphical comparison ofthe ultimate load between the control sample and thetemperature sample. A horizontal line at 400MPa represents the design levelforthe connection (elastic deformations).The other findings are similar to those from M.C. tests:- the ultimate strength of the specimens apparently did not depend on number ofcycles;- the ultimate strength didn’t vary significantly;- the deformations developed during the treatment did not deteriorate the gluebonding;- the ductility of the specimens varied significantly within the temperature sample butthe average ductility ratio (13) was only slightly lower then the one obtained fromthe control sample (15).Although reduced by the procedure described in section 2.4.1.1, an intensive drying of thesurface of the glulam blocks occurred during the heating part of the cycle. In effect, a lot of51cracks formed in the specimens. Certainly the cracking did influence the stiffness of thespecimen. In the real structure, however, the dimensions of glulam pieces are much largerand the temperature changes not so severe (it’s hard to imagine a situation where thetemperaturejumps by 50°Cwithin few minutes). Therefore, cracking in the real lifewouldn’tbe so serious.It can be seen that the effect of the temperature treatment on the specimens wasalmost exactly the same as the effect of the M.C. treatment (described in section 2.3.7). Inboth cases the stresses were introduced into the specimen due to the relative movement ofthe glulam versus the rebar. Both tension and compression stresses were experienced byeach part ofthe connection (i.e. the glulam and the rebar) during each cycle ofthe treatment.Although the magnitude of the stresses during the M.C. treatment was not established itcan be assumed (based on the behaviour of the specimens during the treatment and testing)that these stresses had a similar range to those during the temperature treatment. Thus, thetemperature test could also represent the M.C. changes in the specimen.2.5 CONCLUSIONS.It can be concluded, from the tests described in this chapter, thatthe changes in surrounding environment, Le. humiditv and temperature. influence the glued-inrebar connection not any more than the glulam member itselfThe treated specimens proved to have the same or similar ultimate strength, yield stress,ductility, and general behaviour when compared to the non-treated specimens. Therefore, itcan be assumed that the connections used in the real structures in different environment conditions will have similar properties and behaviour as those tested in the laboratory. Of course,all usual effects of long term loading of glulam will influence the structure and have to beconsidered.It should be emphasized that within the elastic range, which is the design range of the52connection, only the stiffness decreased (about 20%). Similar drop of glulam properties couldbe expected due to higher moisture content. All other properties, such as strength and ductility,remained unchanged.It is necessary, however, to further investigate this subject. Specially, the internal stressesdeveloped during the treatment (both M.C. and temperature) should be studied more precisely.The time seems to be a very important factor here. Also an accumulation of stresses duringlarger number of cycles should be established.The tests described in this chapter were conducted on the specimens which returned totheir original condition (except the sample tested wet). It is important, however, to establishedthe extend to which the internal stresses can be superimposed with the stresses caused by theexternal loads. Therefore, the future research should include the treatment of loaded specimens.Because, in this researchprogram, time ofthe treatment as well as the size ofthe specimenswere limited, the need of full scale, long term tests still exists. Some assumptions have beenmade (for example, occurrence of the Failure Mode III), which can be verified only by testingfull size members.The future research could be based on the temperature tests only since they would alsorepresent the M.C. changes. The temperature test is much easier to conduct and saves a lot oftime.53CHAPTER 3COMPRESSION PERPENDICULARTOTHEGRAIN.3.1 TEST OBJECTIVE.Compression perpendicular to the grain is the second weakest property of timber (tensionperpendicular to the grain being the weakest). It is quite often that this property limits thedesign of an element. Usually, this happens when large concentrated forces have to be transferred from horizontal to vertical members in the structure through bearing type connection.The most common case is a beam supported by a column or a wall, where bearing stresses aredeveloped in the beam.The objective of the tests described in this chapter was to investigate the possibility ofincreasing the bearing capacity of the glulam members, by gluing a rebar into the glulam,perpendicularly to the grain. The problem became apparent when knee joints were tested byRobert Malczyk in summer 1993 (described in chapter 1). Very high compression stresses underthe steel plates were observed during those tests- far beyond the resistance of the glulam -causing excessive overall deformations of the structure. Although other solution was appliedto improve the performance ofthe kneejoint, reinforcing ofthe glulamwas seriously considered.It was also anticipated that similar problems may occur in the future with other joint configurations. It was decided that a separate research program should be developed to obtain moreinformation about bearing reinforcing of the glulam. A part of it was incorporated into thisthesis.The tests started in June 1993 and were conducted with varying intensity untilDecember 1993.543.2 MATERIALS, TEST SETUPS, AND TEST PROCEDURES.3.2.1 Materials.The glulam used for the tests described in this chapter were some 30 years old beamsrecovered from the UBC bookstore when it was demolished. The width of the beams was178mm and so was the width of all test specimens. Both length and height of the specimensvaried according to the needs of the test setup and parameters investigated.Deformed weldable reinforcing bars (rebars) of grade 400 (yield stress = 400MPa)were used. Two sizes of rebars were tested: #10 and #15. Their nominal diameters were11.3mm and 16.0mm, and their nominal areas 100mm2and 200mm respectively (refer tosection 2.2.1 for more information).The epoxy glue IFC-SP manufactured by Industrial Formulators of Canada Ltd.,Burnaby, B.C., was used for the connections. It consists of two parts: resin and hardener.The mixing ratio is 100 to 42 (by weight) respectively. The glued specimens were cured atleast 7 days before testing in order to reach full strength of the glue.The steel plates used for the testing were 12.7mm (½”) thick, cut from flat bars ofA-36 type steel. The specified minimum tensile strength of the steel was Fu =45OMPa.SIDE VIEWS PLAN VIEW•______-[250c)Fig.3.1 Preliminaiy test setup: a) a specimen with bearingplate,b) a specimen with protruding rebai c) cross section.ci) I 250 b) I 250553.2.2 Test Setup.The Baldwin testing machine described in section 2.2,2 was used to perform all tests.This time a compression configuration of the Baldwin machine was used. In most cases themachine was set for medium range (0 to 356 kN). The high range (0 to 1780 kN) was usedonly for multi-rebar tests.Three different test arrangements were used. The preliminary tests were conductedon the glulam blocks supported uniformly on the entire area of the specimen (see Fig.3.1Fig.3.2 Specimen in the testing machine duringpreliminaiy test(note the deformations ofthe lowerpart of the glulam block)56and Fig.3.2). Because some unwanted failure (described in section 3.3.3) occurred duringthese tests, the setup was later changed to the simple supported beam arrangement (seeFig.3.3). The compression force was applied to the specimen in two different ways:a) to the steel plate located on the top surface of the specimen (the rebar was flush withthe surface of the glulam),b) directly to the rebar (the rebar was protruding by the length equal to its diameter).NC.,NC.,200 800Fig.3.3 Simple supported beam setup:a) a specimen with bearingplate,b) a specimen with protruding rebar.v—srsrrsi;200cz)NI Li iiiwm m* I5)57The third arrangement involved a blockwith twobearing plates (placedon theoppositefaces of the glulam). The rebar was going through the entire depth of the specimen and wasflush on both sides (see Fig.3.4). This setup was designed to check the likeithood ofbucklingof the rebar within the specimen.0C,,1200mm or 1150mmFig.3.4 Specimen with top and bottom bearingplates.3.2.3 Test Procedures.Three measurements were recorded by the data acquisition system: a load measuredby the load cell ofthe Baldwin, a relative displacement between the base and middle reactionbeam of the machine, and a relative displacement between the base and the top (or bottom)surface of glulam (see Fig.3.2). The displacements were measured by LVDT gauges. A setof data was recorded every second.The test was considered to be completed when:a) the protruding part of the rebar (or the plate in case of flush rebar) was completely58pressed into the glulam (after 12mm indentation),b) the protruding part of the rebar buckled,c) glulam failure occurred (shear failure or bearing failure at the support).3.3 FLUSH REBAR TESTS.In the tests described in this section a combined action ofthe glued-in rebar and the glulamin bearing was induced.3.3.1 Manufacturing of the Specimens.The manufacturing of the specimens proceeded as follows. First, the hole (or holes)was drilled in the centre of the glulam block. Then the glue was poured into the hole andthe rebar inserted. After the glue had set, the rebar was ground flush with the surface ofthe glulam. The specimens were cured 7 days before testing. Before the test the steel plate(12.7mm thick) was placed on top of the specimen in such way that the centre of the platematched the centre of the glulam block and the centre of the rebar or rebar group (to obtainequal forces in all rebars).In cases where the rebar was welded to the plate, the gluing took place after thewelding.3.3.2 Parameters Investigated.The following parameters were investigated during these tests:- rebar’s size #10 and #15- rebar’s length from 50mm to 372mm- number of rebars from ito 4- size of the plate 76x102, 102x152, 127x152, 152x124 mm59- effect of welding rebar welded to the plate or not- support conditions uniform (Fig.3.1.a), simple supported beam (Fig.3.4.a),uniform with two plates (Fig.3.4.)3.3.3 Preliminary Tests.The results ofthe tests conducted on the plain specimens and the reinforced specimensare presented in Table 3.1.Spec. Rb! Pis Pde Pdu Ste Stu Remarksmm mm kN kN MPa MPa— — —Plain glulam specimens, block size: 178x372x250mm (width x depth x length)Glm6x5a- 127x152 118 149 6.1 7.7 comp.Ibearing failureGlm6x5b - 136 170 7.0 8.8Glm6x6a- 152x152 134 157 5.8 6.8Glm6x6b - 121 153 5.2 6.6Glm6xTh *) - 152x178 164 265 6.1 9.8 bearing failureG1m370a - 152x250 142 180 3.7 4.7 comp./bearing failureGlm200a- 127x152 117 - 6.1 - no failureGlmlOOa - 112 - 58 -— — —Specimens reinforced with one #10 rebar, block size: 178x372x250mmR10-050b 50 127x152 115 136 6.0 7.0 compression failureRb-bOb 100 109 135 5.6 7.0R10-200b 200 110 137 5.7 7.1— — —Table 3.1 continues on the nextpage...60Spec. Rb! Pis Pde Pdu Ste Stu Remarksmm mm kN kN MPa MPa—- — —Specimens reinforced with one #15 rebar, block size: 178x372x250mmR15-lOOb 100 127x152 118 136 6.1 7.0 compression failureR15-lOOa “ 152x250 130 173 3 4 46R15-200a 200 135 187 3 6 49R15-200b “ 127x152 117 135 6.1 7.0R15-300a 300 152x250 130 189 3.4 5.0R15-300b “ 127x152 120 164 6.2 8.5— — —Note: Rbl - embedment length of the rebar [mm]Pls -plate size width x length [mm] (thickness 12.7mm)Pde - maximum load within elastic region (approximate) [kN]Pdu- ultimate load [kNJSte - equivalent stress under the plate (within elastic region) JMPa]Stu- ultimate equivalent stress under the plate [MPa]*) length of the block increased to 600mmTable 3.1 Results ofthe preliminaiy tests with bearingplate andflush rebar.The main goal of these tests was to quantify the difference in bearing capacity betweenplain glulam specimens and the specimens reinforced with glued-in rebars. Unfortunately,the size of the specimens combined with uniform support created compression failure of theglulam in most cases, not bearing failure under the plate. However, it was noticed that thecompression failure occurred beneath the reinforced part ofthe specimen (see Fig.3.2). Thisfact indicated that the glued-in rebar spread the bearing stresses from under the plate deeperinto the glulam. It was suggested that larger glulam blocks would be needed to obtain evidentbearing failures. It was confirmed during a test Glm6xTh (with the length ofthe glulam blockincreased to 660mm).613.3.4 “Beam” Setup Tests.During the tests described in previous section the bearing plates were located directlyabove the support surface of the specimens. In cases where the rebar was glued-in, thebearing stresses were “moved” into the unreinforced part of the glulam. That created relatively short (especially where the rebar was 300mm long), highly stressed zone, where thesupport stresses interfered with the stresses transferred by the rebar. To avoid that situation,itwas decided to move the support location away from the reinforced section of the specimen.To achieve this a short simple supported beam setup was chosen.An initial clearance between the base and the bottom face of the beam was set to51mm (2”). That clearance allowed the beam to deflect freely at midspan, i.e. at the locationof the bearing reinforcement. The clear span was kept approximately twice the depth of thespecimen so the influence of compression stresses at the supports on the midspan crosssection could be diminished. The usual support length was 200mm. The total length of thespecimen was calculated according to the empirical formula:total length> 2 x (rebar’s embedment length + support length)The maximum length of the specimen that could be tested in the Baldwin machine was1300mm. Themajority ofthe specimenshad dimensions 178x372x1200mm. Three specimens(Bm15, Bm17, Bm20) were 178x290x1150mm and one (#1) was 178x372x960mm.The results of the tests conducted on “beam specimens” without reinforcement andwith one #15 rebar glued-in are presented in Table 3.2.62Spec. Rbl Pde Pdf Pdu Ste Stf Stu [ Remarksmm kN kN kN MPa MPa MPaPlain glulam specimens with bearing platesBm5- 60 - 137 7.7 - 17.5 paBm6- 110- 215 7.1 - 13.9 pb#1 - 175 - 295 4.9 - 8.2 * PCSpecimens reinforced with one #15 rebar, bearing plate size: 76x102mmBm9 100 110 116.5 133 14.2 15.0 17.2Bm7 200 143 177.5 140 18.4 22.9 18.1Bm17 148 178.5 160 19.1 23.0 20.6Bm18 155 181.2 170 20.0 23.4 21.9Bm19 150 178.6 155 19.3 23.0 20.0Bm2 300 150 220.0 162 19.3 28.4 20.9Bm3 130 181.0 145 16.8 24.7 18.7Bm12 140 244.0 240 18.1 31.5 31.0 widBm13 372 150 264.6 220 19.3 34.1 28.4 gtBm15 290 140 203.1 155 18.1 26.2 20.0 gtTable 3.2 continues on the nextpage...63Spec. Rb! Pde Pdf Pdu Ste Stf Stu Remarksmm kN kN kN MPa MPa MPaSpecimens reinforced with one #15 rebar, bearing plate size: 102x152mmBm20 100 180 206.0 225 11.6 13.3 14.5Bm8 200 200 227.1 238 12.9 14.6 15.4Bml 300 190 233.5 215 12.3 15.1 13.9Bmll 180 331.5 300 11.6 21.4 19.3 w!dNote: Rbl - embedment length of the rebar [mm]Pde- maximum load within elastic region (approximate) [kN]Pdf- load at failure of reinforced bearing area [kN]Pdu- load at the end of the test [kN]Ste - equivalent stress under the plate (within elastic region) [MPa]Stf - equivalent stress under the plate at failure [MPa]Stu - equivalent stress under the plate at the end of the test [MPa]p - plate size: a) 76x102mm, b) lO2xlS2mm, c) 178x202mm- suppo# length 240mmwld - rebar welded to the plategt - rebar going through the entire depth of the specimenTable 3.2 Results ofthe tests conducted on beam specimens with bearingplate andflushrebar.The comparison of load-deformation curves obtained during these tests is presentedon Fig.3.5 for 76x102mm plate and on Fig.3.6 for 102x152mm plate.The indentation of the plate with respect to the top surface of the glu!am is presentedon the abscissa. The applied load and the corresponding bearing stress under the plate isshown on the ordinate. The stress was calculated as the applied force divided by the areaof the plate (uniform stress distribution under the plate was assumed). It represents a realstress in the g!u!am only in case of non-reinforced specimen. When the rebar is present thisvalue represents the combined action of the glu!am and the rebar. Thus, the values of stressfrom Fig.3.5 and Fig.3.6 cannot be compared directly.64AnFig.3.5 Load-defonnation curves for the specimens with 76x102mm plate.Fig.3.6 Load-deformation cwvesfor the specimens with 102x152mm plate.a(nU)I300250‘m--:.... j. ‘[158AR. 300 mm• I#15n,3mmI: / .35.j 30y 25150 . . 2015100 •....1 II5,200mm1 1058A100rnm•.IaU)U)IU -0 2I I I I I I I P I6 8 10 12 14DENTATION OF PLATE [mm]16 18I220-JJU#15BAIR,2OCnm #15BAJR,30(Ynm250/1 .15200150 1 10im50/PLfiJN GLU1 /1’ .5-V0 2 4 6 8 • 10 • 12 14 • 18 18 20INDENTA11ON OF PLATE [mm]65From the presented curves it is obvious that even the shortest rebar (100mm) improvesthe bearing capacity of the glulam significantly. By gluing the rebar the ultimate capacitycan be increased (almost 2 times) and, what is even more important, the elastic region ofthe curve is 2.5 times larger when compared to the plain glulam specimen. This can beachieved by using just 200mm long rebar. The improvement in stiffness is also very significant- from 1.5 times for 100mm rebar up to 2.5 times for 372mm rebar.Of course, the relative increase of the elastic or ultimate load depends on the size ofthe plate used. The larger the plate the smaller is an influence of a single rebar.Fig.3.7 Comparison ofL-D cuivesforthe specimens with 76x102mm plates:a) plain glulam, b) 300mm rebar not welded to theplate,c) 300mm rebar welded to theplate.a.00ILlC,)IINDENTA11ON OF PLATE [mm]66Another interesting observation was made when load-deformation curves of thespecimenswith the rebars welded to the platewere comparedwith the non-weldedspecimensof the same rebar’s length (see Fig.3.7 and Fig.3.8). Although the elastic load is the same,an additional increase in initial stiffness is observed. Also the ultimate load was over 20%larger for welded specimen than for non-welded. The effect of welding wasn’t of muchinterest at that stage and it wasn’t investigated further because it would require an extraoperation for the manufacturers. However, it appears that providing better contact betweenthe rebar and the plate can additionally increase the bearing capacity of reinforced glulam.Fig.3.8 Comparison ofL-D curves for the specimens with 102x152mm plates:a) plain glulam, b) 300mm rebar not welded to theplate,c) 300mm rebar welded to theplate.00)(0IiiU)IINDENTATION OF PLATE [mm]673.3.5 “Two Plates” Tests.It is possible that the bearing plates are required on opposite sides of the glulammember at the same time. That is the case of a column to beam connection, where the beamis continuous over the joint and the structure has more than one story. The test setupdescribed in this section (see Fig.3.4) was intended to verify if only one rebar can be used toreinforce the compression resistance of the glulam member on both sides.These tests were designed to check if the glulam provides sufficient resistance againstbuckling of the rebar.The glulam block was of the size as the “beam” specimen. A #15 rebar was goingthrough the entire depth of the glulam and was flush on both sides. Two steel plates ofidentical size were placed at the rebar’s location. A compression load was applied until oneof the plates was completely pressed into the glulam. The results of the tests are presentedin Table 3.3 and on Fig.3.9.Spec. Gis Rbl Pis Pde Pdf Ste Stf Remarksmm mm mm kN kN MPa MPa— —Bm14 372x1200 372 76x102 142 259.3 18.3 33.9 gtBm16 290x1150 290 102x152 130 366.5 8.4 22.9 gt— —Note: Gis - glulam size: 178 x depth x length [mm]Rbl- embedment length of #15 rebar [mm]Pls - plate size: width x length [mm] (thickness 12.7mm)Pde- maximum load within elastic region (approximate) [kN]Pdf - load at failure of reinforced bearing area [kN]Ste - equivalent sfress under the plate (within elastic region) [MPa]Stf - equivalent stress under the plate atfailure [MPa]gt - rebar going through the entire depth of the specimenTable 3.3 Results of the tests with two bearingplates andflush rebar.68Fig.3.9 L-D curvesfor the specimens with two bearingplates.The results are very much similar to those obtained from the “beam” tests for the samelength of the rebar, i.e. similar increase in bearing capacity of the glulam was observed.The buckling of the rebar did not take place.TOTAL INDENTATION OP BOTh PLATES [mm]693.3.6 MuIti-rebar Tests.Some specimens with multi-rebar reinforcement were tested as well. Three or four#15 rebars were glued-in under the plate. The “beam” setup was used for testing. The resultsof multi-rebar tests are presented in Table 3.4 and on Fig.3.1O.Spec. Gis Nr Rbl Pis Pde Pdf Pdu Ste Stf Stu Remar.mm mm mm kN kN kN MPa MPa MPa— — — —#1 372x960 - - 178x202 175- 295 4.9 - 8.2 *— — — —Bm21 410x1300 3 200 152x204 410 nf 455 13.2 - 14.7 sfbfBm23 372x1200 3 200 “ 388 419.9 330 12.5 13.5 10.6 sfBm22 3 200 “ 400 433.3 340 12.9 14.0 11.0 sbr#2 372x960 4 300 178x202 480 nf 586 13.3 - 16.3 * wld bfNote: Gis- glulam size: 178 x depth x length [mm] —Nr - number ofrebarsRb! - embedment length of the rebars [mm]Pls- plate size width x length [mm] (thickness 12.7mm)Pde- maximum load within elastic region (approximate) [kN]Pdf - load at failure of reinforced bearing area [kN]Pdu- load at the end of the test [kN]Ste - equivalent stress under the plate (within elastic region) [MPa]Stf - equivalent stress under the plate atfailure [MPa]Stu - equivalent stress under the plate at the end of the test [MPa]nf -failure underplate not observedsf - shearfailure of the beambf - bearingfailure of the supportssbr - shear reinforcement and supports bearing reinforcement added- supports length 240mmwld - rebars welded to the plateTable 3.4 Results of the multi-rebar tests.70Fig.3. 10 Comparison ofL-D curves for theplain glulam specimen and the multi-rebar speci,nens.A large increase in load carrying capacity was observed. It was proportional to theincrease in the plate’s surface and total rebars’ length. The equivalent stress was comparableto that obtainedwith single rebar test. Unfortunately, the limitations ofthe Baldwin machinedidn’t allow larger specimens to be tested and shear failures occurred in two specimens aswell as bearing failures at the supports. To prevent that, the specimen Bm22 was reinforcedby gluing one #15 rebar 350mm long at each support (perpendicularly to the grain). Thoserebars acted as bearing and shear reinforcement simultaneously. This reinforcement allowedfor a clear picture of failure at the location of the compression force since the shear capacityof the glulam was increased.One of the specimens (#2) had its rebars welded to the plate. In that case the increaseof the stiffness was much more significant than in the specimens described in section 3.3.4(about 8 times larger than the plain glulam specimen and 5 times larger than non-weldedspecimens).INDENTA11ON OF PLATE [mm]713.4 PROTRUDING REBAR TESTS.In the tests described in this section the rebars glued-in perpendicularly to the grain wereloaded in compression. There were no bearing on the glulam. The load was carried exclusivelyby the glued-in rebars.The main goal was to establish the minimum embedment length of the glued-in rebarsloaded in compression.3.4.1 Parameters Investigated.The following parameters were investigated during these tests:- size of the rebar #10 and #15- embedment length of the rebar from 50mm to 400mm3.4.2 Preliminary Tests.The preliminarytests were conducted on specimens uniformly supported on the wholebottom surface of the glulam block (ref. to Fig.3. 1). The rebar was glued-in at the centre ofthe specimen. The protruding part of the rebar was kept equal to the diameter of the rebar.The size of the glulam blocks was 178x372x250mm in majority of the specimens. Only twospecimens (R15-400a and R15-400b) had the depth increased from 372mm to 420mm toaccommodate 400mm long rebars.The results of the preliminary tests are given in Table 3.5.72Sp Rbl Pde Pdf Sre Srf Sgf Sbf Remarksmm kN kN MPa MPa MPa MPa— —Specimens reinforced with one #10 rebar, protruding length 11mmR10-050c 50 32 33 320 330 0.7 16.5 embedment failureR10-050d “ 32 32 320 320 0.7 16.0R10-lOOc 100 55 64 550 640 1.4 16.0R10-lOOd “ 54 56 540 560 1.3 14.0R10-150a 150 54 79 540 790 1.8 13.2 rebar bucklingR10-150b “ 49 81 490 810 1.8 13.5R10-150e “ 54 86 540 860 1.9 14.4R10-150c1 “ 49 74 490 740 1.7 12.4R10-200c 200 44 79 440 790 1.8 9.9R10-200d 46 83 460 830 1.9 10.4R10-250a 250 45 86 450 860 1.9 8.6R10-250b “ 47 78 470 780 1.8 7.8— —Specimens reinforced with one #15 rebar, protruding length 16mmR15-lOOc 100 56 58 280 290 1.3 9.7 embedment failureR15-lOOd 59 62 295 310 1.4 10.4R15-150c 150 85 96 425 480 2.2 10.7R15-150d “ 82 88 410 440 2.0 9.8R15-200c 200 86 109 430 545 2.4 9.1R15-200d 89 113 445 565 2.5 9.4R15-250c 250 91 126 455 630 2.8 8.4R15-250d “ 91 130 455 650 2.9 8.7 rebar bucklingR15-320a 300 92 140 460 700 3.1 7.8— — —Table 3.5 continues on the nextpage...73Spec. Rb! Pde Pdf Sre Srf Sgf Sbf Remarksmm kN kN MPa MPa MPa MPa— —R15-300c 300 78 115 390 575 2.6 6.4 rebar bucklingR15-300d “ 91 143 455 715 3.2 8.0R15-350a 350 88 147 440 735 3.3 7.0R15-350b “ 89 134 445 670 3.0 6.4R15-400a * 400 91 151 455 755 3.4 6.3R15-400b * “ 93 152 465 760 34 63— —Note: Rbl- embedment length of the rebar [mm]Pde- maximum load within elastic region (approximate) [kN]Pdf - rebar’s failure load [kN]Sre- stress in the rebar at maximum load within elastic region [MPaJSrf - stress in the rebar atfailure load [MPaJSgf - stress in the glulam atfailure load [MPa]Sbf - bond stress between wood and glue atfailure [MPa]- depth increased to 420mmTable 3.5 Results ofthepreliminaiy tests with protruding rebars.The embedment length established on the basis of these tests was 150mm for #10rebar and 250mm for #15 rebar. In the specimens where the length of the rebar increasedthe embedment length the failure occurred due to buckling of the rebar (see Fig.3.11 andFig.3.12). Where the length was smaller the shear around the glue coating (in the glulam)occurred without any signs of buckling of the rebar (see Fig.3.12).The failure load of the specimens with full embedment length was larger (up to 18%)than the ultimate load for the same rebar tested in tension (compare with the values presented in section 2.2.1). Although the boundary between the elastic region and the plasticregion in the compression test wasn’t so distinct as in the tension test it can be assumed thatit was located at the same load level.74VtV V‘22!V. ‘V2t’’• •‘•l1IS 3OOHnimM.t‘.1=V_____VIVV•preliminarv tesL Note budded rebars.V,.V.rFig.3.12 Specimens after thepreliminaiy test. Note buckled rebar in the specimen with300mm long rebar and glulam deformations around the rebar in the specimenswith 250mm rebars.753.4.3 “Beam” Setup Tests.Although the preliminary tests gave rather consistent results it was decided to verifythem using the “beam” setup. Again, the reason for this was to eliminate possible influenceof the support condition - specially in the specimens with long rebars.The tests were conducted on the glulam beams of uniform size (178x372x1200mm),supported on 200mm long supports at each end. Only #15 rebars were tested. The protruding length of the rebar was equal to its diameter (i.e. 16mm).The results of these tests are presented in Table 3.6. The typical load-deformationcurves obtained during the tests are shown on Fig.3.13.Spec. Rbl Pde Pdf Sre Srf Sbf Remarksmm kN kN MPa MPa MPaBm24 100 60 65.5 300 328 10.9 embedment failureBm25 “ 55 628 275 314 105Bm26 “ 57 62.8 285 314 10.5Bm27 200 73 111.4 365 557 9.3Bm28 “ 88 1120 440 560 94Bm29 “ 90 117.0 450 585 9.8Bm30 300 93 148.0 465 730 8.2 rebar bucklingBm31 “ 92 1543 460 772 86Bm32 “ 87 149.1 435 746 8.3Note: Rbl - embedment length of the rebar [mm]Pde - maximum load within elastic region (approximate) [kN]Pdf -failure load [kN]Sre - stress in the rebar at maximum load within elastic region [MPa]Srf - stress in the rebar atfailure load [MPa]Sbf - bond stress between wood and glue atfailure [MPa]Table 3.6 Results ofthe beam tests with theprotruding rebars.76wILlzCl)Cl)Ui0)Fig.3. 13 Typical load-deformation curvesfor the “beam” specimens with protruding rebars.It can be seen from Table 3.6 that the results conducted on “beam” specimens confirmed results from the preliminary tests: the minimum embedment length for # l5rebarloaded in compressionwas greater than 200mm and smaller than 300mm (embedment lengthestablished previously was 250mm). All specimens with the rebars longer than 200mm failedby buckling of top part of the rebar.It was also noticed that after the embedment failure (100mm and 200mm rebars) thespecimens were able to carry some load due to the friction around sheared surface and somebearing under the rebar. Although this carrying capacity decreased as the indentationincreased but finally it stabilized at 40% of the failure load.The plot of the failure loads for all specimens- regardless of the support condition -is presented on Fig.3.14.INDENTATiON OF REBAR [mm]77Fig.3.14 Plot ofthe failure load versus the embedment length for the specimens with protnsding rebars.It should be emphasized that placing the protruding rebar (or rebars) at the supporthas several advantages:- involves larger volume of the glulam in carrying perpendicular loads than commonbearing;- establishes clear support location;- elevates the glulam member over the support what can provide better access of air andeasier drying of the glulam;- eliminates contact with often moist supports;- increases shear capacity of the glulam at location of high shear stresses (providing thelength of the rebar is sufficient);1112100*az)CrC] mWi 143 b6040...rn X #l5bar3lObarS20 jz•a •50 100 150 200 250 300 350 400 450 500REBAR’S LENGTH [mm]78- allows reliable and convenient connection method with supporting structure (speciallyin case of concrete walls or columns);- can provide some resistance against uplift forces (tension perpendicular to the grain).There are also some disadvantages:- the rebars sticking out of the beam may cause some transportation (or storing) problems;- the faces of the beam (top and bottom) cannot be switched;- the connection is very sensitive to high temperatures and fire when not protected.3.5 GUIDELINES FOR BEARING REINFORCEMENTWITH #15 REBARS.The tests described in this chapter proved that gluing the rebars perpendicularly to thegrain at the location of high concentrated compression force is a very effective way to increasethe bearing capacity of the glulam member.The following could be used as guidelines for bearing reinforcement where #15 rebarsare used:- the rebars should be glued-in perpendicularly to the grain;- the hole diameter should be 19mm (¾”);- for multi-rebar connection the spacing parallel to the grain should be 100mm (4”), and75mm (3”) perpendicularly to the grain;- the steel plate should provide uniform distribution of the compression force to all glued-inrebars;- the protruding part of the rebar should be shorter than the diameter of the rebar;- the resistance of the glulam under the plate should be taken not more than the specifiedvalue in compression perpendicular to the grain;- theelastic resistanceofa single rebar300mm or longer maybetaken as its nominal resistance79i.e. 8OkN (400MPa);- the elastic resistance of a single rebar shorter than 300mm but longer than 100mm may becalculated according to the empirical formula:resistance [kN] = 35 + 0.15 x embedment length [mm]- the combined resistance of the plate and the rebar may bet ken as a summarized resistanceof the plate and the rebar acting separately;- the elastic resistance of up to 4 rebars may be assumed to be directly proportional to thenumber of rebars used.3.6 CONCLUSIONS.As can be seen from the tests presented in this chapter, the gluing of the rebar perpendicularly to the grain is a very effective way of increasing the bearing capacity of the glulam.According to obtained results the failure load and the elastic range of the bearing connectioncan be doubled with respect to the plain glulam bearing. The increase of the stiffness is evenmore significant and can exceed 2.5 times.Although the tests conducted at UBCgave a relatively clear picture of the possible benefitscoming from reinforcing of the glulam they didn’t answer all questions related to that topic.Here are some of the unknowns that should be established during future research:- spacing of the rebars,- the group effect,- evaluation of shear stresses around the rebar,- evaluation of bearing stresses under the rebar,- contribution of bearing reinforcement to the shear capacity of the glulam.80CHAPTER 4BEAM-TO-COLUMN JOINTS.4.1 TEST OBJECTIVES.A moment resisting frame is widely used as a primary structural system for low-rise officeor apartment buildings. It provides significant savings of materials when compared to staticallydeterminate structures. It has also ability to resist lateral loads without additional bracing elements.Unfortunately, glulam has not generally been successfully used for this type of structurein Canada. The main reason for this was a lack of a reliable connection methods which wouldbe able to transfer bending moments under reversed loading conditions, i.e. in case of windloads or seismic loads. Theglued-in rebar connection is onepossible means to solve that problem.The primary objective ofthe tests described in this chapterwas to confirm that expectation.The other objectives were:- develop manufacturing and design techniques for the beam-to-column joints in amoment resisting frame,- investigate the behaviour of glued-in rebar connection under cyclic loading,- verify the effectiveness of the reinforcement perpendicular to the grain asdescribed in chapter 3,• investigate the behaviour of perpendicularly glued rebars under reversed loads(tension/compression).The tests were prepared and conducted at the University of British Columbia betweenJanuary and May 1994.814.2 PRELIMINARY ANALYSIS OF THE FRAME FORA MULTI-STOREY BUILDING.It was decided that the tested connections should imitate the real structure’s joints asclosely as possible. To achieve this, a specific structure was chosen to provide background forlater comparison. It was a 3-storey office building with a steel moment resisting frame as a mainstructural system. The complete design of that building can be found in “Metric Design Notesfor Limit States Design in Steel” (Canadian Institute of Steel Construction, 1979).The goal was to design a similar building using glulam members assuming that a momentresisting joint can be created. Then, a specffic beam-to-column connection was chosen forfurther investigation. It was designed using the glued-in rebar idea (i.e. the rebars glued intothe glulam providing a rigid connection between the beam and the column in the structure). Itwas then manufactured and tested in the laboratory. The last step was to compare the strengthdata collected during the tests with the design loads and deformations obtained from the analysis.A full analysis of the building’s frame can be found in Appendix B. Essential informationand assumptions, as well as the results of the analysis are presented below for convenience.4.2.1 Basic Dimensions.The building was rectangular in shape, 63.Om long and 36.Om wide. The main framewas placed in the longitudinal direction. The span was 9.Om and the spacing 4.5m. Therequirements for the clear heights were: 3.Om for the ground floor, and 2.7m for 2nd and3rd floors. It was found that a storey height equal to 4.Om would satisfy those requirements.It was assumed that the lateral forces in the direction perpendicular to the plane ofthe frame were resisted by bracing between frames.A plan and a cross section of the building are shown on Fig.4. 1.82Fig.4. 1 Overall dimensions ofthe building.4.2.2 Loads and Load Combinations.The following specified loads were calculated to act on the frame:1. floor dead load 2.00 kN/m22. roof dead load 1.15 kN/m2‘4,‘4,—JI I I I I[‘ I 1’ I I I. ‘1 I I I I 1II I I II I I I I-I I I I I.I II I I II I I II I I I I.I I II I I.—. I. I. .1 I II 63mPLANI 11 IIIU U,jII I 1CROSS SECTION833. floor live load 1.90 kN/m24. snow load 2.30 kN/m25. wind load on the windward wall (pressure) 0.43 kN/m26. wind on the leeward wall (suction) 0.32 kN/m7. wind load on the roof (suction) 0.75 kN/m28. first storey seismic force 50 kN9. second storey seismic force 101 kN10. roof seismic force 131 kNThe following factored load combinations were included in the analysis:A. 1.25x(Dead Loads) + 1.5x(Live Loads)B. 1.25x(Dead Loads) + 1.5x(Wind Loads)C. 0.85x(Dead Loads) + 1.5x(Wind Loads)D. 1.25x(Dead Loads) + 1.Ox(Seismic Forces)E. 0.85x(Dead Loads) + 1.Ox(Seismic Forces)F. 1.25x(Dead Loads) + 0.7x{1.5x(Live Loads) + 1.5x(Wind Loads)}G. 1.25x(Dead Loads) + 0.7x{1.5x(Live Loads) + 1.Ox(Seismic Forces)}A load combination giving the largest internal forces in the element was chosen to designthat element.4.2.3 Design Forces.The internal maximum forces governing the design of the elements of the frame aresummarized in Table 4.1.84element internal force governing design load combinationfirst storey beam bending moment 235 kNm Gsecond storey beam bending moment 191 kNm Groof beam bending moment 172 kNm AflrstXstoreyolumn bending moment 159 kNm Daxial force 283 kNsecond storey column bending moment 78 kNm Gaxial force 168 kNthird storey column bending moment 80 kNm Gaxial force 73 kNTable 4.1 Internalforces governing the design ofthe elements.4.2.4 Element Sizes.The sizes of the elements and their resistance are presented in Table 4.2.moment resistance axial resistance sizeelement [kNm] [kN] [mm]first storey beam 250 130x684second storey beam 202- 130x608roof beam 180 130x570first storey column 221 1093 175x532second storey column 99 439 130x418third storey column 99 449 130x418Table 4.2 Calculated resistance and size ofthe elements.854.2.5 Configuration of the Elements.There are two possible arrangements of the beams and the columns in the discussedframe. The first one in which the columns are continuous through the joints and the secondone is where the beams are continuous. The second option was chosen for investigation inthis research.In this situation the beams required some moment resisting splice connections to makethem continuous over the entire length of the building. This kind of the joint, using glued-inrebars idea, was described in chapter 1. The proposed locations of those joints are shownon Fig.4.2 and represent places with small moments.SPLICE CONNECTIONS- Q) ( W Q) () (),i- ,-;.‘, (Th ,-;. ‘w..L) ‘.1) i) W jJ ‘Lj_I 9f I 91 91 9m1Fig.4.2 Locations ofthe splice joints in the beams.The splice connections have one more very important function. It is fairly easy toshape the joint in that way so the steel elements will yield at certain load to create a plastichinge at that location (for example by varying the thickness of the steel plates used in theconnection). From the earthquake design point ofview it is much safer that the plastic hingesare created in the beams than in the columns (so called “weak beams” approach). When anearthquake occurs plastic hinges in the beams do not create collapse of the structure butallow accommodation of large movements in a relatively safe way.864.3 TEST SETUP.It was decided that a connection of a second storey column to a second storey beam wouldbe tested. That was an arbitrary choice and any other would be as good as this one. An axialforce, a bending moment, and a shear force acted in that connection in a real structure. However,reflecting all those forces in a test would be not only difficult, but could also produce unclearpicture of the failure. Because the main goal of the test was to prove the ability of the glued-inrebar connection to transfer bending moments it was decided to design the test setup in suchway that it would create moments similar to those in a real joint. Furthermore, the moment inthe beam was not of much interest because the beam was designed to be continuous over thejoint. In effect, the moment in the column was the only one related to the real joint’s forces.3500REACTIONLOAD CELLGLULAM COLUMN APPLIED_FORCE130x418x1700mm —BLOCKING_____________________r°°°°°°°°Q°°°°°°°°° I1500 1500Fig.4.3 Overall dimensions ofthe reaction frame and the specimen.87Other moments (in the beam) resulted from the moment in the column and their magnitudewas about 30% of those in the real joint. The magnitude of the shear forces created in the beamwas about 40% of the real shear. The shear force in the real column was close to zero but itwasn’t possible to avoid shear in the colunm during the test. In effect, the shear force in thetest column was several times larger than the shear force in the real joint. An axial force in thecolumn was not simulated at that stage of testing.A total of3 tests were performed. The testing took place in a steel reaction frame erectedin the Structures Laboratory at UBC. The dimensions of the specimens were designed to fitthe available space inside the frame and reflect the column moments as closely as possible. Theoverall dimensions of the frame and the specimen are shown on Fig.4.3.The load was applied by a hydraulic jack equipped with a load cell. The working range ofthe jack was 0-445kN (0-100,000lb) and its stoke was 152mm (6”). The load from the jack wastransferred to the column by two channels bolted together on opposite sides of the column andpin connected to thejack. The bearing length between the channels and the column was 200mm.The beam was clamped down on both ends to the reaction frame. The bottom supportswere provided by glulam blocks which created a 300mm bearing length. The top supports wereprovided by 200mm wide channels bolted down to the frame by 1½” threaded bars. The lateralmovement of the specimen was prevented by blocking between the frame and both ends of thebeam.The out-of-plane movement of the specimen was restricted at 3 locations. The angleswere bolted to the frame in front of and behind the beam at both ends. They were placed atthe mid-height of the beam. Additionally, out-of-plane bracing was installed in the frame1300mm above the top face of the beam (300mm below the applied force). The bracing allowedfree in-plane movement ofthe column (within thejack’s stroke). The bracing is shown on Fig.4.4.88The load was sequentially applied in both directions (i.e. pushing and pulling) to obtainreversing of the moments in the joint. The loading procedures varied slightly among the testsand they will be described in the sections related to the specific test (4.6 through 4.8). The forcesacting on the specimen during the test, as well as the resulted moment and shear diagrams, areshown on Fig.4.5.The following data were collected during the test:- applied force measured by the load cell in the hydraulic jack,- displacement of the jack’s cylinder,- three measurements of an overall movement of the specimen (LVDT 7 through 9),- six displacements of various parts of the connection (LVDT 1 through 6),- two displacements of the columnversus the beam (LVDT 10 and LVDT 11) - these readingswere recorded only during the last test (T-3).Fig.4.4 Photograph ofthe specimen T-3 in the reaction frame. Note a bracingjust below thehydraulicjack89rTr?1 RR=O.633xPfRH ZELFEZJf 1.5mRL =O.633xP1.625 1.625MOMENT [kNmJFig.4.5 Loac4 reactions, and resulting moment and shear diagrams for the specimen. Note,only one direction ofthe load is shown.All gauges, except LVDT 7, were placed on the front side of the glulam members. LVDT7 was placed on top of the column at the center of its cross section. The general location of thedisplacement gauges is shown on Fig.4.6 and Fig.4.7.SHEAR [kN]rL__90REACTION FRAMEh-”-jfAPPLIED FORCELVDT7rLJ(I 750 750750 750I—,.’1-fri -ooooo]obo5oI 1500 1500 IFig.4.6 Locations ofthe gauges monitoring the overall movement ofthe specimen.Fig.4.7 Locations ofthe gauges within the connection region.The distances S1,S2,S4 shown on Fig.4.7 varied according to the size of the plates usedin different specimens. They will be quoted later when the specific tests are described.914.4 MATERIALS.GLULAM.The glulam members were donated by Surrey Laminated Products Ltd, Surrey, B.C. Theywere beams and columns, grade 24-f made from Douglas Fir lumber. The elements wererectangular in shape and had the following dimensions:columns: 130x418x1700mmbeams: 130x608x3250mmREBARS.Grade 400 (yield stress = 400MPa), deformed weldable reinforcing bars (rebars) wereused. Two sizes of rebars were used: #15 and #20. Their nominal diameters were 16.0mm and19.5mm respectively. Their nominal areaswere 200mm and 300mm2respectively. The nominaldimensions are equivalent to those of a plain round bar having the same mass per meter as thedeformed bar. The results of the tension tests for #15 rebars have been afready presented inchapter 2 (Table 2.2 and Fig.2.2). The results of the tension tests of #20 rebars are summarizedin Table 4.3.yield ultimateload stress load stress[kN] [MPa] [kN] [MPaJAverage 144.2 481 202.2 674(n=5)Table 4.3 Results of tension testsperformed on #20 rebars.All rebars used in the tests were sandblasted prior to gluing.92PLATES.The steel plates used for the testing were cut from flat bars of A-36 type steel (grade300W). The specified minimum tensile strength of the steel was Fu=45OMPa. The thicknessof the plates was 12.7mm (½”). There was only one location in the joint where the thicker plates(¾” or 1”) were used in two last tests.BOLTS.The bolts used were imperial size 5/8, SAE-Grade 8, medium carbon alloy steel whichwere equivalent to grade ASTM-A490 specified in the steel design code (CAN3-S16.1-M89).Their characteristics were as follows:nominal diameter [mm]: 15.88nominal area [mm2]: 198specified minimum tensile strength, F [MPa]: 1035 (150,000psi)factored shear resistance, [kN/bolt]: 57.6 (12,9441b)factored tensile resistance, Tr [kNlbolt]: 103 (23,1461b)The bolts’ length were 24.5mm (1”).GLUE.The epoxy glue IFC-SP was used for all connections. It is manufactured by IndustrialFormulators of Canada Ltd., Bumaby, B.C. It consists of two parts: resin and hardener. Themixing ratio is 100 to 42 (by weight) respectively.An average amount of glue use for gluing the rebars into the holes was as follows:#15 0.20 g/mm length of the rebar,#20 0.33 g/mm length of the rebar.The pot life of the glue varied from ito 3 hours depending on the outside temperature and thevolume of the glue used. The glued specimens were cured at least 7 days before testing in orderto reach full strength of the glue.934.5 JOINT DESIGN.In this section, the design of the experimental connection is presented.4.5.1 Joint Configuration.The forces taking part in transferring the bending moment from the column to thebeam are schematically shown on Fig.4.8.Fig.4.8 Moment transferfrom the column to the beam.APPLIED LOADISTIFFNED ANGLESbolted— column & beam platesBEAM PLATEREBARSperpendtcularlyto the grainBEAM94It was assumed that the internal forces would be transferred from the column to theinclined, glued-in rebars and then to the column plates. The column plates and the beamplates were bolted to the stiffened angles. The beam plates were welded to the rebars gluedperpendicularly to the grain into the beam. In effect, the moment from the column wastransferred to the beam as a couple of forces.4.5.2 Calculations.It was assumed that the joint should be able to transfer the factored bending momentequal to 8OkNm (see Table 4.1).The first step was to determine the necessary number and size of the inclined rebarsto carry that moment. As the top surface of the column plates was supposed to match thesurface of the glulam it was calculated that the effective arm between the forces in the rebars(parallel to the column axis) was a =377mm for #15 rebars or a =373mm for #20 rebars (seeFig.4.9).NNoxcos3O°Fig.4.9 Forces created at the end ofthe column.COLUMN95The tensile strength of the rebars N0 was established earlier during the material testsfor both rebar sizes. The values used for the design are presented in Table 4.4.yield load [kN] ultimate load [kN]rebar’s 0° 30° 0° 30°size N0 Ny Nuo Nu#15 90 78 130 112#20 140 121 200 173Table 4.4 Capability ofglued-in rebars to transfer tension loads.Itwas desirable that the capacity ofthe rebars was close to the design moment (8OkNm)at the yield level and close to the short term bending resistance of the column at the ultimatelevel. As the normal term loading resistance of the 130x418mm glulam section is lO4kNmthe short term resistance was estimated to be:Mr= 104/0.8=130 [kNm]where 0.8 is the amplification factor due to the short term loading.Two possible combinations of the rebars could be used to resist above moment: three #15rebars (moment =3x 1 l2kN x 0.377m = l27kNm), or two #20 rebars (moment =2x 173kNx 0.373m = l29kNm) per side of the connection. The second option was chosen. It wasanticipated that the rebars would start to yield in tension at about 9OkNm and would be ableto transfer ultimate moment of l29kNm. Ofcourse, to reach the ultimate level, the minimumembedment length for each rebar had to be ensured. For the rebars inclined at 300 to thegrain this length wasLe30=300mm.The second step was to establish the number of rebars glued into the beam. As #20rebars were used in the column it was logical to use the same size rebars in the beam. Thearm of the forces was larger than the arm in the column so the force in the beam rebars wassmaller than in the column rebars. Therefore, it was decided to use two #20 rebars on eachside ofthe connection. The minimum embedment length for those rebars wasLe90=400mm.96The next step was to design the bolted connection between the angle and the columnas well as the beam plates. It was assumed that the column part would be loaded in shearonly. The tension forces in those bolts weren’t expected as any lateral forcewould be transferin bearing between the column plate and the angle on the opposite side of the column. Theshear force in all bolts was equal to the ultimate tension force in two inclined #20 rebars,i.e. 2x173 =346kN. As the factored shear resistance of a single bolt was 57.6kN, the requirednumber of bolts for the column part was:346= 57.6 = 6.0The connection between the angle and the beam plate was loaded in tension (346kN) aswell as in shear. The total lateral force in the specimen at ultimate level was predicted tobe:P==78.lkN, assume80kN.Therefore the shear force to be transferred by one side of the connection was 4OkN.According to the code CAN/CSA-S16.1-M89 the bolts should satisfy the relationship:m2+13T1 0,56213(AbFU)nwhere:V1= total shear force = 4OkNTf = total tension force = 346kNm=number of shear planes = 113= an interaction factor = 0.30=resistance factor = 0.67AbF,= specified minimum tensile force per bolt = 205kN= number of boltstherefore:97flb(+I3T\....0.56213(AbFU)/ 402 \0.5—÷0.30X346flbI 2 I =3•4\0.56x 0.672X0.30X 2052)As the bolted connections were not the parts of thejoint which were intended to investigate,it was decided to overdesign them and use nc=8and b=4•The last step was to check the strength of all welds between the rebars and the platesas well as in the angles. Because each of the specimens was different the sizes and locationofwelds varied accordingly. The general rule was to overdesign the welds and put them overthe entire length provided by the elements.4.5.3 Performance Expectations.It was expected that the specimen would perform during the test as follows:at 78 kNm: nominal yielding point (400MPa) for the column rebars,80 kNm: factored moment (design moment),90 kNm: yielding point for the column rebars obtained from material tests,124 kNm: short term shear resistance of the column (without considering the increaseof the shear capacity due to the glued-in rebars),129 kNm: ultimate strength of the column rebars (from tests),130 kNm: short term bending resistance of the column,148 kNm: nominal yielding point for the beam rebars,173 kNm: yielding point for the beam rebars obtained from tests.It was anticipated that the ultimate failure would take place in the column rebars on thetension side of the connection (both rebars at the same time).98The limits of the deformations for the building were as follows:8 mm (4000mm1500) - total storey drift due to the specffied wind load and gravity loads,80 mm (4000mm/50) - interstorey deflection due to the earthquake loads.The bending moments in the second storey column due to the specffied wind and gravityloads were 37kNm and due to the earthquake loads were 77kNm.As the specimen represented half of the storey height, the displacements measuredduring the tests by LVDT 7 (see Fig.4.6) reflected 50% of the storey drift. Therefore, it wasdesirable that the measured deflection would be less than 4mm at 37kNm and less than 40mmat 77kNm.4.6 TEST T-14.6.1 Joint Configuration.Fig.4.10 Locations of the column’cplates and rebars in the specimen T-1..(___-:d__jIo.j14 102 149I 12012,71f1599Each side of the joint in the specimen T-1 consisted of:- two inclined #20 rebars glued into the column (400mm and 600mm) located in twodifferent planes,- column plate 12.7x102x220mm with 8 tapped holes for 5/8 bolts,- two #20 rebars glued into the beam perpendicularly to the grain (410mm each) locatedin one plane,- beam plate 12.7xlO2x200mm with 4 tapped holes for 5i bolts,- steel angle 12.7x200x220mm with a stiffener.The joint in the specimen T-1 is shown on Fig.4.10 and Fig.4.11.E;110 20200Fig.4.11 Locations ofthe beam plates and rebars in the specimen T-1.1004.6.2 Manufacturing Steps.The specimen T-1 was manufactured and assembled as follows:1. woodworking on the column (routing grooves, drilling holes at 300 angle);2. preparing the rebars and the plates for the column (cutting, bending and fitting therebars, drilling and tapping the holes in the plates);3. tackwelding the rebars to the plates, removing them from the holes and final weldingaway from the glulam;4. sandblasting the steel assembly,5. pouring the glue into the holes and inserting the rebars (one side of the column at atime);6. repeating the procedures 1-5 with respect to the beam (the holes for the rebars weredrilled at 9Q0 angle);7. preparing the angle plates (cutting and drilling) and welding the stiffeners,8. bolting the stiffeners to the column plates and then to the beam plates,9. curing the epoxy in the specimen at least 6 days.The pictures of some of the above steps will be presented in section 4.7.2 as they were takenduring manufacturing of the specimen T-2.4.6.3 Testing Procedures.After the glue set, the beam was positioned in the testing frame (refer to section 4.3),leveled, and bolted down to the frame. The column was then moved into the frame andbolted to the beam. The specimen T-1 in the testing frame is shown on Fig.4.12.101installed.Next, the lateral bracing was installed and the hydraulic jack was connected to thecolumn. Finally the displacement gauges were attached to the specimen and connected tothe data acquisition system. The location of the gauges was described in section 4.3 and isshown on Fig.4.13. The distances shown on Fig.4.7 were: Si=l8Omm, S2=5Omm,S4=200mm.102Fig.4.12 Specimen Ti in the testingframe. Note: the lateral bracing has not been yetThe loading procedures for the test T-1 were as follows:Each cycle started from zero load point. The negative load was applied first (pushing thecolunm from the right to the left) creating compression in the left hand side of the connectionand tension in the right hand side of the connection. After reaching the desired load levelthe unloading started. After passing the zero load point the load was reversed (puffing thecolumn) and increased to the similar level as the negative one. At that moment thecompression was created in the right side of the connection and the tension in the left side.Fig.4.13 Locations of the dzplacement gauges in the specimen T-1.1. load range from -10 to +10 kN2. load range from -20 to +20 kN3. load range from -40 to +40 kN4. load range from -50 to +50 kN5. final loading2 cycles2 cycles4 cycles4 cyclesfrom 0 to -54kN and to +6OkN.103The unloading to the zero load level ended the cycle.As the displacement control was used during this test, the peak loads varied slightlyfrom one cycle to the other.The test was stopped after a characteristic loud sound (indicating rebar fracture) washeard.4.6.4 Results.OVERALL BEHAVIOUR.The specimen T-1 failed by yielding and fracture of one of the inclined column rebarson the left side of the connection (see Fig.4.14 and Fig.4.15).—— - I PFig.4.14 Joint in the specimen T-1 afterfailure.104The ultimate load was 59.5kN which created the moment at the joint equal to 95kNm.Although, the moment was higher than the design requirement for the structure, this wasonly 74% of the expected ultimate moment (the bending stress in the glulam = 25.1MPa).Surprisingly, only one rebar was broken - not both of them as was expected. Thatindicated uneven distribution of the load between the rebars.Very large deformations of the glulam under the column plates were observed duringthe test (see Fig.4.14). An insufficient bearing resistance of the glulam was blamed for thiseffect. Probably, this was also the reason for uneven stress distribution between the columnrebars.MOMENT-DEFLECTION RELATIONSHIP.The relationship between the moment at the joint and the deflection of the end of thecolumn is presented on Fig.4.16.Fig.4. 15 Left side of the column after exposing the rebars. Note the fracture ofthe upperrebar and thefinal incinatk’n oftheplate with respect to the laminations(on ginallyparallel).105.0080604020160 .POSITIVE MOMENT120--_80-120n.-20 -j-40-60-80-.0080-60 -40 -20 0 20 40 60DISPLACEMENT [mmj (LVDT7)Fig.4.16 Deflections ofthe columnfree end during the test T-1(first cycles at each load level arepresented).The moment-deflection (M-D) relationship can be described as linear up to 32kNm(2OkN load level). At that point the deflection of the top of the column reached 7.3mm. Athigher loads, pinching of the M.D curve was observed. Approximately at the same time thecrushing of the glulam under the colunm plates started. The average slope of the curvedecreased as the load increased. At a load level 5OkN it was only 55% of the slope observedin the first cycle (lOkN level).The specimen barely reached the design level of 8OkNm. Immediately after that thecolumn rebars on the right side ofthe connection started to yield (moment =-86kNm). Verylarge deformations of the joint region were observed. After the load was reversed, a similareffect was observed on the left side of the connection. Finally, the specimen failed by rebarfracture at a moment equal to 95kNm.The serviceability limit for the wind loads (at 37kNm) was exceeded by more thanEIzLi0-1-80106twice. The corresponding deflection was 9.4mm. However, the limit for the earthquakeloads was satisfied with only 27.6mm deflection (equal to 69% of the limit). Also the finallarge deformations within the joint prior to the failure were desirable from the earthquakeengineering point of view.THE COMPONENTS OF THE TOTAL DEFLECTION.The total deflection dt (measured by LVDT 7) consisted of three main components:db - displacement due to the rotation of the beam,dc - displacement due to the deformation of the glulam column,d - displacement due to the deformations between the glulam and the plates as well asrebars yielding.There was also a component due to the transverse movement of the whole specimen (in thedirection of applied force) but this was considered to be small with respect to the othercomponents.The percentage share of those three components in the total deflection of the columnis presented in Table 4.5.load moment db dc d dt [100%]kN kNm % % mm10 16 19 48 33 2.720 32 12 35 53 7.540 64 12 27 61 19.350 80 10 22 68 29.460 6.5 13.5 80 58.2Table 4.5 Components of the total deflection in test T-1.In the table above, db was calculated from the test data (based on the readings from LVDT8 and LVDT 9), dc was calculated from the theoretical formula for displacement in a107cantilever glulam beam, dj was calculated as dt-db-dc.It is clear that the component associated with the plates and the rebars behaviourbecame dominant early in the test. At the ultimate load it counted for 75% of the totaldisplacement. It is obvious that reducingd would improve the overall behaviour ofthe joint.THE BEHAVIOUR OF THE COLUMN PART OF THE JOINT.---/ci)COLUMN}\5)IFig.4.17 Behaviour ofthe tension side ofthe connection: a) desirec4 b) observed.108The main problem observed during the test was insufficient resistance of the glulamin bearing under the column plates. In fact, at a load level 4OkN and higher, the connectionbetween the column and the column plates could not be considered rigid. Therefore, thestress distribution in the rebars was different than assumed before the test. That could bethe main reason why only the outer rebar was broken at the end of the test. The differencebetween the desired and observed behaviour of the connection is shown on Fig.4.17.The bolted connection between the column plate and the angle did not cause anyproblems.THE BEHAVIOUR OF THE BEAM PART OF THE JOINT.The bearing problem under the column plates also affected the stress distribution inthe beam rebars. As indicated on Fig.4.17, the tension forces, which were created in thoserebars, were different from the assumed forces. Although the rebars behaved as expected(no yielding) there were problems with the bolted connection and the bending resistance ofthe plate. At the very end of the test the bolts closest to the inner beam rebar failed. In fact,it was the thread in the plate which failed in tension. It turned out that the stiffness of thebeam plate was insufficient for those loads. Due to the overload of the inner bolts themoment created in the plate between the bolts and the inner rebar exceeded the bendingresistance of the plate (see Fig.4.18)Fig.4.18 Bending of the beam plate.1094.6.5 Comments.It was expected that some problems associated with bearing under the plates wouldarise during the test. It was decided, however, that specimen T-1 shouldn’t be reinforcedin compression perpendicular to the grain. The test showed how serious the bearing problemwas and indicated that the reinforcement perpendicular to the grain was necessary underthe column plates.The magnitude of deformations did not come as a great surprise but the ultimate loadwas signfficantly less than expected. The analysis of the data indicated, however, that thecompression perpendicular to the grain was the cause of the problems in both cases.It was also decided that the thickness of the beam plates should be increased in thesubsequent specimens. That would prevent bending of those plates. To avoid pull-out ofthe bolts from the beam plates, it was suggested to increase the number of the beam boltsfrom 4 to 6.4.7 TEST T-2.4.7.1 Joint Configuration.The specimen T-2 was an improved version based upon the experience with thespecimen T-1. The compression reinforcement added was two #15 rebars glued perpendicularly to the grain under the column plates (see Fig.4.19). They were going through theentire depth of the column. Those rebars were flush with the surface of the glulam underthe plates. Some minor changes to the joint configuration are listed below:- beam plates thickness was increased from 12.7 (½”) to 25.4mm (1”),- six instead of four bolts were used to connect the beam plates to the angles,- the beam rebars and the beam bolts were located at the same cross section to reduce110bending moment in the plate,- a 5mm gap between the column and the beam was left to avoid contact between theglulam members during the test.H #20 REBARS14 1511511114pc130Fig.4.19 Locations ofthe column’cplates and rebars in the specimen T-2.The joint in the specimen T-2 is shown on Fig.4. 19 and Fig.4.20.1114.7.2 Manufacturing Steps.The specimen T-2 was manufactured and assembled in a similar way as the specimen T-1(refer to section 4.6.2). The only change was the gluing of the perpendicular rebars in thecolumn. This activity was performed after routing the grooves for the column plates andbefore gluing the inclined rebars into the column.I I-J L.120 f0I 200Fig.4.20 Locations of the beam plates and rebars in the specimen T-2.112Different phases of the manufacturing process ofthe column are presented on Fig.4.21through 4.25.Fig.4.21 Routing grooves. Fig.4.22 Drillingperpendicular holes.113Fig.4.23 Gluingperpendicular rebars.Fig.4.24 Drilling inclined holes.1144.7,3 Testing Procedures.The assembly of the specimen T-2 in the testing frame proceeded exactly the same asit was described for the specimen T-1 (see section 4.6.3).The location of the gauges was slightly different than previously. The distances shownon Fig.4.7 were: S1 =200mm, S2=SOmm, S4=200mm.115Fig.4.25 Fitting inclined rebars andplates.The loading procedures for the test T-2 were as follows:1. load range from -10 to +10 kN2. load range from -20 to +20 kN3. load range from -30 to +30 kN4. load range from -40 to +40 kN5. load range from -50 to +50 kN6. load range from -60 to +60 kN7. fmal loading4 cycles4 cycles4 cycles4 cycles4 cycles4 cyclesfrom 0 to -6lkN and to +4OkN.Load control was used throughout the test (except the final loading) with sinusoidalwave form (see Fig.4.26).z‘-ITIME [mm]Fig.4.26 Loadingprocedures for the specimen T-2.116The period of the load cycles was 1 minute up to 40 kN load and 2 minutes over 40 kN load.Each cycle started from zero load point. The positive load was applied first (pullingthe column from the left to the right) creating compression in the right hand side of theconnection and tension in the left hand side of the connection.Displacement control was used during final loading (not shown on Fig.4.26) becauseof large displacements. In fact, the test was terminated without obvious failure of the glulamor steel elements due to the limitation of the stroke of the hydraulic jack.4.7.4 Results.OVERALL BEHAVIOUR.A rebar fracture in the specimen T-2 was not observed during the test. However thedeformations of the joint (see Fig.4.27) indicated some kind of hidden failure.Fig.4.27 Joint in the specimen T-2 duringfinal loading.117After the test, the column rebars were exposed and it turned out that the weld betweenthe outer rebar and the left column plate started to tear (see Fig.4.28). The fracture alsoinvolved some parts of that rebar so it was not a clear weld failure.The ultimate load during the test reached 60.5kN which created the moment at thejoint equal to 97kNm (the bending stress in the glulam = 25.6MPa). This is almost identicalresult to the previous test T-1. Also the deformations of the joint were at the same level asbefore.Apparently, the reinforcement which was used in the specimen T-2 was not effectiveas far as the overall behaviour of the specimen is concerned. It worked, however, locally.MOMENT-DEFLECTION RELATIONSHIP.The relationship between the moment at the joint and the deflection of the end of thecolumn is presented on Fig.4.29.I’WELDFig.4.28 Failure of the specimen T-2.118Fig.4.29 Deflections of the column ‘S free end during the test T-2(first cycles at each load level arepresented).The linear character of the M.D curve was observed only at the lowest load level(l6kNm, lOkN). Immediately after that the pinching started. The average slope of thecurve decreased by 26% with respect to the original. However the slope at lOkN load levelwas only 75% of the slope observed in the specimen T-1.The specimen barely reached the design level of 8OkNm. Immediately after that thecolumn rebars started to yield. The specimen survived all 4 cycles at 6OkN load level (moment=96kNm) but the deflections recorded by LVDT7 were 2.2 times larger than at the previouslevel (5OkN).The serviceability limit for thewind loads (at 37kNm) was exceeded almost three times.The corresponding deflection was 11.6mm. The limit for the earthquake loads was satisfied,similar to T-1 test, with only 27.0mm deflection (68%).EIzUI0160—POSITIVE MOMENT120--80”II0080604020•-40-60-80n- I1W I-100 -80 -60 -40 -20 0 20 40DISPLACEMENT [mm] VDT7)I I IW60 80 100119THE COMPONENTS OF THE TOTAL DEFLECTION.The percentage share of three components in the total deflection of the column ispresented in Table 4.6.load moment db d d dt [100%]kN kNm % % % mm10 16 19 29 52 4.520 32 15 29 56 9.130 48 12 27 61 14.640 64 11 25 64 20.950 80 10 22 68 30.060 96 6.5 12 81.5 65.0Vote: dt - total deflection;db - displacement due to the rotation of the beam;d displacement due to the elastic deformation ofthe column;dj - displacement due to the deformation of the joint.Table 4.6 Components of the total deflection in test T-2.The specimen T-2 appears to be even more flexible than T-1 - specially within thelower range of the load. But the percentage share of three components is almost exactly thesame as before.THE BEHAVIOUR OF THE COLUMN PART OF THE JOINT.The comparison of the LVDT 3 readings (refer to Fig.4.7 or 4.27 for the location)during the tests T-1 and T-2 is presented in Table 4.7 (positive moment).120load moment LVDT 3 [mm]kN kNm T-1 T-2 T2/T110 16 0.05 0.03 0.6020 32 0.12 0.09 0.7530 48- 0.12 -40 64 0.40 0.18 0.4550 80 0.90 0.24 0.2760 96 3.20 0.46 0.14Table 4.7 Local compression deformations of the glulam at the end ofthe column.The data collected by LVDT 3 shows that the bearing reinforcement did work.Unfortunately, its effect was only local and did not influence the overall behaviour of thespecimen. The compression deformations in the column at the location ofthe reinforcementwere much smaller than those recorded during T-1 test. The reinforcement became moreand more effective as the load (and bearing stresses) increased. At the 6OkN load level thedeformations of the glulam were only 14% when compared to unreinforced specimen.load moment LVDT 4 [mm]kN kNm T-1 T-2 T2JT110 16 0.06 0.09 1.5020 32 0.45 0.51 1.1330 48 - 0.99 -40 64 1.39 1.68 1.2150 80 2.43 2.96 1.2260 96 4.70 11.3 2.40Table 4.8 Local “tension” deformations between the column and the column plate.121At the same time, the gauge located on the top end of the column plate (LVDT 4)recorded the uplift displacements of the plate versus the glulam. Those readings are compared in Table 4.8.The deformations were over 20% larger during the test T-2 than during the test T-1(unlike the readings from LVDT 3). At the ultimate load they were 2.4 times larger but thiswas the effect of the weld failure.The tension forces perpendicular to the plate were carried (in both specimens) by theinclined rebars only. Apparently, this was not enough. Some kind of”tension reinforcement”turned out to be necessary.THE BEHAVIOUR OF THE BEAM PART OF THE JOINT.Fig.4.30 Partialpullout of the beam rebar in the specimen T-2.122The bending of the beam plates was successfully avoided in the specimen T-2. Theresistance of the bolts was also sufficient. However, a new problem appeared at the veryend of the test. The inner beam rebar on the right side of the joint was pulled out of thebeam (see Fig.4.30).It was very difficult to determine the reason for that kind of failure. It is possible thatthis particular rebar was not adequately cleaned before gluing. The corresponding rebar onthe other side of the joint did not show any signs of pullout failure, although it experiencedthe same tension force. To avoid similar accidental situation, it was decided that the beamrebars in the next specimen should be extended from 400mm to 550mm.4.7.5 Comments.The lack of improvement in the behaviour of the specimen T-2 was obvious. In someaspects, T-2 was even worse than T-1. Although the compression deformations under thecolumn plates were significantly reduced, the reinforcement caused even larger “tension”deformations.After this test it became clear that the column plates and the glulam should be ableto resist not only compression forces but also tension forces in the direction perpendicularto the axis of the column. Therefore, the plates should be permanently connected to therebars glued perpendicularly to the grain.1234.8 TEST T-3.4.8.1 Joint Configuration.I 418—I—I #20 REBARS/Ii //ii I / //1//Ill I /II //\////‘0////II #15REBARSo1Ia - -ilIIj I —— —Hi Io - -V/In cD_//(Ii (UIJ 11) —1 cI(__I u151 51 12711120127Fig.4.31 Locations of the column’cplates and rebars in the specimen T-3.124Thejoint in the specimen T-3 was redesigned. The inclined column rebars (#20) wereplaced in one row and inclined out of the plane by 3° to allow crossing with the rebars fromthe other side. The length of the column plates was increased to 240mm and their width to127mm (equal to the width of the column). Four #15 rebars were placed in the holes drilledperpendicularly to the grain. Those rebars went through the entire depth of the column andtheir ends were welded to the column plates, and were also glued to the glulam. In this way,the #15 rebars acted not only as bearing reinforcement (as it was in the specimen T-2), butwere also able to resist tension forces. The column part of the joint is shown on Fig.4.31.Fig.4.32 Locations ofthe beam’c plates and rebars in the specimen T-3.125Some changes were also introduced to the beam part of the joint (see Fig.4.32.). Thewidth of the beam plates was increased, as it was done in the colunm, to cover the full widthof the beam. Three #20 rebars (instead of 2) were welded to the plate and glued into thebeam. Their length was also increased to 550mm to eliminate any accidental pull-out (referto section 4.7.4). In order to accommodate three rebars in one row it was necessary toincrease the length ofthebeam plate to 235mm. As the bending ofthe platewas not observedin the specimen T-2 the thickness of the beam plates in the specimen T-3 was decreased to19mm (Y4”).The dimensions of the angle with the stiffener were increased accordingly to the size of theplates.4.8.2 Manufacturing Steps.The specimen T-3 was manufactured in a way similar to the specimen T-1. However, someextra operations were necessary due to the perpendicular reinforcement in the columns.After the column plates with the inclined rebars (#20) were glued into both sides of thecolumn, the holes perpendicular to the grain (¾” diameter) were drilled through the entiredepth of the glulam (the plates were predrilled earlier). #15 rebars were placed in thoseholes and glued to the glulam. After the glue set, each end of the rebar was welded to thecolumn plate.4.8.3 Testing Procedures.The assembly of the specimen T-3 in the testing frame proceeded in exactly the samemanner as it was described for the specimen T-1 (see section 4.6.3).The location of the gauges was slightly different than previously. The distances shownon Fig.4.7 were: S1 =200mm, S2=lOmm, 54=220mm. The gauges LVDT1O and LVDT11were used during this test.The loading procedures for the test T-3 were as follows:1261. load range from +10 to -10 kN2. load range from +20 to -20 kN3. load range from +30 to -30 kN4. load range from +40 to -40 kN5. load range from +50 to -50 kN6. load range from +60 to -60 kN7. load range from +70 to -70 kN8. load range from +80 to -80 kN9. final loading806040204 cycles4 cycles4 cycles4 cycles4 cycles4 cycles4 cycles4 cyclesfrom 0 to +9OkN and to -86kN.Load control was used throughout the test with sinusoidal wave form (see Fig.4.33).zI°°JGYL.ES WITh lrn!n.PERIOD CYCLES WiTH 2nhIr PERIOD————-i-100— . ——— . . . ——0 5 10 15 20 25 30 35 40 45 50 55 60 65TIME [mm]Fig.4.33 Loadingproceduresfor the specimen T-3.The period of the load cycles was 1 minute up to 40 kN load and 2 minutes over 40 kN load.Each cycle started from zero load point. The positive load was applied first (pullingthe column from the left to the right) creating compression in the right hand side of the127connection and tension in the left hand side of the connection.The test was terminated after a shear failure occurred in the glulam column, withoutany sign of the failure in the rebars. The failure included delamination of the column (failureof the glue line) on the length equal to approximately 213 of the column height.4.8.4 Results.Fig.4.34 Specimen T-3 duringfinal loading.128OVERALL BEHAVIOUR.Thebehaviour ofthe specimenT-3 was muchbetter than the behaviouroftwo previousspecimens. Both, the stiffness and the strength of the connection were improved. Theglulam column failed in shear at -86kN load (110% of the calculated resistance), howeverthe specimen survived the load of +9OkN (116%, shear stress = 2.48MPa). The shear tookplace on both sides of the column but away from the joint region, i.e. away from the locationof inclined and perpendicular rebars. That proved the positive effect of gluing the rebarson the shear capacity of the glulam member. The shear failure of the specimen T-3 is shownon Fig.4.34 and Fig.4.35.After the test, the column rebars were exposed to check where the fracture occurred.The inclined rebars as well as the perpendicular reinforcement did not show any sign ofyielding. Also the welds performed very well.Fig.4.35 Joint in the specimen T-3 afterfailure.129The ultimate load during the test reached 90.OkN which created the moment at thejoint equal to l44kNm. This is almost 50% more than in two previous tests and 80% morethan the design requirement for the structure. The bending stress in the column reached:144000000=38.OMPa130X 4182The deformations of the joint were approx. half of those recorded previously at theultimate level.Apparently, the welding of the perpendicular rebars was the key for the improvedbehaviour of the joint.MOMENT-DEFLECTION RELATIONSHIP.The relationship between the moment at the joint and the deflection of the end of thecolumn is presented on Fig.4.36.EIzUi0Fig.4.36 Deflections ofthe column vfree end during the test T-3(first cycles at each load level arepresented).DISPLACEMENT [mm] VDT7)130The linear character of the M.D curve was observed only up to the moment of 32kNm(2OkN load level). Immediately after that some pinching started. The average slope of thecurve was over 30% greater than observed during the test T-1. The slope decreased duringthe test by 45%, but its fmal magnitude (at 8OkN level) was equal to the slope of the specimenT-1 at 4OkN (2OkN for T-2).The specimen quite easily reached the design level of 8OkNm (5OkN). The ultimatemoment - l44kNm - was greater than anticipated moment resistance of the joint by 12.5%and greater than calculated short term moment resistance of the column by 11%.The serviceability limit for the wind loads (4mm at 37kNm) was exceeded 2.5 times.The corresponding deflection was 9.9mm. The limit for the earthquake loads was satisfied- at 77kNm the deflection was only 19.7mm (50% of the limit).THE COMPONENTS OF THE TOTAL DEFLECTION.load moment db dc d dt [100%]kN kNm % % % mm10 16 20 50 30 2.620 32 15 37 48 7.130 48 13 31 56 12.840 64 12 31 57 16.950 80 13 30 57 21.760 96 12 30.5 57.5 25.970 112 12.5 30 57.5 30.980 128 12 28 60 37.4\Tote: dt - total deflection;db - displacement due to the rotation of the beam;dc - displacement due to the elastic deformation of the column;dj - displacement due to the deformation of the joint.Table 4.9 Components ofthe total deflection in test T-3.131The percentage share of three components in the total deflection of the column ispresented in Table 4.9.The joint deformation component (d’) is slightly smaller than in previous tests. Thedifference is more significant at lower load levels (by 40% at lOkN) than at higher (by 10%at 5OkN when compared to T-2). That indicates greater stiffness of the specimen T-3.THE BEHAVIOUR OF THE COLUMN PART OF THE JOINT.The comparison of the LVDT 3 readings during the tests is presented in Table 4.10(positive moment).The data collected by LVDT 3 shows further improvement of the bearing reinforcement. The compression deformations in the specimen T-3 were 66% smaller than thecorresponding deformations in the specimen T-2 and over 80% smaller than thedeformations in the specimen T-1! The ultimate deformation of 0.14mm in the reinforcedspecimen (T-3) was reached in the unreinforced specimen (T-1) at about 22kN (at 4 timeslower load!).load moment LVDT 3 [mm]kN kNm T-1 T-2 T-3 T3)T210 16 0.05 0.03 0.02 0.6720 32 0.12 0.09 0.03 0.3330 48 - 0.12 0.04 0.3340 64 0.40 0.18 0.06 0.3350 80 0.90 0.24 0.08 0.3360 96 3.20 0.46 0.09 0.2070 112 -- 0.11 -80 128 -- 0.12 -90 144 -- 0.14 -Table 4.10 Local compression deformations of the glulam at the end ofthe column.132load moment LVDT 4 [mm]kN kNm T-1 T-2 T-3 T3/T210 16 0.06 0.09 0.02 0.2220 32 0.45 0.51 0.02 0.0430 48- 0.99 0.03 0.0340 64 1.39 1.68 0.04 0.0250 80 2.43 2.96 0.05 0.0260 96 4.70 11.3 0.06 0.0170 112 - - 0.07 -80 128 - - 0.12 -90 144 - - 0.49 -Table 4.11 Local tension deformations at the end ofthe column.Even more astonishing results were collected by the gauge LVDT 4 (see Table 4.11).The tension deformations were reduced to 2% of the values recorded during the test withthe unreinforced specimens.What is even more important, the compression and the tension deformations in thespecimen T-3 had identical amplitudes. Additionally, their graphs were linear throughoutentire test.The welding of the perpendicular rebars to the plates turned to be a very effective wayof decreasing compression and tension deformations between the glulam column an thecolumn plates.THE BEHAVIOUR OF THE BEAM PART OF THE JOINT.The specimen T-3 did not experience any problems in the beam part of the connection.The beam’s plates and rebars behaved perfectly well. The gauges measuring the deformations between the beam and the plates recorded linear relationship throughout the entiretest. The deformations at the ultimate load level reached 0.5mm in tension and 0.1mm incompression. It was also observed that the required behaviour of the beam plates was133achieved, i.e. the outer end of the plate experienced larger deformations than the inner end(as shown on Fig.4.17a). In this situation the additional third rebar under the beam platewas not needed.4.8.5 Comments.The specimen T-3 showed great improvement in comparison to the other specimens.It survived the highest load and, at the same time, it displayed the smallest deflections ofthe column. The deformations measured by all gauges located within the joint region weresmaller than any other recorded earlier. Compression and tension deformations had similaramplitudes and their behaviour was linear throughout the test. The overall stiffness of thejoint, measured by the slope of the moment-deformation curve, was over 30% greater thanpreviously recorded. Both, the shear resistance and the bending resistance of the column,calculated according to the code values, were exceeded during the test. The failure tookplace in the glulam which indicates that the joint can be made stronger than the memberitself.Most of the improvement described above should be credited to theperpendicular reinforcement used in this specimen. The welding ofthe perpendicular rebarsto the plates seems to be the proper way ofdesigning the column part of the connection. Inthe specimen T-3 the rebars were glued before they were welded to the plates. That causedsome problems: the penetration of the weld was not complete and some toxic fumes werecreated during the process. Although the welds performed adequately during the test, inthe future, welding should take place before gluing. This can be done by injecting the gluebetween the glulam and the rebar through an additional, small hole made in the side of thespecimen.This specimen met all design criteria except one: serviceability at low load levels.Further research should try to eliminate that problem.1344.9 CONCLUSIONS.The tests described in this chapter showed that the glued-in rebar connections are ableto transferbending moments under reversed loading conditions in multi-storey frame structures.In fact, the connection may be stronger than the glulam member which is connected (glulamfailure in the test T-3). However, to achieve an acceptable behaviour of the connection a verycareful detailing is necessary. The reinforcement perpendicular to the grain under the plates isessential for obtaining satisfactory stiffness of the joint. This reinforcement should be able totransfer compression forces as well as tension forces.135CHAPTER 5DESIGN GUIDELINESFOR GLUED-IN REBAR JOINTS.5.1 OBJECTIVE.The objective of this chapter is to provide basic information about a design method formoment resisting joints using glued-in rebars. Although the information is based on thebeam-to-columntests described in chapter 4, it maybe also used to design other moment resistingjoints, such as column to foundation joint or beam splice.The main aim for using the glued-in rebar connection method is to obtain timber connections with reliability equal to reliability of the steel joints. Whenever the embedment lengthof the rebar is larger than required minimum length, the failure takes place in the steel.Therefore, the connection maybe designed as the steeljoint. As a consequence ofthis approach,the designer should rely only on steel parts of the connection. Contact between the glulammembers should be ignored because it may change the pattern of load transfer. For example,putting the column in contact with the beam would decrease the lever arm between the tensionforce and the resultant of the compression forces. It would also be very difficult to predict thelocation ofthat resultant. In effect, the connectionwould be less reliable than similar connectionwhere the column and the beam are not in direct contact.However, there are locationswhere relying only on steel would be too conservative. Thatis the case of bearing plates, which are widely used in the glued-in rebar joints. Usually therebar is welded to the plate and, where subjected to compression load, the plate and the rebaract together. The compression load is transferred by the bearing under the plate and shear stressaround the rebar. The tests described in Chapter 3 showed that the compression deformations136of the rebar glued perpendicularly to the grain and the deformations of the bearing plate aresimilar within elastic range. Therefore, with sufficient degree of accuracy, it may be assumedthat the load carrying capacity of the plate with the rebar in compression perpendicular to thegrain is the sum of the capacity of the rebar and the plate acting separately (refer to Chapter 3for more details).5.2 DESIGN PHILOSOPHY.The design philosophy of the glued-in rebar connection is based on the following:- the moment in the connection is transferred by the rebars glued into the glulam at an angleto the grain to involve the whole cross section of the element (or a significant portion ofthat section),- the failure of the connection takes place in the steel,- all forces are transferred through the steel parts of the connection (the contact betweenthe glulam members is disregarded),- wherever the bearing stress occur under the plate, the glulam should be reinforced bygluing additional rebars perpendicularly to the grain; furthermore, those rebars should bepermanently attached to the plate so they can transfer tension and compression forces,- the welding of the rebars to the plates should take place before gluing of the rebars, preferably away from the glulam,- the connection should be designed in such way that all the gluing procedures take place inthe plant, before delivery of the elements to the site,- field welding should be avoided,- bolted connection of the elements (their steel parts) on the site is preferred.1375.3 DESIGN ASSUMPTIONS.The following assumptions have been made:1. The failure of the glued-in rebar connection is a tension failure in the rebar.2. The bending moment is transferred by the glued-in rebars acting in tension or compressionand forming a moment couple (see Fig.5.1). The compression interaction of the glulamand the rebars takes place exclusively under the beam plates subjected to the compressionforces (bearing).3. An axial force in the column (not included during the tests described in chapter 4) istransferredby the glued-in rebars in the beam in conjunction with bearing under the plates.4. The shear forces (usually insignificant in this configuration of the beam and the column)are carried by shear in the beam’s rebars, bearing under the column plate, and tensionin the column bolts.a)Mfb)4Ffc)Fig.5.1 Forces in the connection:a) bendin& b) axia4 c) shear.1385. The minimum embedment length is provided for all rebars in theconnection, according to Table 5.1.rebar size embedment length at an angle to the grain [mm]9()O 300#10 200 150#15 300 250#20 400 300Table 5.1 Minimum embedment lengthfor various size ofthe rebars.6. The diameter of the holes drilled in the glulam is 3-5mm larger than the nominal diameterof the rebars.7. The glue provides enough resistance to yield the rebar in tension when the embedmentlength is greater than the length specified in Table 5.1.8. Weldable rebars are used.9. The buckling of the rebars subjected to compression loads is restrained by the glulam,therefore the compression resistance of the rebars may be calculated as being equal tothe tension resistance of the rebars.10. The design is in the elastic deformations range.11. The stresses created in the rebars due to the factored loads should be smaller than thefactored yield strength of reinforcing steel (4 X F ).In the design guidelines presented below, it is assumed that the elements themselves(beams and columns) have already been designed according to section 6 of Canadian WoodDesign Code CAN/CSA-086.1-M89. Therefore the applicable formulas referring to theresistance of the glulam members are not quoted.The formulas, which are introduced in the following sections, pertain to the resistance of thejQjflt and its elements (i.e. resistance of the rebars, plates, and the glulam in direct contact withthe plates) caused by the forces acting in the joint (i.e. bending moment, axial force, shear force).1395.4 MOMENT RESISTANCE OF THE JOINT.The moment resistance of the joint should be greater than the factored bending momentacting in the joint:M1 (Mrc, Mrb)The moment resistance of the beam to column connection should be calculated as thelesser of the following:whereMrc=x F,orMrb= 1b FbMrc - moment resistance of the column part of the joint,Mrb - moment resistance of the beam part of the joint,Fig.5.2 Geomehy ofthe beam to column joint140- distance (centre to center) between the column rebars welded to the plates,- distance between the beam rebars (or between groups of rebars),Fc - resistance of the colunm rebars on one side of the joint equal to:F = ,x AX FyrX cosaFb - resistance of the beam rebars on one side of the joint equal to:F b = 4) x A b X F .•- resistance factor (suggested value 0.9),F - specffied yield strength of reinforcing steel,Ac - total area of the rebars glued in the column on one side of the joint,Ai, - total area of the rebars glued in the beam on one side of the joint,a - angle of inclination of the column rebars.The resistance of the glulam members has to be at least equal to the calculated resistanceof the joint. The resistance of the column may be calculated as the moment resistance of thetotal cross section (based on gross area) of the member. The resistance of the beam should becalculated based on the net area of the member at the location of the beam rebars. This is aconservative approach but it should be adopted until the reduction of the moment capacity ofthe glulam member, caused by drilling the holes at 90° and gluing the rebars, is established.5.5 BEARING REINFORCEMENT OF THE COLUMN PLATES.The tests described in Chapter 4 showed that the bearing resistance of the glulam underthe column plates was not sufficient. The bearing reinforcement, in the form ofthe rebars gluedperpendicularly to the grain, was used to decrease local compression deformations ofthe glulam.The rebars were continuous over the entire depth of the column and connected by welding tothe column plates on the opposite sides of the column.This reinforcing method is recommended for the connection as means of providing sufficient stiffness. The rebars glued at 90° angle to the grain have to transfer the horizontal forces141(A X Fyr x sjn cxi) coming from the inclined rebars (see Fig. 5.3). Bearing in the glulam is notaccounted for because the horizontal forces may create compression as well as tension in theperpendicular rebars, depending on the direction of the moment applied to the column.A F,,.sin ocFig.5.3 Forcesperpendicular to the grain in the columnThe tension and compression resistance of the perpendicular rebars in the column maybe calculated as:whereF perp = x A prp x FFperp resistance of the perpendicular rebars,- resistance factor (suggested value 0.67),F - specffied yield strength of reinforcing steel,Aperp total area of the perpendicular rebars glued in the column and welded to the plates.A0FA0 Fcos oc01425.6 AXIAL RESISTANCE OF THE JOINT.It is assumed that the axial force is carried by the steel parts only without direct contactbetween the column and the beam. The axial force resistance should be taken as the lesser ofthe following:P= 2xF,orPb=2X(Fb+Qf)andP1(P, Pb)wherePc - axial resistance of the column part of the joint,- axial resistance of the beam part of the joint,Fc - resistance of the column rebars on one side of the joint, calculated as in section 5.4,Fb - resistance of the beam rebars on one side of the joint, calculated as in section 5.4,Or - bearing resistance of the glulam under one beam plate, calculated according toCAN/CSA-086.1-M89 section 6.5.9.5.7 COMBINED AXIAL AND BENDING RESISTANCEOF THE JOINT.In case of combined axial loads and bending the following interaction equations shouldbe satisfied:—+ l.OF lXF.orP,_________—+ 1.OFb lbX(Fb+Qr)143wherePf - applied axial load,Mf - applied bending moment,F, Fb, Qr - as in section 5.6,ic’ ‘b - as in section 5.4.5.8 SHEAR RESISTANCE OF THE JOINT.It is assumed that shear force from the column is transferred to the stiffened angles (bybearing between the column plate on one side of the column, and tension in the bolts on theother side of the column) and then to the beam plates (creating shear in the bolts connectingthe angles and the beam plates). Then the shear is transferred from the plates to the beamrebars. This is the critical section, where the resistance of the joint due to shear force shouldbe checked.The shear resistance of the joint should be calculated as follows:Vrb4<2) CK— 1 EOSKSEKT 1 11400x1.0x1.02C F 2x21.2 20.4Pr=0.9X20.4X 175x342x0.62=681300N=681.3kNO.K6.3.2.2 second storeyKel.OL = 4000 - 570 = 3430 mmCK = 20.6151Try glulam section: 175 x 190 mmC== 19.6‘o CK— 1 EOSKSHKT 1 11400x1.0x1.02C F 2x22.7 20.4Pr=0.9X20.4X 175x152x0.54=263700N=263.7kNO.K1526.3.3 Diagonals.The gross section tensile resistance of the diagonals is calculated in this section.KSt = 1.0 - service condition factor for tensiontg = 15.3 MPa - specified strength in tension parallel to the grain at gross sectionFtg = ftg(KDKHKStKT) = 15.3 x(1.0) = 15.3 MPa6.3.3.1 first storeyAssume glulam section: 80 x 304 mm.The factored tensile resistance parallel to the grain is:TrFtgAg’O.9X15.3)<80 304=334900N=334.9kNO.K6.3.3.2 second storeyAssume glulam section: 80 x 266 mm.The factored tensile resistance parallel to the grain is:TrFtgAgO.9)<15.3’<80 266293000N293.OkNO.K6.3.3.3 third storeyAssume glulam section: 80 x 152 mm.The factored tensile resistance parallel to the grain is:Tr=FtgAg=0.9X 15.3x80 152= 167400N= 167.4kNO.K1536.4 COST COMPARISON.The comparison between the amount ofthe glulam used for both structures (hinged frameand rigid frame) is presented in Table 6.2. The numbers in the table represent the volume ofthe glulam, in m3, needed for one frame.elements hinged frame rigid framebeams 18.46 15.17columns 3.52 5.60diagonals 2.25 -TOTAL 24.23 20.77Table 6.2 Volume ofthe glulam usedfor the frames [m3].The volume of the glulam used for the rigid frame is 14.3% smaller than the volume needed forthe hinged frame. Assuming the present price of the glulam in B.C. at approximately $1,120 percubic meter, the savings on the glulam can reach $3,900 per one frame ($35,000 on the wholebuilding).However, the glued-in rebar connections are more costly than the traditional non-rigid connections. The cost of materials used for the joint T-3, which was described in section 4.8, is asfollows:material unit unit price quantity costepoxy glue 1 $11 1.5 $17rebars kg $0.65 12 $8steel plates kg $0.90 25 $23bolts (5/8”) pcs $0.42 28 $12total $60Table 6.3 Cost ofthe beam-to-columnjoint usingglued-in rebar connection.154The cost of the materials for a traditional joint between a beam and a column is estimated to beapproximately $40 (the joint consists of 4 shear plates and one steel beam hangers). This costshould be compared to the cost of joining a column to a beam in the rigid frame, which is $60.Although the rigid connections are more expensive than the traditional ones, the total cost ofthe materials for the structure and the joints is smaller in case of the rigid frame with glued-injoints than in case of the traditional frame (see table 6.4).per frame per buildingglulam joints glulam jointshinged frame $27,140 $1,680 $244,000 $15,000rigid frame $23,260 $2,640 $209,000 $24,000cost difference (h-r) $3,880 -$960 $35,000 -$9,000total savings $2,920 $26,000Table 6.4 Cost comparison between hingedframe and rigidframe buildings.Above analysis does not take into consideration the cost of labor, which is very difficult toestimate at this stage. The glued-in rebar connections require more work in the plant, but theerection costs are much smaller than erection costs of the traditional frame. Therefore, it maybe assumed that the combined cost of manufacturing and erection is similar in both cases.155CHAPTER 7CONCLUSIONSThe effect of the moisture content changes and the temperature changes on the glued-in rebarconnection has been investigated. The relative movement taking place at the interface betweenthe glulam and the steel bar was examined. The tests showed that the behaviour of the specimens tested after cyclic temperature changes or cyclic moisture changes was similar to thebehaviour ofthe specimens testedwithout any treatment. The only propertywhichwas affectedby the treatment was a stiffness (decreased about 20%). The other properties, such as strengthand ductility, remained unchanged.Therefore, the glued-in rebar connection may be considered reliable in the temperature andmoisture conditions which can occur in the building. The connectionwas found to be influencedby those conditions to a lesser degree than the glulam structure itself.2. A new method of increasing bearing capacity of the glulam has been developed. The gluingof the reinforcing bars under the bearing plate can increase the compressive resistance of theglulam in the direction perpendicular to the grain by 100%. This is possible because of thedistribution of the local bearing stresses along the glued-in rebar through a significant portionof the glulam cross section.3. The rebars glued-in perpendicularly to the grain have an additional effect of increasing theshear capacity of the glulam members.1564. A beam to column connection for glulam structures, based on the glued-in rebar idea, hasbeen designed and tested, and its stiffness estimated. The connection is capable to transferaxial loads, shear loads and bending moments. Thus, it can be used in statically indeterminatestructural systems, i.e. in the structures where the glulam was not successfully used before.The connection proved to have a bending and shear resistance equal to or greater than theresistance of the glulam members which were joined. This allows more effective use of theglulam in the structure and “capacity design” approach in the design of the members.A ductile behaviour of the connection prior to the failure was observed.5. A suggested design method for the beam-to-column connection is outlined in the thesis.The glued-in rebar connections can be designed as the steel connections whenever the minimum embedment length of the rebars glued into the glulam is ensured.6. The rigid frame with the glued-in rebar joints is compared with a hinged frame utilizing thetraditional connection methods. The savings on the glulam volume can reach 15%.7. However, while the glued-in rebar connection methods seems to work very well in variousconfigurations, there are still some questions to be amplified and thus the need for furtherinvestigation still exists.Future research may include:- the rebars’ contribution to the shear capacity of the glulam,- detailed studies of the embedment mechanism,- development of the finite element model of the connection.157BIBLIOGRAPHY1. Townsend, P.K., Buchanan, A.H., Moss, PJ.“Steel Dowels Epoxy Bonded in Glue Laminated Timber”, Report 90-11, Department ofCivil Engineering, University of Canterbury, Christchurch, New Zealand, June 1990.2. Fairweather, R.H., Buchanan, A.H., Dean, J.A.“Beam Column Connections for Multi-Storey Timber Buildings”, Report 92-5, Department of Civil Engineering, University of Canterbury, Christchurch, New Zealand, July1992.3. Barber, J., Buchanan, A.H.“Fire Resistance of Epoxied Steel Rods in Glulam Timber”, Report 94-1, Department ofCivil Engineering, University of Canterbury, Christchurch, New Zealand, May 1994.4. Buchanan, AH., Fairweather, R.H.“Epoxied Moment-Resisting Connections for Timber Buildings”, Proceedings, International Workshop on Wood Connectors, Las Vegas, Nevada, USA, 1992.5. Riberholt, H.“Steel Bolts Glued into Glulam”, Proceedings, Meeting of IUFRO Wood EngineeringGroup, Oxford, U.K., 1980.6. Riberholt, H.“Glued Bolts in Glulam”, Department of Structural Engineering, Technical University ofDenmark, Series R, Number 210, 1986.7. Riberholt, H.“Glued Bolts in Glulam, Part 2”, Department of Structural Engineering, Technical University of Denmark, Series R, Number 228.1588. Riberholt, H.“Glued Bolts in Glulam. Proposals for CIB Code”, Proceedings, CIBMeeting 21,Parksville, British Columbia, Canada, 1988.9. Foschi, R.O., Folz, B.R., Yao, F.Z.“Reliability Based Design of Wood Structures”, Report No.34, Department of CivilEngineering, University of British Columbia, Vancouver, Canada, 1990.10. Madsen, B., Malczyk, R., Wiktor, R.“Epoxied Dowel Connections”, Proceedings, International Workshop onWood Connectors, Las Vegas, Nevada, USA, 1992.11. Madsen, B.“Structural Behaviour of Timber”, Timber Engineering Ltd., Vancouver, Canada, 1992.12. Malczyk, R.“Glued-in Rebar Connections”, M.A.Sc. Thesis, Department of Civil Engineering, University of British Columbia, Vancouver, Canada, October 1993.13. McIntosh, K.A.“From Theory to Reality -30 Years in Glulam Manufacture”, Proceedings of the SecondPacific Timber Engineering Conference, Auckland, New Zealand, 1989.14. Turkovskij, S.“Prefabricated Joints of Timber Structures with Inclined Glued-in Bars”,Proceedings ofthe 1991 International Timber Engineering Conference, TRADA, London, U.K., 1991.15. Turkovskij, S., Lukyanov, EL, Pogoreltsev, A.A.“Use of Glued-in Bars for Reinforcement of Wood Structures”, Proceedings of the 1991International Timber Engineering Conference, TRADA, London, U.K.,1991.16. Turkovskij, S.B., Kurganskij, V.G., Pochernjaev, B.G.“Opyt primienienija kliejenyh dierieviannyhkonstrukcij vMoskovskoj oblasti” (in Russian),NTO strojindustrii, Moscow, USSR, 1987.15917. Bodig, 3., Meyer, R.W., Kellogg, R.M.“Structural Use of Wood in Adverse Environments. Moisture Effects on Structural Useof Wood”, Society of Wood Science and Technology, New York, USA.18. Schniewind, A.P.“Concise Encyclopedia of Wood & Wood-Based Materials”, Pergamon Press, New York,1989.19. Charleson, A.W., Patience, D.B.“Review of Current Structural Timber Jointing Methods”, New Zealand Timber DesignJournal, Issue 3, Volume 2, 1993.20. Forest Products Laboratory“Wood Handbook: Wood as an Engineering Material”, U.S. Department of Agriculture,Washington, D.C., 1974.21. Wangaard, F.F.“Wood: Its Structure and Properties”, The Pensylvania State University, 1979.22. Canadian Institute of Steel Construction“Handbook of Steel Construction”, Canadian Institute of Steel Construction, Willowdale,Ontario, Canada, 1991.23. Canadian Institute of Steel Construction“Metric Design Notes for Limit States Design in Steel”, Canadian Institute of Steel Construction, Willowdale, Ontario, Canada, 1979.24. PZ1TB“Poradnik inzyniera i technika budowlanego. Tom 5” (in Polish), Polish Association of CivilEngineers and Technicians, Warsaw, Poland, 1986.25. Canadian Wood Council“Wood Design Manual”, Canadian Wood Council, Ottawa, Ontario, Canada, 1990.160APPENDIX ATHEORETICAL DEFORMATIONS AND STRESSESIN A GLUED-IN REBAR SPECIMENDURING FREEZING/HEATING CYCLEThis Appendix presents detailed calculations of the stresses and deformations shown onFig.2.20 in section 2.4.3. Please, refer to chapter 2 for information about the specimen and testingprocedures.1. ASSUMPTIONS.The following assumptions have been made:- there are no internal stresses in the specimen at the beginning of the cycle,- the manufacturing temperature is +20°C and so is the temperature of the specimen (rebarand glulam) at the beginning of the cycle,- a thin layer of glue is not taken into consideration in this calculations,- the changes of the ambient temperature are instantaneous,- linear elastic behaviour of both materials is considered,- elastic modulus of steel (rebar) isE5=200 000 MPa,- elastic modulus of glulam perpendicularly to the grain is Eg500 MPa (test value),- linear thermal expansion coefficient of steel is cs=11.3x106PC,- linear thermal expansion coefficient of wood in the direction perpendicular to the grain iscg=40x106PC,- thermal conductivity of steel is 58 W/mJK,- thermal conductivity of wood perpendicularly to the grain is 0.16 WImIK,- the dimensional changes of the rebar due to the change of the outside temperature arecompleted before the glulam starts to contract or expand - this is not exactly accurate but,due to a very big difference between the conductivities of both materials, is acceptable.161C00)CO2aE0C)aCl)Cl)wI—Cl)C0COC$C0COCO2aE00CuaCl)U)wccICl)C0COC$STRESS IN REBARCC’-C’-Sn,0 _rri EE I.E.D-IA, FE—100’ I I I I I I I-40 -30 -20 -10 0 10 20 30 40 50 (TEMPERARTURE [ C]STRESS IN GLULAM0.6:EE.‘: :zz z.zz0-40 -30 -20 -10 10 20I30 40 ‘ 50TEMPERARTURE [C]Fig. Al Internal stresses in the specimen due to temperature changes.601622. FREEZING CYCLE: from +20°C to -30°C.There are no stresses in the rebar nor in the glulam. Outside temperature is +20°C. Thestarting point of the graph (Fig.A.l) is 0.phase 1 - line OAThe outside temperature changes from +20°C to -30°C (LsT=-50°C). Only the rebaris affected by the changed temperature in this phase.The rebar contracts and puts the glulam into compression. The tension stress develops inthe rebar. An equilibrium is reached when the tension force in the rebar is equal to thecompression force in the glulam.The final contraction of the whole specimen in phase 1, Xj, is calculated by solvingthe equation below [A4].F5g [Al]gAg [A2]EsE3A=E9gAg [A3][A4]-E A = -Eq Agwhere subscripts s g refer to steel and glulam respectively, andF - force at equilibrium,- stress at equilibrium,A - area,- strain in the material,L - length of the specimen,- change of the length.The glulam is subjected to compression deformation gXl.163The rebar’s tension deformation is equal to the difference between free contraction of therebar due to the temperature change z and the shortening of the specimen at equilibriumX1 (equation [A5]).[A5]-5Ox1..lREBARI [GLUL4M405mmFig.A.2 Deformation ofthe specimen inphase 1 (T=-5O°C).Free contraction of the rebar isr1T [A6]-5O= 11.3 x 1 06 x 405 X 50=0.23mmSubstituting the following values:E8=200 000 MPa, Eg=500 MPa, L=405 mm4%=200 mm2, Ag=175x250=43 750 mm2,the equation [A4] becomes(0.23—X1) x1 [A7]405X 200000x 200 = 500x 43750Solving equation [A7] for X1 givesX1 = 0.149 mm (contraction)164Now, the compression stress in the glulam can be found:g=Eg=°9X500=0.l84MPaTherefore the equilibrium force is:Fg=agAg=O.184X4375O=8O5ON =FFinally, the tension stress in the rebar can be calculated:F 8050aSA 200=40.25MPaphase 2- line ABThe glulam cools and shrinks, releasing the tension stress in the rebar and then,shrinking even more, puts the rebar into compression. The equilibrium is reached whenthe temperature ofthe glulam matches the surrounding temperature, i.e. -30°C (it is assumedthat the rebar reached -30°C in phase 1).The final contraction of the specimen after phase 2, X2, can be found by solving thesame relationship as before (equation [A4]). Now, however, the change of the length ofthe rebar and the glulam are as follows (see Fig.A.3):=x2—z5° [A8]Ag=t;50—X [A9]where50- free contraction of the rebar (T=-50°C),;50- free contraction of the glulam (zT=-50°C).—50=0.23mmz50=4 X106X405X50=0.8lmm1654-50TFig.A.3 Deformation ofthe specimen in phase 2 (LsT=-50°C).After substituting relations [A8] and [A9] the equation [A4] may be written as:(X2—0.23) (0.81 -X2)405 x200000x200= 405 x500x43750After solving above, the contraction of the specimen in phase 2 is:X2 = 0.435 mm (contraction)Now, the tension stress in the glulam can be found:(0.81-0.435)405X500=0.463MPaTherefore, the equilibrium force is:F0=ciAO.463X4375O=2O256N= FFinally, the compression stress in the rebar can be calculated:F 20256SA 200 =101.2SMPa[AlO]II II II IjBllARGLULAMx2H’— 405 mm166phase 3- line BCThe outside temperature changes from -30°C to +20°C (zT= +50°C). Similarly as itwas in phase 1, only the rebar is affected by the changed temperature at first.At the end of phase 2 the rebar was in compression. Now, it wants to expand due tothe rising temperature but is restrained by the glulam. The compression stress in the rebarincreases and so does the tension stress in the glulam. The total shortening of the specimen(X320m2,the floor live load can be reduced multiplying the specified occupancyload by factor (tr) calculated below:980.5 98 0.5tro.3+(t) o.3+(45) ‘0.79therefore the floor live load is:1f=0.79X2.40= 1.90 kN/m2.b) roof live loadS5 specified ground snow load 2.50 kN/m2Sr associated rain load 0.30 kN/m2Cb basic roof snow load factor 0.80C wind exposure factor 1.00Cf roof slope factor 1.00 (flat roof)Ca accumulation factor 1.00176The specified snow load, S, can be calculated follows:S=Ss(CbCwCfCa)+SrS = 2.5x(0.8x1.Oxl.Oxl.0)+0.3 = 2.3 kN/m2.III. Wind Loads.(l/3O) reference velocity pressure (1/30) 0.55 kNIm2(1/1O) reference velocity pressure (1/10) 0.45 kN/m2h reference height of the building 13.0 mIt is assumed that exposure factor Ce is constant over the entire height of the building andis equal to:i.oa) wind pressure on the windward wallCpCg peak pressure coefficient 0.75The specffied external pressure is equal to:p = q Ce CpCgthereforeP(113o) = O.55X 1 .OSX 0.75 = 0.43 kN/m2P(I,io) = 0.45 x 1 .OSX 0.75 = 0.35 kN/m2b) wind suction on the leeward wallCpCg peak pressure coefficient 0.55The specified external suction is equal to:177p =qCeCpCgthereforeP(L/3o) = 0.55 X 1 .05X 0.55 = 0.32 kJJ/m2p(1110)-0.45X 1.OSxO.55=0.26 kNIm2c) wind suction on the roofCpCg peak pressure coefficient 1.30The specified external suction is equal to:p =qCeCpCgthereforep(1/30)—0.55X 1.05x 1.30=0.75 kN/m2P(l/Io) = 0.45x 1 .05x 1.30 = 0.61 kN/m2IV. Earthquake Loads.U calibration factor 0.6R force modffication factor 2.0 (moment-resisting wood frame with ductile connectors)Note: The values of the factor R, permitted by the code, do not include the case ofglued-in rebar connections. Assuming rigidity of thejoints and failure in the steel,this structure could be treated as a steel structure.v zonal velocity ratio 0.2 (Vancouver)Za acceleration seismic zone 4 (Vancouver)Zv velocity seismic zone 4 (Vancouver)178F foundation factor 1.0 (dense soil)I seismic importance factor 1.0 (nominal)N number of storeys 3The fundamental period of the building, T, is:T=0.1N=0.1x3 =0.3stherefore, the seismic response factor, S, is equal to:S = 3.0-3.6x(T-0.25) = 3.0-3.6x(0.3-0.25) = 2.82The weight, W, of the structure is estimated to be as follows:25% of the snow load 0.25x2.30x4.5x63.0 164 kN100% of roof dead load 1.15x4.5x63.0 326 kN100% of floors dead load 2x1.90x4.5x63.0 1077 kN100% of self weight of frame 20m3x5.3kN/m 101 kNTOTAL: 1668 kNThe equivalent lateral seismic force, Ve, can be calculated in accordance with the followingformula:Ve =vSIFWthereforeVe = 0.2x2.82x 1.Ox 1.Ox 1668 = 941 kN.The minimum lateral seismic force (base shear), V, is equal to:V = (Ve/R)UthereforeV = (94112.0)xO.8 = 282 kN.The above seismic force at the base of the structure, V, is calculated for one frame, i.e. itis 1/8 of the total base shear of the whole building.179The distribution of the minimum lateral seismic force along the height of the building iscalculated according to the formula given below:F(V—F)Wha) roof seismic forceThe concentrated force at the top Ft = 0, because period T<0.7s.The weight of the roof (including snow) is equal to:Wr = 164 + 326 + 101/3 524kNThe seismic force at the roof level is:F(VF1)Wrhr— (282-O)X524x 12.0 —r— 524X12.0*572X8.0.572X4.0 —b) second storey seismic forceThe weight of the floor is equal to:W = 107712 + 10113 = 572 kNThe seismic force at the second floor level is:F(V-F)Wh5— (282—0)x572x8.0 —S— 524X12.0.572X8.0.S72x4.0 —180c) first storey seismic forceThe weight of the roof (including snow) is equal to:WfWs =572kNThe seismic force at the first floor level is:- (V — F)W,h,F1— NZWhC — (282-O)X572X4.O— 5o‘ I 524X 12.O*572X8.O-572X4.O —3. LOAD COMBINATIONS.The following factored load combinations are included in the analysis:a. 1.25x(Dead Loads) ÷ 1.5x(Live Loads)b. 1.25x(Dead Loads) + 1.5x(Wind Loads)c. O.85x(Dead Loads) + 1.5x(Wind Loads)d. 1.25x(Dead Loads) + 1.Ox(Seismic Forces)e. O.85x(Dead Loads) + 1.Ox(Seismic Forces)f. 1.25x(Dead Loads) + O.7x{1.5x(Live Loads) + 1.5x(Wind Loads))g. 1.25x(Dead Loads) + O.7x{1.5x(Live Loads) + 1.Ox(Seismic Forces)}A load combination giving the largest internal forces in the element is chosen to design thatelement.The load combinations involving partial loading of the structure were found to give less criticalinternal forces than full loading.1814. FRAME ANALYSIS.The static analysis of the frame was performed using a finite element method program calledPLANE 3. The model of the structure used for the computations is presented on Fig.B.2.Y J RIGID JOINT NODE NUMBER[m] ELEMENT NUMBER0 PINNED JOINT NODE NUMBER6___8 2 12 8 /6______ ___96___3 9 9 45__11 /____19____42 8 44__6 0 10 6 14 18 40 ?1 7 3 /9 3 40_______5__________9 13 /7 2/ 2 29//////////// ///////////// ///////////// ///////////// ///////////// ///////////// /////////////0 9 18 ‘27 36 ‘45 54 63Fig.B.2 Model of the frame used in the numerical analysis.The factored loads applied to the frame in each load case are shown in Table B.1.uniformly distr. load combinationsloads [kN/m] element No. direc. a b c d e f g— — — — — — —roof beams 6,12,18,24,30,36,42 -Y 22.0 1.4 -0.7 6.5 4.4 13.8 17.3floor beams 4,5,10,11,16,17,22,23, -Y 24.1 11.3 7.7 11.3 7.6 20.2 20.228,29,34,35,40,41windward columns 1,2,3 X - 2.9 2.9 - - 2.0 -leeward columns 43,44,45 X - 2.1 2.1 - - 1.5 -— — — — — — —Table B.1 ...continues on the nextpage...182concentrated load combinationsloads [kNJ node No. direc. a b c d e f g— — — — — — —[st St. seismic force 2,6,10,14,18,22,26 X - - - 7.2 7.2 - 5.02nd St. seismic force 3,7,11,15,19,23,27 X - - - 14.4 14.4 - 10.1roof seismic force 4,8,12,16,20,24,28 X - - - 18.7 18.7 - 13.1Table B.1 Load cases used in the numerical analysis ofthe frame.5. RESULTS OF THE ANALYSIS.The internal forces in the frame, due to the load combinations described in section 4, arepresented below.Load combination “A” 1.25*D+1.5*LBeam Forcesmember axiall shearl bml axial2 shear2 bm2[kN] [kN] [kNm] [kN] [kN] [kNm]1 -290.9 -15.9 0.0 -290.9 -15.9 -63.72 -189.8 -26.3 49.9 -189.8 -26.3 -55.23 -88.6 -34.6 60.2 -88.6 -34.6 -78.34 10.3 101.0 -113.6 10.3 -115.9 -180.35 8.3 101.2 -115.3 8.3 -115.7 -180.36 -34.6 88.6 -78.3 -34.6 -109.4 -172.07 -659.7 1.6 0.0 -659.7 1.6 6.4Table B.2 ...continues on the nextpage...1838 -434.4 2.5 -5.3 -434.4 2.5 4.99 -209.9 5.7 -9.5 -209.9 5.7 13.310 9.4 109.4 -168.6 9.4 -107.5 -160.411 5.2 108.9 -165.8 5.2 -108.0 -161.812 -28.9 100.5 -158.7 -28.9 -97.5 -145.513 -628.4 -0.3 0.0 -628.4 -0.3 -1.314 -412.5 -0.2 0.4 -412.5 -0.2 -0.415 -196.2 -0.2 -0.1 -196.2 -0.2 -0.716 9.3 108.4 -162.1 9.3 -108.5 -162.817 5.1 108.3 -162.1 5.1 -108.6 -163.118 -29.1 98.7 -146.2 -29.1 -99.3 -149.419 -632.3 0.0 0.0 -632.3 0.0 0.020 -415.4 0.0 -0.1 -415.4 0.0 0.021 -198.3 0.3 -0.5 -198.3 0.3 0.722 9.2 108.5 -162.8 9.2 -108.5 -162.823 4.8 108.5 -162.6 4.8 -108.5 -162.624 -28.8 99.0 -148.7 -28.8 -99.0 -148.725 -632.3 0.0 0.0 -632.3 0.0 0.026 -415.4 0.0 0.1 -415.4 0.0 0.027 -198.3 -0.3 0.5 -198.3 -0.3 -0.728 9.3 108.5 -162.8 9.3 -108.4 -162.129 5.1 108.6 -163.1 5.1 -108.3 -162.130 -29.1 99.3 -149.4 -29.1 -98.7 -146.231 -628.4 0.3 0.0 -628.4 0.3 1.332 -412.5 0.2 -0.4 -412.5 0.2 0.433 -196.2 0.2 0.1 -196.2 0.2 0.734 9.4 107.5 -160.4 9.4 -109.4 -168.6Table B.2 ...continues on the nextpage...18435 5.2 108.0 -161.8 5.2 -108.9 -165.836 -28.9 97.5 -145.5 -28.9 -100.5 -158.737 -659.7 -1.6 0.0 -659.7 -1.6 -6.438 -434.4 -2.5 5.3 -434.4 -2.5 -4.939 -209.9 -5.7 9.5 -209.9 -5.7 -13.340 10.3 115.9 -180.3 10.3 -101.0 -113.641 8.3 115.7 -180.3 8.3 -101.2 -115.342 -34.6 109.4 -172.0 -34.6 -88.6 -78.343 -290.9 15.9 0.0 -290.9 15.9 63.744 -189.8 26.3 -49.9 -189.8 26.3 55.245 -88.6 34.6 -60.2 -88.6 34.6 78.3Support Reactionsnode Rx Ry M[kN] [kNI [kNm]1 15.9 290.9 0.05 -1.6 659.7 0.09 0.3 628.4 0.013 0.0 632.3 0.017 0.0 632.3 0.021 -0.3 628.4 0.025 1.6 659.7 0.029 -15.9 290.9 0.0Table B.2 Internalforces due to the load combination “a”.185Load combination “B” 1.25*D+1.5*WoBeam Forcesmember axiall shearl bml axial2 shear2 bm2[kN] [kN] [kNm] [kN] [kN] [kNm]1 -93.8 2.5 0.0 -93.8 -9.1 -13.22 -51.2 -5.3 17.0 -51.2 -16.9 -27.33 -6.3 0.2 12.7 -6.3 -11.4 -9.84 -3.8 42.6 -30.2 -3.8 -59.1 -104.85 -17.1 45.0 -40.0 -17.1 -56.7 -92.96 -11.4 6.3 -9.8 -11.4 -6.3 -10.27 -225.6 7.8 0.0 -225.6 7.8 31.38 -119.1 6.5 -11.7 -119.1 6.5 14.19 -12.2 1.9 -4.3 -12.2 1.9 3.210 -2.5 47.4 -61.8 -2.5 -54.3 -92.611 -12.5 50.2 -74.5 -12.5 -51.5 -80.712 -9.6 5.8 -7.0 -9.6 -6.8 -11.213 -215.0 6.7 0.0 -215.0 6.7 26.914 -113.8 4.1 -7.4 -113.8 4.1 8.915 -12.8 1.2 -2.0 -12.8 1.2 3.016 0.2 46.9 -58.4 0.2 -54.8 -93.917 -9.6 49.5 -69.8 -9.6 -52.2 -82.318 -8.3 6.0 -8.3 -8.3 -6.6 -10.819 -216.2 6.9 0.0 -216.2 6.9 27.520 -114.4 4.2 -7.6 -114.4 4.2 9.3Table B.3 ...continues on the nextpage...18621 -12.6 1.3 -2.3 -12.6 1.3 2.922 2.8 47.0 -58.8 2.8 -54.7 -93.723 -6.7 49.6 -70.6 -6.7 -52.1 -82.024 -7.0 6.0 -7.9 -7.0 -6.6 -10.925 -216.1 6.9 0.0 -216.1 6.9 27.626 -114.4 4.0 -7.2 -114.4 4.0 8.927 -12.6 1.3 -2.2 -12.6 1.3 2.928 5.7 47.0 -58.9 5.7 -54.7 -93.529 -3.9 49.7 -70.9 -3.9 -52.0 -81.330 -5.8 6.0 -8.0 -5.8 -6.6 -11.131 -215.2 7.0 0.0 -215.2 7.0 28.032 -113.8 4.2 -7.3 -113.8 4.2 9.333 -12.7 1.3 -2.5 -12.7 1.3 2.934 8.5 46.7 -58.1 8.5 -55.0 -95.535 -1.1 49.1 -69.5 -1.1 -52.6 -85.436 -4.4 6.1 -8.2 -4.4 -6.5 -10.437 -223.5 6.5 0.0 -223.5 6.5 26.038 -118.9 2.3 -4.2 -118.9 2.3 5.039 -12.7 0.6 -0.1 -12.7 0.6 2.140 12.7 49.6 -65.3 12.7 -52.1 -76.641 0.6 53.6 -80.2 0.6 -48.1 -55.142 -3.9 6.1 -8.3 -3.9 -6.5 -9.843 -106.6 15.7 0.0 -106.6 7.3 46.044 -54.5 20.0 -30.6 -54.5 11.6 32.745 -6.5 12.3 -22.4 -6.5 3.9 9.8Table B.3 ...continues on the nextpage...187Support Reactionsnode P.x Ry M[kN] [kN] [kNm]1 -2.5 93.8 0.05 -7.8 225.6 0.09 -6.7 215.0 0.013 -6.9 216.2 0.017 -6.9 216.1 0.021 -7.0 215.2 0.025 -6.5 223.5 0.029 -15.7 106.6 0.0Table B.3 Internalforces due to the load combination “b”.Load combination “C” O.85*D+1.5*WoBeam Forcesmember axiall shearl bml axial2 shear2 bm2[kN] [kN] [kNm] [kN] [kN] [kNm]1 -54.3 4.9 0.0 -54.3 -6.7 -3.52 -27.3 -1.1 9.2 -27.3 -12.7 -18.33 2.3 4.2 4.9 2.3 -7.4 -1.74 -5.6 27.1 -12.7 -5.6 -41.3 -77.15 -16.8 29.6 -23.2 -16.8 -38.8 -65.06 -7.4 -2.3 -1.7 -7.4 4.0 6.0Table B.4 ...continues on the nextpage...1887 -136.5 7.6 0.0 -136.5 7.6 30.48 -64.5 5.9 -10.7 -64.5 5.9 13.19 7.7 1.3 -3.1 7.7 1.3 2.010 -4.0 30.6 -36.0 -4.0 -37.8 -68.011 -12.2 33.4 -48.8 -12.2 -35.0 -55.912 -6.2 -3.7 8.0 -6.2 2.6 2.713 -129.9 6.8 0.0 -129.9 6.8 27.114 -61.9 4.1 -7.4 -61.9 4.1 9.015 6.0 1.2 -2.0 6.0 1.2 3.016 -1.3 30.3 -33.5 -1.3 -38.1 -68.917 -9.3 32.8 -45.0 -9.3 -35.6 -57.218 -4.9 -3.4 5.7 -4.9 2.9 3.419 -130.6 6.9 0.0 -130.6 6.9 27.520 -62.1 4.2 -7.6 -62.1 4.2 9.321 6.4 1.3 -2.2 6.4 1.3 2.822 1.4 30.3 -33.9 1.4 -38.1 -68.723 -6.4 32.9 -45.7 -6.4 -35.5 -57.024 -3.7 -3.5 6.2 -3.7 2.8 3.325 -130.6 6.9 0.0 -130.6 6.9 27.626 -62.1 4.1 -7.3 -62.1 4.1 8.927 6.3 1.3 -2.3 6.3 1.3 3.028 4.2 30.3 -33.9 4.2 -38.1 -68.629 -3.6 33.0 -45.8 -3.6 -35.4 -56.530 -2.3 -3.5 6.3 -2.3 2.8 3.031 -130.1 7.0 0.0 -130.1 7.0 27.832 -61.9 4.1 -7.3 -61.9 4.1 9.2Table B.4 ...continues on the nextpage...18933 6.1 1.3 -2.5 6.1 1.3 2.834 7.1 30.2 -33.5 7.1 -38.2 -69.735 -0.8 32.6 -44.8 -0.8 -35.8 -59.636 -1.0 -3.3 5.8 -1.0 3.0 4.637 -134.4 6.7 0.0 -134.4 6.7 26.938 -64.4 2.8 -5.2 -64.4 2.8 6.139 7.2 1.1 -1.3 7.2 1.1 3.340 11.0 31.8 -37.6 11.0 -36.6 -59.141 0.8 35.7 -52.2 0.8 -32.7 -38.342 0.1 -4.2 7.9 0.1 2.1 -1.743 -67.2 13.3 0.0 -67.2 4.9 36.344 -30.6 15.8 -22.8 -30.6 7.4 23.745 2.1 8.3 -14.6 2.1 -0.1 1.7Support Reactionsnode Rx Ry M[kN] [kN] [kNm]1 -4.9 54.3 0.05 -7.6 136.5 0.09 -6.8 129.9 0.013 -6.9 130.6 0.017 -6.9 130.6 0.021 -7.0 130.1 0.025 -6.7 134.4 0.029 -13.3 67.2 0.0Table B.4 Internalforces due to the load combination “c’190Load combination “D” 1.25*D+1.O*EqBeam Forcesmember axiall shearl bml axial2 shear2 bm2[kN] [kN] [kNm] [kN] [kNI [kNm]1 -73.2 19.5 0.0 -73.2 19.5 78.12 -54.8 6.7 -9.9 -54.8 6.7 17.03 -21.2 -0.5 3.0 -21.2 -0.5 1.04 5.6 18.4 88.0 5.6 -83.3 -204.05 -7.2 33.6 14.0 -7.2 -68.1 -141.46 -19.2 21.2 1.0 -19.2 -37.3 -71.47 -283.1 39.8 0.0 -283.1 39.8 159.28 -171.3 34.9 -66.8 -171.3 34.9 72.89 -62.9 20.3 -38.4 -62.9 20.3 42.910 3.3 28.5 21.9 3.3 -73.2 -179.011 -7.0 40.3 -30.1 -7.0 -61.4 -125.112 -17.6 25.6 -28.5 -17.6 -32.9 -61.613 -259.4 37.3 0.0 -259.4 37.3 149.414 -158.9 31.5 -59.7 -158.9 31.5 66.215 -57.9 17.6 -33.2 -57.9 17.6 37.216 1.9 27.3 30.1 1.9 -74.4 -182.017 -7.5 39.6 -25.6 -7.5 -62.1 .126.618 -18.7 25.0 -24.5 -18.7 -33.5 -62.919 -262.3 37.7 0.0 -262.3 37.7 150.720 -160.4 31.8 -60.3 -160.4 31.8 66.7Table B.5 ...continues on the nextpage...19121 -58.6 17.8 -33.6 -58.6 17.8 37.722 0.7 27.5 29.0 0.7 -74.2 -181.523 -8.0 39.7 -26.3 -8.0 -62.0 -126.224 -19.6 25.1 -25.2 -19.6 -33.4 -62.625 -262.0 37.7 0.0 -262.0 37.7 150.726 -160.3 31.6 -60.0 -160.3 31.6 66.527 -58.6 17.6 -33.3 -58.6 17.6 37.328 -0.5 27.4 29.2 -0.5 -74.3 -181.929 -8.4 39.8 -26.5 -8.4 -61.9 -125.930 -20.6 25.2 -25.3 -20.6 -33.3 -62.231 -261.9 37.6 0.0 -261.9 37.6 150.432 -159.8 31.5 -59.7 -159.8 31.5 66.433 -58.2 17.6 -33.2 -58.2 17.6 37.334 -1.6 27.8 28.2 -1.6 -73.9 -179.435 -8.9 39.7 -26.3 -8.9 -62.0 -127.036 -21.7 24.9 -24.8 -21.7 -33.6 -64.037 -263.2 38.3 0.0 -263.2 38.3 153.138 -163.7 31.5 -60.3 -163.7 31.5 65.839 -60.4 16.5 -31.3 -60.4 16.5 34.940 -2.0 25.5 34.0 -2.0 -76.2 -193.941 -8.3 41.3 -29.9 -8.3 -60.4 -116.042 -23.8 26.8 -29.1 -23.8 -31.7 -50.943 -168.3 34.2 0.0 -168.3 34.2 136.844 -92.1 32.2 -57.1 -92.1 32.2 71.545 -31.7 23.8 -44.5 -31.7 23.8 50.9Table B.5 ...continues on the nextpage...192Support Reactionsnode R.x Ry M[kN] [kN] [kNm]1 -19.5 73.2 0.05 -39.8 283.1 0.09 -37.3 259.4 0.013 -37.7 262.3 0.017 -37.7 262.0 0.021 -37.6 261.9 0.025 -38.3 263.2 0.029 -34.2 168.3 0.0Table B.5 Internalforces due to the load combination “d”.Load combination “E” O.85*D+1.O*EqBeam Forcesmember axiall shearl bml axial2 shear2 bm2[kN] [kN] [kNm] [kN] [kN] [kNm]1 -33.7 21.9 0.0 -33.7 21.9 87.72 -30.8 10.9 -17.7 -30.8 10.9 26.03 -12.7 3.5 -4.8 -12.7 3.5 9.14 3.8 2.9 105.5 3.8 -65.5 -176.35 -6.9 18.2 30.8 -6.9 -50.2 -113.46 -15.2 12.7 9.1 -15.2 -26.9 -55.2Table B.6 ...contiizues on the nextpage...1937 -194.0 39.6 0.0 -194.0 39.6 158.38 -116.7 34.4 -65.8 -116.7 34.4 71.79 -43.0 19.7 -37.3 -43.0 19.7 41.610 1.8 11.7 47.8 1.8 -56.7 -154.411 -6.7 23.5 -4.4 -6.7 -44.9 -100.312 -14.2 16.0 -13.6 -14.2 -23.6 -47.713 -174.3 37.4 0.0 -174.3 37.4 149.614 -107.0 31.5 -59.7 -107.0 31.5 66.315 -39.1 17.6 -33.3 -39.1 17.6 37.216 0.5 10.6 55.0 0.5 -57.8 -157.017 -7.2 23.0 -0.8 -7.2 -45.4 -101.618 -15.3 15.6 -10.5 -15.3 -24.0 -48.619 -176.7 37.7 0.0 -176.7 37.7 150.820 -108.2 31.7 -60.2 -108.2 31.7 66.721 -39.7 17.8 -33.6 -39.7 17.8 37.622 -0.8 10.8 54.0 -0.8 -57.6 -156.623 -7.7 23.1 -1.3 -7.7 -45.3 -101.324 -16.2 15.6 -11.0 -16.2 -24.0 -48.425 -176.4 37.7 0.0 -176.4 37.7 150.726 -108.1 31.6 -60.0 -108.1 31.6 66.527 -39.6 17.7 -33.3 -39.6 17.7 37.328 -1.9 10.7 54.2 -1.9 -57.7 -157.029 -8.1 23.1 -1.5 -8.1 -45.3 -101.130 -17.2 15.7 -11.1 -17.2 -23.9 -48.231 -176.8 37.5 0.0 -176.8 37.5 150.232 -107.9 31.5 -59.6 -107.9 31.5 66.3Table B.6 ...continues on the nextpage...19433 -39.5 17.6 -33.2 -39.5 17.6 37.334 -3.1 11.3 52.8 -3.1 -57.1 -153.635 -8.6 23.1 -1.5 -8.6 -45.3 -101.236 -18.3 15.6 -10.9 -18.3 -24.0 -49.037 -174.1 38.5 0.0 -174.1 38.5 154.138 -109.2 32.1 -61.3 -109.2 32.1 66.939 -40.5 17.1 -32.4 -40.5 17.1 36.140 -3.8 7.7 61.7 -3.8 -60.7 -176.441 -8.1 23.4 -1.9 -8.1 -45.0 -99.242 -19.9 16.5 -12.9 -19.9 -23.1 -42.743 -128.8 31.8 0.0 -128.8 31.8 127.144 -68.1 28.0 -49.3 -68.1 28.0 62.545 -23.1 19.9 -36.7 -23.1 19.9 42.7Support Reactionsnode Rx Ry M[kN] [kN] [kNm]1 -21.9 33.7 0.05 -39.6 194.0 0.09 -37.4 174.3 0.013 -37.7 176.7 0.017 -37.7 176.4 0.021 -37.5 176.8 0.025 -38.5 174.1 0.029 -31.8 128.8 0.0Table B.6 Internalforces due to the load combination “e”.195Load combination “F” 1.25*1) +O.7*(1.5*1+ 1.5Wo)Beam Forcesmember axiall shearl bml axial2 shear2 bm2[kN] [kNI [kNm] [kNj [kN] [kNm]1 -220.5 -6.5 0.0 -220.5 -14.5 -41.92 -139.2 -16.9 37.2 -139.2 -24.9 -46.63 -55.9 -19.3 41.4 -55.9 -27.3 -51.74 2.5 81.3 -79.1 2.5 -100.5 -165.25 -5.7 83.3 -88.0 -5.7 -98.5 -156.66 -27.3 55.9 -51.7 -27.3 -68.3 -107.67 -509.6 6.3 0.0 -509.6 6.3 25.28 -320.1 5.7 -10.6 -320.1 5.7 12.09 -131.0 4.6 -8.2 -131.0 4.6 10.010 3.1 89.0 -129.4 3.1 .92.8 -146.611 -4.6 90.6 -136.4 -4.6 -91.2 -139.112 -22.7 62.7 -97.6 -22.7 -61.5 -92.613 -485.2 4.5 0.0 -485.2 4.5 18.214 -304.3 2.7 -4.9 -304.3 2.7 6.015 -123.2 0.8 -1.5 -123.2 0.8 1.616 4.9 88.1 -123.5 4.9 -93.7 -148.817 -2.6 89.9 -131.6 -2.6 -91.9 -140.718 -21.9 61.7 -91.0 -21.9 -62.5 -94.619 -488.2 4.8 0.0 -488.2 4.8 19.220 -306.3 3.0 -5.3 -306.3 3.0 6.5Table B.7 ...continues on the nextpage...19621 -124.4 1.1 -1.9 -124.4 1.1 2.422 6.8 88.2 -124.2 6.8 -93.6 -148.623 -0.8 90.0 -132.3 -0.8 -91.8 -140.324 -20.8 61.9 -92.2 -20.8 -62.3 -94.325 -488.2 4.8 0.0 -488.2 4.8 19.326 -306.3 2.8 -5.1 -306.3 2.8 6.327 -124.4 0.7 -1.3 -124.4 0.7 1.628 8.8 88.2 -124.3 8.8 -93.6 -148.129 1.4 90.1 -132.7 1.4 -91.7 -139.630 -20.1 62.1 -92.6 -20.1 -62.1 -92.931 -485.4 5.1 0.0 -485.4 5.1 20.332 -304.3 3.0 -5.4 -304.3 3.0 6.733 -123.2 1.0 -1.6 -123.2 1.0 2.534 10.8 87.5 -122.5 10.8 -94.3 -152.935 3.4 89.5 -131.3 3.4 -92.3 -143.936 -19.1 61.0 -90.4 -19.1 -63.2 -100.037 -508.1 3.7 0.0 -508.1 3.7 14.938 -320.0 0.5 -0.5 -320.0 0.5 1.439 -131.3 -2.9 5.2 -131.3 -2.9 -6.340 14.1 93.8 -137.6 14.1 -88.0 -111.541 6.7 96.4 -147.7 6.7 -85.4 -98.642 -22.0 68.2 -106.3 -22.0 -56.0 -51.743 -229.5 19.2 0.0 -229.5 13.2 64.844 -141.5 27.3 -46.7 -141.5 21.3 50.445 -56.0 28.0 -48.2 -56.0 22.0 51.7Table B.7 ...continues on the nextpage...197Support ReactionsBeam ForcesTable B.7 Interiwlforces due to the load combination “f.Load combination “G” 1.25*D +O.7*(1.5*L+ 1.O*Eq)member axiall shearl bml axial2 shear2 bm2[kN] [kN] [kNm] [kN] [kN] [kNm]1 -205.8 5.5 0.0 -205.8 5.5 22.02 -141.4 -8.5 18.3 -141.4 -8.5 -15.63 -66.1 -19.6 34.5 -66.1 -19.6 -43.94 9.0 64.4 3.7 9.0 -117.4 -234.75 1.0 75.3 -50.1 1.0 -106.5 -190.66 -32.7 66.1 -43.9 -32.7 -89.6 -149.9Table B.8 ...continues on the nextpage...node Rx Ry M[kN] [kN] [kNm]1 6.5 220.5 0.05 -6.3 509.6 0.09 -4.5 485.2 0.013 -4.8 488.2 0.017 -4.8 488.2 0.021 -5.1 485.4 0.025 -3.7 508.1 0.029 -19.2 229.5 0.01987 -549.1 28.7 0.0 -549.1 28.7 114.78 -356.0 25.6 -49.3 -356.0 25.6 53.29 -165.8 17.5 -32.2 -165.8 17.5 37.810 7.0 75.7 -70.7 7.0 -106.1 -207.111 -0.9 83.7 -105.3 -0.9 -98.1 -170.212 -28.3 76.2 -112.2 -28.3 -79.5 -127.413 -515.7 26.0 0.0 -515.7 26.0 103.814 -335.3 22.0 -41.6 -335.3 22.0 46.215 -154.2 12.2 -23.4 -154.2 12.2 25.616 6.0 74.4 -61.6 6.0 -107.4 -210.517 -1.3 83.0 -100.6 -1.3 -98.8 -171.818 -29.2 74.7 -101.8 -29.2 -81.0 -130.619 -519.9 26.4 0.0 -519.9 26.4 105.520 -337.9 22.2 -42.3 -337.9 22.2 46.721 -156.0 12.7 -23.8 -156.0 12.7 26.822 5.2 74.5 -62.7 5.2 -107.3 -210.123 -1.9 83.1 -101.3 -1.9 -98.7 -171.324 -29.6 74.9 -103.8 -29.6 -80.8 -130.025 -519.6 26.4 0.0 -519.6 26.4 105.526 -337.8 22.2 -42.0 -337.8 22.2 46.627 -156.0 12.2 -23.0 -156.0 12.2 25.728 4.4 74.5 -62.6 4.4 -107.3 -210.029 -2.0 83.2 -101.7 -2.0 -98.6 -170.930 -30.5 75.2 -104.3 -30.5 -80.5 -128.231 -517.5 26.5 0.0 -517.5 26.5 105.932 -335.9 22.2 -42.1 -335.9 22.2 46.7Table B.8 ...continues on the nextpage...19933 -154.4 12.4 -23.2 -154.4 12.4 26.634 3.6 74.3 -62.0 3.6 -107.5 -211.735 -2.3 82.9 -101.0 -2.3 -98.9 -173.136 -31.2 73.9 -101.6 -31.2 -81.8 -137.037 -535.2 26.0 0.0 -535.2 26.0 103.838 -350.7 21.0 -39.8 -350.7 21.0 44.039 -164.1 8.3 -16.7 -164.1 8.3 16.740 3.6 76.9 -68.0 3.6 -104.9 -193.741 0.2 87.7 -112.5 0.2 -94.1 -141.242 -36.0 82.3 -120.3 -36.0 -73.4 -80.243 -272.3 32.1 0.0 -272.3 32.1 128.344 -167.5 35.7 -65.3 -167.5 35.7 77.645 -73.4 36.0 -63.6 -73.4 36.0 80.2Support Reactionsnode Rx Ry M[kN] [kN] [kNm]1 -5.5 205.8 0.05 -28.7 549.1 0.09 -26.0 515.7 0.013 -26.4 519.9 0.017 -26.4 519.6 0.021 -26.5 517.5 0.025 -26.0 535.2 0.029 -32.1 272.3 0.0Table B.8 Internalforces due to the load combination ‘200The internal maximum forces governing the design of the elements of the frame aresummarized in Table B.9.element internal force governing design load combinationfirst storey beam bending moment 235 kNm Gsecond storey beam bending moment 191 kNm Groof beam bending moment 172 kNm Afirst storey column bending moment 159 kNm Daxial force 283 kNsecond storey column bending moment 78 kNm Gaxial force 168 kNthird storey column bending moment 80 kNm Gaxial force 73 kNTable B.9 Internalforces governing the design ofthe elements.6. DESIGN OF THE MEMBERS.The design calculations are performed according to section 6 of CAN/CSA-086.1-M89.The glulam 24f-E, D.Fir-L. shall be used as material for the frame’s members.I. Design of the Beams.The following factors and material properties are used in calculations ofthe factored bendingmoment resistance and the factored shear resistance:4 resistance factor 0.9KD load duration factor 1.0 (standard term)KH system factor 1.0 (nominal)KSb service condition factor for bending 1.0 (dryservice conditions)201Ks service condition factor for shear 1.0 (dryservice conditions)KSE service condition factor for M.O.E. 1.0 (dryservice conditions)KT treatment factor 1.0 (non-treated glulam)KX curvature factor for glulam 1.0 (straight members)KN notch factor 1.0 (free of notches)i, specified strength in bending 30.6 MPaspecified strength in shear 2.0 MPaE Modulus of elasticity 13100 MPaThe factored moment resistance of the beam is given by the formula:Mr = FbSKLKxwhereFb= fb(KDKHKSbKT) = 30.6 MPaS section modulusKL lateral stability factorCK= (0.97EKSEKT/Fb)°•5= 20.38The factored shear resistance of the beam is calculated according to the formula:Vr = 4FO.6AKNCvZ°’8 WIwhereFv= fv(KDKHK5vKT)=2.O MPaA cross-sectional area of beam, mm2C shear load coefficientZ Beam volume, m3202a) first storey beamMf=235kNmAssume glulam section 130x684mm.A= 130x684=88920mm2S = 130x6842/ = 10.14 x106mm3Z = 0.13x0.684x9.0 = 0.8 m3theoretical length 1 = 9000 mmMOMENT RESISTANCEThe unsupported length of the beam, lu’ is assumed to be 1/3 because this is theapproximate length of the bottom edge of the beam subjected to compression stresses.The effective length of the beam, Le, is:Le = ‘•921u = 1.92x9000/3 =5760mmThe slenderness ratio of the beam, CB, is:1Ld’\°5 (5760x684’\°C8=I—I =1 I =15.27b2) 130 Il0CBCKThe lateral stability factor is, therefore, equal to:iIc84 1(15.27’\KL=l—3--) 132o38) =0.89Now, the moment resistance may be calculated:Mr = 0.9X 30.6X 10.14X 106X 0.89x 1.0 = 249.8x lO6Nmm 249.8kNmMr>Mf235kNm O.K.203SHEAR RESISTANCEThe sum of all factored loads, Wf, acting on the beam is:Wf=9.0x24.1 =216.900kNThe shear load coefficient is:C = 3.74Now, the shear resistance of the beam may be calculated:Vr = O.9X 2.OX O.6X 88920x 1 .OX 3.74 x 0.8-0.18 = 378300NVr>Wf O.K.b) second storey beamMf= 191 kNmAssume glulam section 130x608mm.A= 130x608 = 79040mm2S = 130x6082/ = 8.01 x106mm3Z = 0.13x0.608x9.0 = 0.71 m3theoretical length 1 = 9000 mmMOMENT RESISTANCEThe effective length of the beam, Le, is:Le = 1921u = 1.92x9000/3 = 5760mmThe slenderness ratio of the beam, CB, is:(Led\°5 I5760x6O8’105C9=(—i-) t.. 1302 J =14.401OC8C<204The lateral stability factor is, therefore, equal to:KL= 1_()4=1_(:)4= 0.92Now, the moment resistance may be calculated:Mr0.9X30.6X8.Ol x 106x0.92x 1.O=202.3x lO6Nmm= 202.3kNmMr>Mf=l9lkNm QLSHEAR RESISTANCEThe sum of all factored loads, Wf, acting on the beam is:Wf=9.0x24.1 =216.900kNThe shear load coefficient is:C = 3.70Now, the shear resistance of the beam may be calculated:Vr = 0.9X 2.0X 0.6X 79040X 1 .0x 3.70X 0.71 -0J8 = 339000NVr>Wf O.K.c) roof beamMf= l72kNmAssume glulam section 130x570mm.A = 130x570 = 74 100 mm2S = 130x5702/6= 7.04 x106mm3Z = 0.13x0.570x9.O = 0.67 m3theoretical length 1 = 9000 mm205MOMENT RESISTANCEThe effective length of the beam, Le, is:Le = 1.92 l = l.92x9000/3 = 5760 mmThe slenderness ratio of the beam, CB, is:(Led’°5 A”5760X570°C8=I—1 =1 1 =13.94b2J \. 1302,,,10CBCKThe lateral stability factor is, therefore, equal to:1(13.94KL=l—3--) 13l2o38) =0.93Now, the moment resistance may be calculated:Mr = 0.9X 30.6X 7.04X 106X 0.93X 1.0= 179.7x lO6Nmm 179.7kNmMr>Mfl72kNmSHEAR RESISTANCEThe sum of all factored loads, Wf, acting on the beam is:Wf = 9.0 x 22.0 = 198.900 kNThe shear load coefficient is:= 3.69Now, the shear resistance of the beam may be calculated:Vr = 0.9X 2.0X 0.6X 74100x 1.OX 3.69x 0.670.18 = 320000NVr>Wf206II. Design of the Columns.The following factors and material properties are used in calculations of the factoredresistance to combined bending and axial load:4> resistance factor 0.9KD load duration factor 1.0 (standard term)KH system factor 1.0 (nominal)KSb service condition factor for bending 1.0 (dry service conditions)KSc service cond. factor for compr. parallel 1.0 (dry service conditions)KSE service condition factor for M.O.E. 1.0 (dry service conditions)KT treatment factor 1.0 (non-treated glulam)KX curvature factor for glulam 1.0 (straight members)fi3 specified strength in bending 30.6 MPac specified strength in compr. parallel 20.4 MPaE modulus of elasticity 13100 MPaE05 modulus of elasticity - 5th percentile 11400 MPaThe factored compressive load resistance parallel to the grain, r, of the column is givenby the formula:1r = 4FAKwhereFc= fc(KDKHKScKT) = 20.4 MPaA cross-sectional area of member, mm2KC slenderness factorThe factored bending moment resistance of the column is calculated as in section 6.L207The columns should satisfy the following interaction equation:PM—+— 1.01r MrwherePf factored compressive axial loadMf factored bending moment amplified due to axial loadsa) first storey columnPf=283kNM’= l59kNmAssume glulam section 175x532mm.A=175x532=93100mm2S = 175x5322/6= 8.26 x106mm3I = 175x5323/12 = 2.20 x109mm4theoretical length 1 = 4000 mmclear length l = 4000- 684/2 = 3658 mmThe effective length factor is assumed to be Ke=O.85, because the column can rotatefreely at the bottom end and the top end rotation is restrained by the moment resistingconnection to the beam.IN-PLANE RESISTANCE (MOMENT AND AXIAL FORCE COMBINED)The effective length of the column, Le, is:Le = Kele =3658x0.85 =3109mmThe slenderness ratio of the column, CC, is:Le3109 <50 QLd 532208thereforeKC=1.0Now, the compression resistance parallel to the grain may be calculated:Pr0•9)<20.4)<93l00’< 1.0= 1709300NThe bending resistance of the column is determined as follows:C8_(1.92l)_(1.92x3658><1752)_11.0CK = 20.4, as before1(cD4 iIii.o4KL=1—5—) =1—-) =0.97Mr = 0.9X 30.6X 8.26X 106X 0.97X 1.0 = 220.7X lO6NmmThe amplification factor due to the axial load is equal to:1 1 —P 283000 —1— 1— 211400.20X19(Le) 3109Mf = 159x1.O1 = 161 kNmNow, the interaction equation may be checked:283 + 161=090Pr Mr 1709.3 220.7OUT OF PLANE RESISTANCE (AXIAL FORCE ONLY’The effective length factor is Ke=l.O and:Le =Kelc =3658x1.0 =3658mm209The slenderness ratio of the column, CC, is:LQ 3658 <50 QJ.175 20.9OandCk_(F)_(11400x i.ox 1.0)0.52060CC>Cktherefore— 1 EQSKSEKT 1 11400X1.0X1.0F 2X20.90 20.4—0.64Now, the compression resistance parallel to the grain may be calculated:Pr0•9X204X93100X0.64 1093000NPr>Pf283000N O.K.b) second storey columnPf= 168kNM’= 78kNmAssume glulam section 130x418mm.A = 130x418 54 340 mm2S = 130x4182/6= 3.79 x106mm3I = 130x418/12 = 791 x106mm4theoretical length 1 = 4000 mmclear length l = 4000- 684/2- 608/2 = 3354 mm210The effective length factor is assumed to be Ke=O.8O, because the column can notrotate freely at any end. The amount of rotation of the top and bottom joints dependson the relative stiffness between all members adjacent to those joints.IN-PLANE RESISTANCE (MOMENT AND AXIAL FORCE COMBINED)The effective length of the column, Le, is:Le =Kelc =3354x0.80 =2683mmThe slenderness ratio of the column, CC, is:LQ 2683 <50 QJ.d - 418 —6.4thereforeK=1.ONow, the compression resistance parallel to the grain may be calculated:Pr0.9X20.4X54340X 1.0997700NThe bending resistance of the column is determined as follows:C8(1.92l)(1.92x3354x 12.6CK = 20.4, as beforeKL= 11(cB)41 1(12.60)4095Mr = 0.9X 30.6x 3.77X 106X 0.95X 1.0 = 98.6X lO6NmmThe amplification factor due to the axial load is equal to:P 168000 = 1.011— 1—2 2 791X106it it 11400 2(Le) 2683211Mf = 78x1.O1 = 79 kNmNow, the interaction equation may be checked:168 +=097O.KPr Mr 997.7 98.6OUT OF PLANE RESISTANCE (AXIAL FORCE ONLY’The effective length factor is Ke=1.O and:Le = K I = 3354x1.0 = 3354mmThe slenderness ratio of the column, CC, is:LQ 3354 <50 QJ.130=25.80andCC> Ck = 20.40therefore— 1 E05KsEKT 1 11400x1.0x1.0K——-—F 2X25.8 20.4=0,42Now, the compression resistance parallel to the grain may be calculated:= 0,9)< 20.4 x 54340x 0.42 = 419000NPr>Pf16S000N O.Kc)thirdstoreycolumnPf =73 kNM?=8OkNmAssume glulam section 130x418mm, the same as in b).212A = 54340 mm2S = 3.79 x106mm3I = 791 x106mm4theoretical length 1 = 4000 mmclear length l = 4000- 608/2- 570/2 = 3411 mmThe effective length factor is assumed to be Ke=O.75, because the column can notrotate freely at any end. The amount of rotation of the top and bottom joints dependson the relative stiffness between all members adjacent to those joints, which is greaterthanmb).IN-PLANE RESISTANCE (MOMENT AND AXIAL FORCE COMBINED)The effective length of the column, Le, is:Le = Ke l = 3411x0.75 = 2558mmThe slenderness ratio of the column, CC, is:LQ2558 <50 Q.jd 418 6.1thereforeKC=1.0Now, the compression resistance parallel to the grain may be calculated:Pr0.9)<20.4)<54340X 1.0997700NThe bending resistance of the column is determined as follows:213CK = 20.4, as before1IC8\4 1(12.7’\KL=l-3--) 131..204) =0.95Mr = 0.9x 30,6x3.77x 106X 0.95x 1.0 = 98.6x lO6NmmThe amplification factor due to the axial load is equal to:1 1 —P 73000 —1— 1— 6fl2El 2791x10(Le) 2558Mf = 80x1.01 =81 kNmNow, the interaction equation may be checked:+ 81=0.89O.K.Pr Mr 997.7 98.6OUT OF PLANE RESISTANCE (AXIAL FORCE ONLY’The effective length factor is Ke=1.0 and:Le Kelc3411x1.03411mmThe slenderness ratio of the column, C, is:L 3411 <50 O.K.130 =26.24andC> Ck = 20.40therefore— 1 EOSKSEKT 1 11400x1.Oxl.0F 2X26.24 20.4 =0.41214Now, the compression resistance parallel to the grain may be calculated:Pr0*9X204)<54340)<041 409000NPr>Pf73000N215