"Applied Science, Faculty of"@en . "Civil Engineering, Department of"@en . "DSpace"@en . "UBCV"@en . "Malczyk, Robert"@en . "2008-09-15T19:14:18Z"@en . "1993"@en . "Master of Applied Science - MASc"@en . "University of British Columbia"@en . "The aim of this research was to develop a reliable structural timber connection, which would break by failure of its steel components, thus taking advantage of the lower variability and higher strength of steel. The basic concept of the joint was to glue the reinforcing bars in pre-drilled holes and develop a bond strength stronger than tensile resistance of the bar. This report describes the testing and design of such connections. The joints consist of re-bars and steel plates. Weldable #10 and #20 re-bars were used. The steel plates were cut from flat bar. The re-bars are welded to the steel plates and inserted into the holes, and fixed to the wood by epoxy glue. The holes are drilled on an angle to the grain; 30 degree angle was found to be the most effective. The testing consisted of preliminary pull-out tests used to establish the embedment length needed to induce steel failures. Additional testing examined the behavior of glued-in rod joints used as a beam splice connection, a column foundation joint loaded with horizontal forces, and knee joints as used in a portal frame, loaded with positive and negative moment. Steel failures were observed consistently. A typical portal frame was designed in order to tie the research to practical applications. Internal forces from this analysis were then used as a guide to the magnitude of the forces that can be expected in real life situations.\r\nIn the last part of this report the necessary design steps for two configurations of glued-in rod joints are presented: 1.Column foundation or beam splice joints. 2.Knee joints. \r\nThis was performed in order to inform structural engineers how glued-in rod joints\r\nwork under load, and how they should be designed. Glued-in rod connections have several advantages over presently used systems. They can be manufactured in the glulam plant, using conventional tools, reducing the field work to merely bolting of the joints. The construction method becomes very similar to the erection methods of the steel structures. The elements are connected by bolts or welding. Glued-in rod connections have a great potential to become a safe connection method for the statically indetermined structures. The testing of knee joints showed the possibility of developing a moment resisting connection, which would make glulam more competitive against steel in the area of portal frames widely used for commercial buildings."@en . "https://circle.library.ubc.ca/rest/handle/2429/1972?expand=metadata"@en . "18817600 bytes"@en . "application/pdf"@en . "GLUED-IN RE-BAR CONNECTIONbyRobert MalczykM.Sc.(Civ.Eng), Warsaw Technical University, Warsaw, Poland,1990.A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF CIVIL ENGINEERINGWe accept this thesis as conformingquired standardTHE UNIVERSITY OF BRITISH COLUMBIAOctober 1993\u00C2\u00A9 Robert Malczyk, 1993In 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 4, ii/L --#1161,1:45re./it)6The University of British ColumbiaVancouver, CanadaDate DE-6 (2/88)11AbstractThe aim of this research was to develop a reliable structural timber connection,which would break by failure of its steel components, thus taking advantage of the lowervariability and higher strength of steel.The basic concept of the joint was to glue the reinforcing bars in pre-drilled holesand develop a bond strength stronger than tensile resistance of the bar.This report describes the testing and design of such connections. The joints consistof re-bars and steel plates. Weldable #10 and #20 re-bars were used. The steel plateswere cut from flat bar. The re-bars are welded to the steel plates and inserted into theholes, and fixed to the wood by epoxy glue. The holes are drilled on an angle to the grain;30 degree angle was found to be the most effective.The testing consisted of preliminary pull-out tests used to establish the embedmentlength needed to induce steel failures. Additional testing examined the behavior ofglued-in rod joints used as a beam splice connection, a column foundation joint loadedwith horizontal forces, and knee joints as used in a portal frame, loaded with positive andnegative moment. Steel failures were observed consistently.A typical portal frame was designed in order to tie the research to practical applica-tions. Internal forces from this analysis were then used as a guide to the magnitude of theforces that can be expected in real life situations.In the last part of this report the necessary design steps for two configurations ofglued-in rod joints are presented :1. Column foundation or beam splice joints.2. Knee joints.This was performed in order to inform structural engineers how glued-in rod joints111work under load, and how they should be designed.Glued-in rod connections have several advantages over presently used systems.They can be manufactured in the glulam plant, using conventional tools, reducing thefield work to merely bolting of the joints. The construction method becomes very similarto the erection methods of the steel structures. The elements are connected by bolts orwelding. Glued-in rod connections have a great potential to become a safe connectionmethod for the statically indetermined structures.The testing of knee joints showed the possibility of developing a moment resistingconnection, which would make glulam more competitive against steel in the area of por-tal frames widely used for commercial buildings.IntroductionLiterature ReviewPreliminary analysisMaterials used for testingPull-out testsBeam testRacking testsKnee joint testsDesign of glued-in rod joints91521475579107118119120122125ivTABLE OF CONTENTSAbstractTable of ContentsList of FiguresList of TablesList of PhotographsAcknowledgementChapter OneChapter TwoChapter ThreeChapter FourChapter FiveChapter SixChapter SevenChapter EightConclusionsFuture ResearchSummary of test dataBibliographyAppendix AVTable of FiguresFigure 1.1^Moment resisting column foundation joint, Riberholt, (1986)^3Figure 1.2^Moment resisting knee joint, Riberholt, (1986)^ 4Figure 1.3^Knee and column foundation joint, Turkowskij, (1991)^5Figure 1.4^Beam splice joint, Townsend, Buchanan, (1990) 6Figure 1.5^Knee joint using structural steel bracket, Townsend, Buchanan,^6(1990)Figure 2.1 Frame geometry.^ 9Figure 2.2^Forces in the glued-in re-bar connection.^ 11Figure 3.1^Compression perpendicular to grain specimen. 14Figure 3.2^Force displacement relation for #10 re-bar tested in tension.^15Figure 3.3^Glued-in rod connection^ 16Figure 4.1^Manufacturing steps for the preliminary connection.^21Figure 4.2^Prefabrication of glulam for the pre-welded joint. 23Figure 4.3^Force displacement relation for Specimen BL-10, typical for^31tension failure in re-bar.Figure 4.4^Force displacement relation for specimen SP-9.^ 37Figure 4.5^Force distribution along the re-bar for increasing loads.^37Figure 4.6^Stress distribution along the re-bar.^ 38Figure 4.7 Force displacement relation for Specimen CON-1^42Figure 4.8 Force displacement relation for Specimen CON-2 42Figure 5.1^Beam splice connection using glued-in rods.^ 45Figure 5.2^Beam test setup and instrumentation. 47viFigure 5.3Figure 5.4Figure 6.1Figure 6.2Figure 6.3Figure 6.4Figure 6.5Figure 6.6Figure 6.7Figure 6.8Figure 6.9Figure 6.10Figure 6.11Figure 6.12Figure 6.13Figure 6.14Figure 6.15Figure 6.16Figure 6.17Figure 6.18Figure 6.19Figure 6.20Mid span deflection and bending moment relation for all beam tests 47loadings.Gap movement and bending moment relation, beam test, loading^49#4.Instrumentation of the racking test setup.^ 53Specimen 2.^ 54Specimen 3. 54Loading force and time relation during racking tests.^56Bending moment and displacement relation for Specimen 1.^57Bending moment and displacement relation for Specimen 2.^58Bending moment and displacement relation for Specimen 3.^59Bending moment and displacement relation for Specimen 4.^60Forces in racking tests specimen.^ 62Vertical displacement of the joint area and bending moment for^66Specimen 3.Movements of the steel plates during loading.^67Horizontal movements of the joint area, Specimen 3.^67Displacement components of the racking tests specimen.^68Deformations due to hinge rotation.^ 69Deformations due to sliding shear. 70Calculated and measured displacements of racking tests specimens. 71Calculated displacement components, Specimen 1.^72Calculated displacement components, Specimen 2. 72Calculated displacement components, Specimen 3.^73Calculated displacement components, Specimen 4. 73viiFigure 7.1^Manufacture steps of a knee joint.^ 78Figure 7.2^Knee joint test setup.^ 79Figure 7.3^Loading history during knee joint tests.^ 82Figure 7.4 Displacement between pins and bending moment relation for Speci- 84men 1.Figure 7.5 Displacement between pins and bending moment relation for Speci- 86men 2.Figure 7.6 Displacement between pins and bending moment relation for Speci- 89men 3.Figure 7.7 Displacement between pins and bending moment relation for Speci- 89men 4.Figure 7.8 Displacement between pins and bending moment relation for Speci- 90men 5.Figure 7.9 Gap opening in Specimen 2.^ 94Figure 7.10 Gap opening in Specimen 3. 95Figure 7.11 Gap opening in Specimen 4.^ 96Figure 7.12 Gap opening in Specimen 5. 96Figure 7.13 Movements of steel plates in Specimen 2.^ 99Figure 7.14 Movements of steel plates in Specimen 4. 100Figure 7.15 Movements of steel plates in Specimen 5.^ 103Figure 8.1^Geometry of the splice and foundation joint. 105Figure 8.2^Forces in the steel plates.^ 106Figure 8.3^Compression parallel to grain in the splice or foundation joint.^108Figure 8.4^Forces on the exterior part of the knee joint.^ 108Figure 8.5^Forces on the interior part of the knee joint. 110viiiFigure 8.6^Forces in the steel plates.^ 113Figure 8.7^Compression parallel to grain in the knee joint.^114Figure 8.8^Forces in the inside stiffener.^ 116ixList of TablesTable 2.1 Internal forces obtained from analysis.^ 10Table 3.1 Compression perpendicular to grain test data. 14Table 3.2 Re-bar tension tests results.^ 15Table 4.1 Pull-out tests results, #10 bars, pull-out failures.^27Table 4.2 Pull-out tests results, #10 bars, tension failure in re-bars.^29Table 4.3 Pull-out tests results, #20 bars, compression perpendicular to grain^33failures.Table 4.4 Pull-out tests results, #20 bars, bar pull-out failures.^33Table 4.5 Pull-out tests results, #20 bars, tension failures in re-bars.^34Table 4.6 Pull-out tests results, pre-welded joints.^ 40Table 6.1 Design yield forces and bending moments in racking test specimens.^63Table 6.2 Tests results of ultimate forces and bending moments in racking test^63specimens.Table 6.3 Maximum bearing forces from tests and from analysis.^64Table 7.1 Ultimate bending moments at joint for knee joint specimens.^86Table 7.2 Gap opening results.^ 88Table of PhotographsPhotograph 4.1 The hole drilling technique.^ 19Photograph 4.2 Preliminary joint.^ 22Photograph 4.3 Pull-out test setup 25Photograph 4.4 Instrumented #10 re-bar.^ 37Photograph 4.5 Pre-welded joints after failure. 40Photograph 5.1 Instrumentation of the beam test specimen.^46Photograph 5.2 Re-bar tension failure in the beam specimen.^48Photograph 6.1 Racking test setup.^ 52Photograph 6.2 Specimen 4 during manufacturing.^55Photograph 6.3 Specimen 1 after failure.^ 61Photograph 6.4 Specimen 3 after failure. 61Photograph 6.5 Location of LVDT's during racking tests.^65Photograph 7.1 Knee joint test setup.^ 71Photograph 7.2 Instrumentation of the knee joint test specimen.^81Photograph 7.3 Specimen 1 after failure.^ 85Photograph 7.4 Specimen 2 after failure, phase 1.^87Photograph 7.5 Specimen 2 after failure, phase 2. 87Photograph 7.6 Specimen 3 after failure.^ 90Photograph 7.7 Specimen 4 after failure. 91Photograph 7.8 Specimen 5 after failure.^ 91Photograph 7.9 Instrumentation of the specimen to measure the joint 93rotation.xiAcknowledgementThe author is very grateful to his supervisor Professor Borg Madsen for his guidance,valuable suggestions, and encouragement throughout this research.The author express his gratitude to Professor R.O. Foschi for theoretical input andfor reviewing the manuscript, and Professor R. Sexsmith for reviewing the manuscript.Appreciation is extended to Mr. H. Kulmann, Mr. P. Symons, and Mr. R. Wiktor forthe helpful participation in preparing and maintaining instrumentation used in the exper-imental part of this research, and for necessary assistance during the various investiga-tions.Research grant from British Columbia Science Council is gratefully acknowledged.Extensive testing described in this report was possible with the help of the followingcompanies: Industrial Formulators of Canada, Structurlam, Surrey Laminaters, and TrussJoist MacMillanThe author wish to thank his family and friends for their support throughout hisgraduate career.CHAPTER 1INTRODUCTION AND LITERATURE REVIEW1.1 INTRODUCTIONIn today's construction market timber structures are less developed than steel andconcrete structures. There are several reasons for this. For years timber buildings weredesigned as statically determined structures, because there were no tools to calculateindeterminate structures accurately. Because of wide use of computers, this is no longervalid.The real reason however for timber structures being less developed than structuresmade from steel and concrete is a lack of reliable connections.In indeterminate structures the efficient use of material depends on the availability ofrigid connections. To make timber more competitive new kinds of rigid joints are underdevelopment. There are several approaches to this problem and many research programsare presently under way dealing with rigid joints.One technique is to make a connection where the failure is forced to occur in the steelthus utilizing its small variability. Properly designed connections will fail by tension failurein the steel dowel, without affecting the wood.The idea originated in Scandinavia where steel rods, were inserted in holes drilledparallel to grain and fixed to the wood by adhesives. Research modifying this method ispresented in the following literature review.21.2 LITERATURE REVIEW1.2.1 GENERAL COMMENTSA comprehensive overview of research related to the topic of this thesis is presentedin chronological order. Studies on glued-in rod connections are presently performed (October, 1993 ) in Canada, Finland, Russia, Australia and New Zealand.Two different approaches are represented. Russian and Finnish researchers are usingreinforcing bars glued in to glulam at an angle. In New Zealand and in Australia rigidconnections using rods glued parallel to grain are being tested.The first research on glued-in steel rods in laminated timber was performed in Swedenaround 1965.1.2.2 RIBERHOLT.The first major work on dowels glued in glulam was performed by Riberholt (1980,1986, 1988.) In Denmark. Riberholt used threaded rods, glued parallel to grain.His report covers testing with pull-out tests where axial loads were applied for bothshort and long term period. In addition rods subjected to lateral loads were mentioned. Thisinformation is then applied to spliced beams tested under both wet and dry service conditions.Column foundation joints were also developed, which could transmit moments and shearforces in addition to axial loads. Finally the experience was applied to connections wheremoment resistant joints were required.In order to achieve this it was necessary to establish the minimum embedment lengthsthat would cause yielding in the steel rather then bonding failure in the glue when subjectedto an axial tension load.The effect of the rod diameter was also investigated as well as the influence of the holesize. The importance of the density of the glulam was also investigated.The physical dimensions such as end distances were established such that the failureiroiroNmi.\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2Tampedm or tar3would take place in the wood. Several glue types were investigated, and while some dif-ferences were observed both Polyurethane and Araldite glue were found to be acceptable.From his tests with the effect of moisture content it was found that wet specimenswere generally 15 percent weaker than the dry ones.Riberholt also noted that long term pull-out tests specimens exposed to outdoorconditions have half the strength of the matching specimens tested under dry conditions.Summary : 1. Equations for maximum pull out force depending on hole diameter, embedment length,and specific density were determined2. An edge distance of 1.5 times dowel diameter was found to be sufficient to secure fullstrength of glued in dowel.Figure 1.1 shows the moment resistant column foundation joint that was investigated byRiberholt.Joint^ConcreteFigure 1.1 Moment resisting column foundation joint. Riberholt, (1986).In 1986 Riberholt started testing of moment resisting knee joint connection. Figure1.2 presents moment resistant knee joint tested in this program. The capacity of the jointreached 80% of the glulam cross-section capacity.4Figure 1.2 Moment resisting knee joint. Riberholt, (1986).1.2.3 TURKOWSKIJ.Connections using reinforcing bars glued into glulam on an angle to grain have beenintroduced in Russia in 1975, Turkowskij, (1991). Re-bars were glued into pre-drilled holes4 mm larger than their diameter. Epoxy was used as the adhesive. The re-bars were thenwelded to the slotted steel plate, connecting to the next element.Several tests were performed, checking the behavior of these joints in tension, com-pression, bending, and shear.Several buildings have been completed using this connection method. The list includessport centre, where glued-in re-bars were used as knee and column foundation joints. SeeFigure 1.3 This method was also used for reinforcing existing glulam structures.For architectural reasons Canadians would find this connection unacceptable becauseof the appearance and the welding up against glulam. However the concept of glueing dowelson an angle to grain was used in the connection presented in this thesis.5Figure 1.3 Knee and column-foundation joint, Turkowskij (1991)1.2.4 TOWNSEND AND BUCHANAN.An extensive program of glued in dowels testing was carried out by Townsend andBuchanan, (1990), in New Zealand. It should be noted that they used high strength rein-forcing bars glued parallel to grain.Their testing program duplicates the experiments conducted by Riberholt but usingradiata pine wood. Unlike Riberholt, Townsend and Buchanan used reinforcing bars as wellas threaded rods.During the next series of tests, Townsend examined the behaviour of several beamsplice joints using reinforcing bars. See Figure 1.4. The capacity of the connection washigher than the capacity of the glulam cross-section. The beam splice used in this programis shown on Figure 1.46Figure 1.4 Beam splice joint, Townsend, Buchanan, (1990)Several types of knee joints were tested by Buchanan and Townsend. The most suc-cessful one, using structural steel bracket is already patented. See Figure 1.5Figure 1.5 Knee joint using structural steel bracket, Townsend, Buchanan (1990)Conclusions :1. Equations for the maximum pull-out force were established in the function of : bar7diameter, embedment length, ratio of hole diameter to bar diameter, ratio of edge distanceand bar diameter. The equations developed by Riberholt were confirmed during Townsendand Buchanan tests.2. In dry wood specimens, adhesive Araldite K-2005, (similar to polyurethane used byRiberholt), behaved similar to Araldite K-80 (more rigid epoxy).3. The resistance of the specimens with K-80 epoxy, tested after wet and dry cycles was 20%weaker than the ones with K-2005 adhesive.4. In over 300 tests there was no glue failure of the adhesive bond between the epoxy andthe wood surface. Pull out failures, when they did occur were caused by a shear failure inthe surrounding wood, away from glue line.1.2.5 CONCLUSIONS1. The number of research programs presently under way using glued-in dowels testifies tothe need for better connection methods.2. Reinforcing bars can be used for the glued-in rod connections. A cost comparison ofre-bars and threaded rods favours the re-bars.3. Inclined bars engage more of the glulam cross-section, than parallel glued ones, whichbehave more like skin connectors. The re-bars also increase the shear capacity of the glulamcross-section.4. Another advantage of the inclined bars is that in contrast to parallel bars they do notcause air pockets during the gluing operation, so they are easier to make and more reliable.5. Holes 4 mm larger, than the re-bar diameter were found acceptable by mentionedresearchers.8CHAPTER 2PRELIMINARY ANALYSIS2.1 INTRODUCTIONThe aim of this chapter is to establish what internal forces can be expected in a heavilyloaded portal frame. The span of the frame was chosen to be 12 m, spacing - 6 m. A 175 mmwide, 20f-EX glulam sections were used. Roof slope was selected to be 8 degrees to checka demanding geometry resulting in high bending moments at the knee joints.Three frame types with different hinge locations were analyzed:1. Fixed frame.2. Two hinged frame - foundation hinges.3. Three hinged frame.The frame with foundation hinges, rigid knee joints and rigid apex joint was found tobe the most effective. This frame was subsequently designed according to CAN/CSA086.1-M89. Several load cases were checked including snow, wind, earthquake, as well astheir combinations. Bending moment and shear force envelopes were established. Theknowledge of internal forces found in this chapter was used as background for the exper-imental part of this thesis. Four full size tests of the knee joints, from the designed framewere tested, as final experiments described in this report. See Chapter 7.2.2 FRAME GEOMETRYMany commercial buildings need portal frames with clear span of 10 to 15 m. Theshape of the frame was selected to be a demanding one, where high values of internal forcesare present. The spacing was chosen to be 6 m for the snow loads in Vancouver. For otherareas, where the snow load is higher, the spacing should be proportionally decreased.The final geometry, see Figure 2.1, was developed in conjunction with the glulammanufacturers. In their opinion there is a great need for development of this type of joints9for these structural frames.Figure 2.1 Frame geometry.2.3 LOADINGVancouver B.C. was chosen for the location of the building. The climatic data wasestablished according to Canadian Building Code.1. Snow load : Ss = 2.5 kPa, Sr = 0.3 kPa.2. Wind load : q = 0.55 kPa3. Earthquake load : v = 02The frame was loaded with four load combinations.- dead load + full snow load- dead load + wind load- dead load + earthquake load- dead load + wind load + 1/2 snow loadThe limit states load factors were used according to CAN/CSA 0862.4 SUMMARY OF THE ANALYSISA two hinge glulam frame of the geometry shown on Figure 2.1 was designed for anindustrial building in Vancouver. Appendix A presents the loads, and the design of deckingand purlins, as well as design of the frame.10The following table presents the maximum internal forces obtained during analysis atknee joint for all load combinations.Load combination Bending momentkNmShear forcekNAxial forcekN1. Full snow 219.5 111.4 147.92. Wind 51.6 26.5 18.53. Earthquake -22.6 3.3 22.94. Wind + half snow 76.6 24.4 25.3Table 2.4 Internal forces obtained during analysis.The maximum bending moment occurred during full snow load - positive momentresulting in tension on outside of the frame, and during earthquake load - negative bendingmoment resulting in tension on the inside part of the knee joint area. Negative momentfrom earthquake load was around 10% of the positive moment from full snowload. Maximumshear and axial forces were found during full snow load.2.5 PRELIMINARY CONNECTION DESIGNThe analysis provided the external forces that can be expected in the heavily loadedportal frame. Maximum bending moment under full snow load was 219.5 kNm. This resultsin a tension force in re-bars Fr = 391.4 kN. The following figure shows the forces in glued-inrod connection, and the free body diagram, for a very simple concept of joint behaviour,used as a first approximation.11Figure 2.2 Forces in the glued-in re-bar connection.F u \u00E2\u0080\u0094 M 219.4 \u00E2\u0080\u0094 339.6 k N-^1^0.646F R = F M x cos(a) = 339.6 x cos(30) = 295.6kNIn the calculation of force Fr the following assumption was made :Shear force V2 is assumed small in comparison to Fm and it was neglected.Assuming standard variation - 15%, increase for the fifth percentile equals to 1.33F R = 1.33x 295.6 = 392.4 k NTherefore the embedment length of the re-bars during the pull-out tests has to beenough to ensure the tension failure of the re-bars, at the pull out force above 400 kN .Assuming that the re-bars would yield at stress of 400 MPa, and #20 re-bars will be used inthe joint the required re-bar area can be calculated.A r e cl, \u00E2\u0080\u0094 FR\u00E2\u0080\u0094 400x 103\u00E2\u0080\u0094 1000rnm2F,,^400A r:g 1000n reb =3 .3 .A \u00E2\u0080\u009Eb 300Four #20 re -bars are needed at the knee joint connection. The re -bars will be stressed to1283% of the yield stress.2.6 SUMMARY1. This chapter has established that a suitable joint being able to resist the bending momentof 219.4 kNm requires four #20 re-bars.13CHAPTER 3MATERIALS USED FOR TESTING3.1 INTRODUCTIONThis chapter describes the types of materials used in this program for making speci-mens. The test types were :- pull-out tests (embedment length)- beam test (splice joint)- racking tests (column foundation joint)- knee joint tests (frame structures)Materials readily available in the building supply trade were chosen.3.2 GLULAMThree sources of glulam were used for the tests described in this report. The firstsource was some 30 year old beams recovered from UBC bookstore, that was demolished.This material was used for getting preliminary information from pull out tests.The next source was Surrey Laminators Ltd. in Surrey B.C. The glulam grade chosenwas 24f-EX, Douglas Fir. This material was used for the beam test, racking tests and onepreliminary knee joint test.The final part was prepared by Structurlam Ltd. in Penticton B.C., and used for theremaining knee joint tests. The material grade was 24f-EX and glulam was made of DouglasFir lumber.The inclination of the hole in the glued-in re-bar joint results in compression per-pendicular to grain so this material property was tested. Compression perpendicular to grainwas tested for four specimens. The compressed area was 130 x 130 mm See Figure 3.1.The height of the specimen was 376 mm. Results are summarized in Table 3.1.14Specimen # Maximum stresscompression perp. to grainMPa1 5.22 5.73 6.74 6.3Average 6.0Table 3.1 Compression perpendicular to grain test data.This value is similar to 5.9 MPa stated in CAN/CSA-089.Figure 3.1 Compression perpendicular to grain specimen.3.3 PARALLAMParallam 2.0E-SP PSL was provided by Truss joint MacMillan Ltd. Parallam was onlyused for a prewelded connection and tested in pull-out test. Because of a problem with glueleaks through the pores inside the cross-section, the use of this material was discontinuedat this time. Thicker glue might be used for parallam to avoid the leaks. Because the dataobtained from tests with parallam specimens was not reliable the results are not presentedin this report.zw0cc0Li.2^3^4DISPLACEMENT (mm)153.4 STEEL RODSRegular strength, deformed weldable reinforcing bars were used in the tests. Two bardiameters were tested : 10 mm and 20 mm. Reinforcing bars were chosen, because of theirlow price compared to threaded rods used by some other researchers for similar applications.Re-bars can be easily bent and cut to the desired length. Tension tests were performed forboth re-bar diameters. Three specimens of each diameter were tested. Table 3.2 summarizesthe results showing the average values. Figure 3.2 shows the typical load deformation relationfor #10 re-bar.Bar diameter - mm 10 20Yield force - kN 45.2 144.2Ultimate force - kN 67.5 202.2Yield stress - MPa 452.1 481.5Ultimate stress - MPa 675.0 674.1Table 3.2 Re-bar tension tests results.Figure 3.2 Force displacement relation for #10 re-bar tested in tension.INSIDE PLAT \u00E2\u0080\u0094BARSSTRAP\u00E2\u0080\u00A2\u00E2\u0080\u00A2 \u00E2\u0080\u00A2\u00E2\u0080\u00A2 \u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2 \u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2163.5 STEEL PLATESAll steel plates used for glued-in rod joints were type A-36, cut from flat bars. Specifiedminimum tensile strength was Fu = 400 MPa. The following flat bar dimensions were usedfor the different tests : 1/4\" x 2\" (6.4 x 50.8 mm ) , 1/2\" x 3\" ( 12.7 x 76.2 mm ) , 1/2\" x 5\" ( 12.7x 127 mm ) , 1/2\" x 6\" ( 12.7 x 152.4 mm )3.6 BOLTS AND STUDSStuds (See Figure 3.3), cut from threaded rods, were used in specimens for :1. Pull-out tests.2. Beam test.3. Racking tests.The diameter of the rod was 5/8\" ( 15.9 mm) The rod was made of high strength steel( ultimate tensile stress - 830 Mpa), with thread type NF ( National Fine ).The studs were replaced by the bolts in knee joints, because of the geometry of a kneejoint. Connectors between the glulam steel plate and the outside strap had to be insertedfrom outside through the strap. The strap is a steel plate, which connects two glulamelements. The glulam steel plates are the internal parts of the connection fixed to glulamin the plant. The bolts were of No 7 grade with NF thread.Figure 3.3 Glued-in re-bar connection used as a beam splice joint.173.7 EPDXYOnly one type of epoxy - IFC-SP, was used for all the connections tested in this researchprogram. This adhesive was manufactured by Industrial Formulators of Canada Ltd., Bur-naby B.C. Raisin and hardener were shipped in separate bottles and mixed in proportion42 to 100 by weight before application. The pot life was approximately 2 hours dependingon the outside temperature. After gluing all specimens were left for 7 days before testing.The glue was poured in to the holes before the re-bars were inserted. Because of theinclination of holes, there was no problem with air pockets inside the hole and no plugs wereneeded. Extra glue from the hole was used to ensure the proper contact between the glulamsteel plate and the glulam surface, because of the presence of the compression perpendicularto grain If needed additional glue was poured between the glulam steel plate and the glulamsurface after the re-bars were inserted, to assure a tight fit.The inclined holes eliminated shaking of the specimen to ensure the proper gluedistribution, and plugs at the hole ends, needed when the holes are drilled parallel to grain.18CHAPTER 4PULL-OUT TESTS4.1 TESTS OBJECTIVES.The first objective of the pull-out tests was to establish the embedment lengths of there-bars required to ensure failure in the steel. The holes for the re-bars were drilled atdifferent angles to grain. The first series of pull-out tests described in this thesis was per-formed to repeat experiments conducted by the mentioned researchers but using Canadianmaterials.The concept of gluing bars at the angle was first used in Russia, see Turkowskij, 1990.The appearance of the Russian connection however is unacceptable from a Canadianarchitectural point of view. Further more, the exposed weld, and the fact that the weldingis performed close to the glulam surface, burning the wood surface is undesirable.The second objective was therefore to eliminate the welding close to glulam, possiblydamaging the epoxy glue, and improve the appearance of the connection. This was achievedby developing the \"pre-welded\" joint, in which the welding is performed before the re-barsare glued, away from glulam, and on the underside of the steel plate, making the weld invisible.The third objective of this series of tests was to develop a reliable method of man-ufacturing glued-in rod joints.It was decided to exclude the glue type and the hole size from the test objectives.4.2 EVOLUTION OF PULL-OUT TEST SPECIMENThis section describes the evolution of the pull-out test specimen.1. Preliminary joint.The preliminary joint was similar to one used by Turkowskij, (1990). The re-bars wereglued in and then welded to the slotted steel plate. The preliminary joint is presented onPhotograph 4.1. Note that the appearance is not attractive192. Pre-welded connection. The re-bars are here pre-welded to the steel plate before glueing, as described earlier.Daps in the glulam are cut so the flat part of glulam and the steel plate are touching eachother. Additional glue is added for better contact, between the steel plate and glulam. Thisjoint has several advantages compared to the preliminary connection.- increase of the bearing area.- avoid heating of epoxy glue while welding, which would decrease the bond stress.- improved appearance since weld is hidden and the wood is not burned by welding.The only disadvantage of the pre-welded joint is more labor intensive since greateraccuracy is needed for proper fitting.Photograph 4.1. Preliminary joint.4.3 PARAMETERS.During pull-out tests the effect of three parameters on the load displacement behaviorwas examined :201 Embedment length of the reinforcing bar. The aim was to find the embedment lengths of the re-bars, which will ensure a strongerbond between re-bar and glulam than the tensile resistance of the rod. This will result intension failure of the re-bars, for all cases when the embedment length is longer than theboundary value.2. Inclination of the hole.Flatter angles result in smaller residual stresses in the bent section of the re-bar, butrods glued in at small angle to grain engage less of the glulam cross-section. This is similarto the behavior of the skin connectors. Larger angles, (up to 45 degrees), increase shearcapacity of the cross-section and get hold of more glulam but result in higher residual stressesat the bent section of the re-bar.3. Diameter of the reinforcing bar. The first series of tests was done with #10 re-bars. The second series used #20 bars.The embedment lengths to ensure the re-bar tension failure were found for both diameters.4.4 FABRICATION OF SPECIMENSFABRICATION STEPS OF THE PRELIMINARY JOINT.This section will describe the manufacturing steps of the preliminary joint, as done byTurkowskij, 1990.1. Drilling of holes in glulam. The hole diameter was 4 mm larger than diameter of re-bars. This was recommendedby previous research - see Turkowsij, 1990, and Townsend, Buchanan, 1990. The drillingtechnique is presented on Photograph 4.2. The steel drilling guide was used to ensure theproper location and angle of the hole.2. Cutting of re-bars to the desired length. Length of re-bars was equal to the inside length of the hole plus 150 mm which wasbent down.3. Glueing of re-bars into the pre-drilled holes. Epoxy was poured into the holes and re-bars were inserted. Due to the inclination of21the hole there were no problems with the air pockets.4. Bending of re-bars Re-bars were bent, without heating, to make the outside part parallel to the glulamsurface.5. Welding of the re-bars to the slotted steel plate. The strap, was inserted on top of the re-bar and welded around the slot. The strapwas a steel plate with slotted hole of the same length as a bent down re-bar. While man-ufacturing the specimens with #20 re-bars fabrication process was improved :A. The re-bars were bent before glueing.B. The slotted plate was replaced by the solid plate.C. The re-bars were welded to the lower surface of the strap.Figure 4.1 presents the manufacturing steps for the preliminary connection.Figure 4.1 Manufacturing steps for the preliminary connection.22Photograph 4.2. The hole drilling technique.FABRICATION STEPS FOR THE \"PRE -WELDED\" JOINT.The following section will describe the manufacturing steps for the \"pre-welded\" joint.1. Cutting out groves in glulamlThe pre-welded connection was designed to increase the bearing area available forcompression perpendicular to grain, and to improve the appearance of the joint. The ideawas to hide the connection inside glulam. The glulam steel plate was flush with the glulamsurface. To achieve this daps for the bent re-bar sections and the plate had to be prepared.The shape of the openings is presented on Figure 4.2.23Figure 4.2 Prefabrication of the glulam for the pre-welded joint.2. Drilling of holes in glulam. The holes in glulam were drilled using the same technique as for the preliminary joints,after the grooves were cut.3. Cutting and bending of re-bars. The re-bars were cut and bent before gluing.4.Tackwelding of the inserted re-bars to the steel plates. The re-bars were inserted in the holes, steel plates placed on top of them, and thenthey were tackwelded. The purpose of this was to ensure the proper positioning of the re-barin relation to the steel plate.5. Welding of re-bar to the steel plate. The re-bars and the steel plate were then removed from the hole and fully welded tothe steel plate, away from the specimen, forming an assembly. The weld was positioned onthe inside face of the plate.6. Gluing of the re-bars in the holes. The epoxy was mixed and poured in to the holes, and the assembly was inserted. Thesteel plate which was welded to the re-bar was resting against the glulam surface. Additionalglue was poured between the steel plate and glulam surface to ensure the contact, andincrease the bearing area. All steel surfaces, plates and re-bars, had been sand blasted beforegluing. Failure between steel and the glue was not observed in any of the tests.244.5 TEST PROGRAMPull-out tests were performed for the following test configurations :1. Tests with #10 re-bars.The effect of hole inclination, and re-bar embedment length were examined. Thefailure modes of specimens using #10 bars were observed.2. Tests with #20 re-bars. The effect of hole inclination, and re-bar embedment length were examined. Thefailure modes of specimens using #20 bars were observed.3. Instrumented bar test. The stress distribution along the embedded length of re-bar was examined.4. Pre-welded connection. The effect on the bearing stresses of the steel plate being pre-welded to re-bars andfailure modes was examined. A manufacturing technique for the pre-welded joint wasdeveloped.4.6 TEST SETUP AND TESTING METHOD.All the tests described in this report were performed at the structures lab of theDepartment of Civil Engineering of U.B.C. In Vancouver. The test setup (See Photograph4.3.), Used for the pull-out tests consists of :1. Hydraulic jack and its support.2. Horizontal support.3. Vertical support.The hydraulic jack was fixed by bolts to the steel frame which in turn was bolted to theconcrete floor of the lab. The test specimen was placed against the horizontal support, whichwas also bolted to the floor. The vertical support was then placed on the two bolts stickingout from the floor and tightened against the specimen, to prevent the rotation of thespecimen. The load cell was connected to the end of the hydraulic jack. A steel strapconnected the specimen and the load cell, via pin. On the other end, the strap was fixed tothe specimen via studs.25The force from the hydraulic cylinder, travelled from hydraulic jack through the steelplates to the re-bars. It was then transferred to the glulam by the glue bond between there-bars and the wood. The load was increased at the constant rate, using a manual gage,until the failure of the specimen occurred.Two sets of readings were recorded by the data acquisition system. One channel ofdata contained reading of the force from the load cell. The other was readings of the LVDT,which measured the relative displacement between the specimen and the steel strap. Thebody of the LVDT was fixed to the upper surface of glulam, while the moving part was restingon a magnetic base, which was sitting on the top of the strap.Photograph 4.3. Pull-out tests setup.4.7 TESTS WITH THE PRELIMINARY CONNECTION.Three types of experiments were preformed for the preliminary connection usingsingle bars :261. Tests with #10 re-bars.2. Tests with #20 re-bars.3. Test with the instrumented bar.4.7.1 TESTS WITH #10 RE -BARSThe effects of two parameters were checked in this series of tests. The first one wasthe embedment length of the re-bar. The embedment length was measured along the re-barfrom the glulam surface to the end of the rod. The re-bars were glued in on the entire lengthof the hole. Three embedment lengths were tested : 130, 300, and 380 mm.The second parameter was inclination of hole relative to the direction of the grain.Three angles examined were : 15, 30, and 45 degrees.The analysis of the tests results is presented according to the two failure modes, whichare defined below.1. Pull-out of the re-bar. The bond between the glue and glulam fails and the bar is pulled out of the specimen.Failure between the steel and epoxy was not observed.2. Tension failure in re-bar. The failure takes place in the steel without affecting the wood. The re-bar yields andbreaks in tension at the glulam surface, where it may have been weakened by the bendingand by the welding.PULL-OUT FAILURES. Only specimens with the smallest embedment length of 130 mm failed by pull-out.Tests results are presented in Tables 4.1-A, 4.1-B, and 4.1-C, for 15, 30 and 45 deg respec-tively.The following symbols are used in the Tables :SPEC # - specimen numberLd - embedment lengthALPHA - hole inclination to grain.27Fmax - maximum force applied during testDmax - displacement of the steel plate relative to glulam at maximum force.STRESS - stress in re-bar at failure.SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(MPa)BL-8 130 15 59.7 2.5 573BL-23 130 15 66.7 2.5 640BL-24 130 15 67.8 2.4 651AVERAGE 64.7 2.5 621Table 4.1-A. Pull out tests results. #10 re-bars. ALPHA = 15 deg. Pull-out failures.SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(MPa)BL-1 130 30 69.7 9.1 606BL-2 130 30 73.6 4.5 640BL-13 130 30 69.0 4.5 600AVERAGE 70.8 6.0 616Table 4.1-B. Pull out tests results. #10 re-bars. ALPHA = 30 deg. Pull-out failures.SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(MPa)BL-5 130 45 68.3 11.1 483BL-6 130 45 68.7 11.0 486BL-22 130 45 733 9.7 518AVERAGE 70.1 10.6 496Table 4.1-C. Pull out tests results. #10 re-bars. ALPHA = 45 deg. Pull-out failures.SUMMARY1.The embedment length of 130 mm results in pull-out failures for all tested hole inclinations.2. Essentially the same maximum forces were obtained for 30 and 45 degrees, while thevalues for 15 degrees were 7% lower.3. The displacement of the plate at failure increases with increase of the hole inclination.Smallest displacements were observed for 15 deg angle and the largest for 45 deg.4. The stresses at failure in the re-bars were above the yield stress - 400 MPa, for all specimens.28TENSION FAILURE IN RE-BARS.All specimens with an embedment length of 300 and 380 mm failed as tension failurein the re-bar. The bars broke in tension just outside the hole, close to the glulam surface.The tests results are presented in Tables 4.2-A, 4.2-B, and 4.2-C.The following symbols are used in the Tables :SPEC # - specimen numberLd - embedment lengthALPHA - hole inclination to grain.Fmax - maximum force obtained during testDmax - displacement of the steel plate relative to glulam at maximum force.STRESS - maximum stress in the re-bar29SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(Mpa)BL-9 300 15 70.6 7.4 678BL-17 300 15 84.1 6.4 807BL-25 300 15 74.9 8.6 719Average 76.5 7.5 734BL-16 380 15 84.1 6.7 807BL-19 380 15 78.3 7.7 752BL-26 380 15 76.2 7.4 731Average 79.5 7.3 768I AVERAGE I FOR I ALPHA=15 I 78.0 I 7.4 I 748 ITable 4.2-A. Pull out tests results. #10 re-bars. ALPHA = 15 deg. Re-bar tension failures.SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(Mpa)BL-3 300 30 86.2 6.3 750BL-4 300 30 76.4 8.3 665BL-12 300 30 91.1 14.6 793Average 84.5 9.3 735BL-14 380 30 82.2 9.0 715BL-11 380 30 79.7 14.8 693BL-10 380 30 72.1 13.1 627Average 78.0 12.3 679I AVERAGE I FOR I ALPHA=30 I 81.3^11.0 I 747 ITable 4.2-B. Pull out tests results. #10 re-bars. ALPHA = 30 deg. Re-bar tension failures.30SPEC # Ld(mm)ALPHA(deg)Fmax(IN)Dmax(mm)STRESS(Mpa)BL-15 300 45 79.3 6.2 561BL-21 300 45 77.3 9.7 547BL-27 300 45 78.9 7.3 558Average 78.5 7.7 555BL-7 380 45 78.1 123 552BL-18 380 45 75.9 16.3 537BL-20 380 45 76.2 14.8 539Average 763 14.5 542I AVERAGE I FOR I ALPHA=45 I 77.6 I 11.1 I 548 Table 4.2-C. Pull out tests results. #10 re-bars. ALPHA = 45 deg. Re-bar tension failures.SUMMARY1. Embedment lengths of 300 and 380 mm produce bond between re-bar and glulam strongerthan tensile resistance of the rod. This causes re-bar tension failure.2. The highest values of the force were obtained for angle alpha=30 degrees. Forces forother angles were close to each other and 5% less than the average of the specimens with30 degree angle.3. Displacement of the plate was the same for 30 and 45 degree angles. Displacement atfailure for specimens with 15 degree angle was 30% smaller.31FORCE DISPLACEMENT RELATION FOR RE-BAR TENSION FAILURES The following graph presents the relation between the force and displacement forspecimen BL-10, which is typical for failures of re-bars. The embedment length was 380mm. The angle of hole inclination was 30 degrees.60^0' t^0 2.6^6^715^10^12.6DISPLACEMENT (mm)Figure 4.4 Force displacement relation for specimen BL-10, typical for re-bar tension fail-ures.The relation between force and displacement can be divided into two areas. The firstone for forces up to 60 kN which is a linear relationship. For the forces higher than 60 kN,(stress in re-bar - 520 MPa), the re-bar starts to yield, and displacement is growing almostwithout the increase of force. The re-bar breaks after 15 mm elongation which is desirablefor ductile connections. However the maximum force was 71 kN at 12.5 mm, causing a stressof 617.SUMMARY AFTER TESTING #10 RE-BARS.1. The tension failure occurs for the re-bars having an embedment length more than 300mm.2. 30 degree inclination results in the slightly higher force at failure.3. Force displacement diagrams show a combination of linear behavior and a long ductilityrange.16^17.6^20324.7.2 TESTS WITH #20 RE-BARSThe main purpose of repeating the pull-out tests using #20 bars was to establish theembedment length needed to obtain tension failure in the re-bars for the larger bar diameter.To establish the embedment length needed to obtain the tension failure of re-bar more exactthan for the #10 bars, five embedment lengths were tested :130, 200, 260, 300, and 380 mm.As for previous tests three different inclinations were tested : 15, 30, and 45 degrees.The analysis of the tests results is also presented according to the failure modes. Threefailure modes were observed.1. Compression perpendicular to grain failure. The portion of re-bar close to the glulam surface crushes in the wood. The reason isthat a bearing area is too small.2. Pull-out of the re-bar. The bond between the glue and glulam fails and the bar is pulled out of the specimen.3. Tension failure in re-bar. The failure takes place in the steel without affecting the wood. The re-bar yields andbreaks in tension close to the glulam surface, weakened by the bending and by the weldingCOMPRESSION PERPENDICULAR TO GRAIN FAILURE. Compression perpendicular to grain failure occurred in the first three specimens,because not enough attention had been given to this failure mode. For all the remainingtests with #20 re-bars, the bearing area was increased, by inserting the steel plates betweenthe strap and the glulam surface.The following table presents tests results of the specimens, which failed by compressionperpendicular to grain.33SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(MPa)SP-1 380 45 222.0 6.4 532SP-2 200 30 114.9 13.7 333SP-5 260 45 134.7 5.7 317Table 4.3 Pull out tests results. #20 re-bars. ALPHA = 30, 45 deg. Compression perpen-dicular to grain failures.SUMMARY1.A different type of failure was observed. For the higher forces compression perpendicularto grain initiates the first failure, when insufficient bearing area is provided.2. This shows the importance of checking compression perpendicular to grain forces.PULL-OUT FAILURE OFF THE RE -BAR. Three specimens failed by pull-out failure of the re-bar. The following table presentsthe tests results.SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(MPa)SP-7 200 45 117.9 23.8 278SP-10 260 15 185.3 15.6 593SP-12 130 45 75.4 3.6 177Table 4.4 Pull out tests results. #20 re-bars. ALPHA = 15, 45 deg. Bar pull-out failures.SUMMARY1. Embedment length needed to cause re-bar tension failure is bigger than 260 mm for 45and 15 degree angles of hole inclination.2. The force required to pull out the re-bar, increases with the increase of embedment length.TENSION FAILURE OF RE-BARTension failure of the re-bar occurred for #20 re-bars only after additional steel plateswere inserted between the strap and the glulam surface, this increased the bearing area. Sixspecimens failed by re-bar tension failure. Tests results are presented in Tables 4.3-A, and4.3-B.34SPEC # IA(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(Mpa)SP-9 260 30 179.8 19.0 521SP-11 260 30 173.4 17.8 503SP-4 300 30 164.8 16.5 478SP-8 380 30 164.4 22.2 477Average 170.6 18.9 495Table 4.5-A. Pull out tests results. #20 re-bars. ALPHA = 30 deg. Bar tension failures.SPEC # IA(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(Mpa)SP-3 380 15 190.9 11.0 608SP-6 300 45 137.1 20.9 324Table 4.5-B. Pull out tests results. #20 re-bars. ALPHA = 15,45 deg. Bar tension failures.SUMMARY1. The tests showed with reasonable degree of confidence that the embedment length tocreate the tension failure in the re-bar was 260 mm for 30 degree angle. (4 tests).2. Embedment lengths for other lengths and angles are less certain.3. The more the bar is bent the higher are the residual stresses, which reduce the tensilestrength of the re-bar. The highest force was recorded for 15 degree angle and the lowestfor the 45 degree angle. The presence of welding in this area of the re-bar also decreasesthe tensile strength of the bar.4. The longest displacement at failure was observed for the 45 degree angle and the shortestfor 15 degree angle. The reason for that is that the steeper angles create higher compressionperpendicular to grain stresses and hence longer displacements.FORCE DISPLACEMENT RELATION FOR TENSION FAILURE IN THE RE-BARS.The following graph presents the force displacement relation for Specimen SP-9,typical for the tension failure of the re-bars. The embedment length was 300 mm. Theinclination was 30 degrees.00^2.5^5^7.5^10 12.5 15 17.5 20 22.5DISPLACEMENT (mm)2535Figure 4.5 Force displacement relation for specimen SP-9.The relation between force and displacement can be divided in to two areas. The firstone up to the force of 120 kN is a linear relation. Then a non linear part consists of yieldingof the rod and strain hardening, similar to the straight tension tests. (No bent in re-bar)SUMMARY FROM TESTING OF #20 RE-BARS.1. When the embedment length is more than 260 mm for 30 degree angle of hole inclination,the failure occurs in the re-bars.2. The more the rod is bent, the less is the force needed to break it in tension. This is causedby the combination of welding stresses, cold bending, and compression perpendicular tograin component.3. The force displacement diagrams show the following behavior : linear relation, followedby yielding, strain hardening, and rupture.364.7.3 TEST WITH AN INSTRUMENTED BARAfter testing the #10 re-bars, it was observed that the re-bar will yield if the embedmentlength is at least 300 mm. The test however did not show the minimum embedment lengthneeded to create the tension failure in the re-bar.Another way of determining the effective length of the re-bar, would be to find thestress distribution along the embedded length. That was the objective of this test.The instrumented bar test was performed with an embedment length of 380 mm, andan inclination of 30 degrees. A reinforcing bar was equipped with strain gages. Flat surfaceswere machined, on one side of the bar, used as bases for the strain gages. Six strain gageswere attached to the bar using a special glue. The first was placed at 45 mm from the glulamsurface, others at 55 mm, 70 mm, 90 mm, 130 mm, and 210 mm For the location of straingages see Photograph 4.4, and Figure 4.8.Figure 4.7 represents forces calculated from the readings of the strain gages at thedifferent locations showed on X-axis. The applied force can be read on Y-axis.Six load levels were used. They are : 4.2 kN, 8.5 kN, 12.7 kN, 16.8 IN, 19.8 kN, and25.8 kN.37Photograph 4.4. Instrumented #10 re-bar.Figure 4.7 Force distribution along the re-bar for increasing loads.The initial increase of the force is caused by the fact that the readings were taken only38on one side of the re-bar. The shape of the lines on the graph indicates rapid decrease offorce up to strain gage 4 situated 90 mm inside glulam. See Figure 4.7.At this point 70 of the force is already transferred to glulam. From here the decreaseof the force is less. The last strain gage was situated 210 mm inside glulam. Its readingindicated that at this point 93 % of the force was transferred to glulam. Assuming the samerate of decrease , the force was fully transferred to glulam by the re-bar length of 250 mm.Figure 4.8 presents the distribution of the stress along the re-bar loaded with a hori-zontal force of 26 kN.Figure 4.8 Stress distribution along the re-bar.CONCLUSIONS 1. The tensile force in the glued-in re-bar is transferred to the glulam mostly by the sectionof the bar closest to the glulam surface. (70 % of the force is transferred by the first 90 mmof re-bar length.)2. The transfer of the force is less along the next length - 120 mm of the re-bar. (23 % is39transferred by this section of re-bar)3. Approximately 250 mm of # 10 re-bar length is needed to transfer 100 % of the tensileforce to the glulam, under the load of 26 IN, which is 30% of ultimate.CONCLUSIONS FROM TESTS WITH \"PRELIMINARY CONNECTION\"1. Safe embedment lengths to insure tension failure in the re-bar were established :250 mm for #10 re-bar300 mm for #20 re-bar2. The inclination was chosen as 30 degrees.3. Sufficient bearing area for compression perpendicular to grain has to be provided.4. Appearance of the connection must be improved, in order to have an acceptable product.404.8 PRE-WELDED CONNECTION4.8.1 PRE-WELDED CONNECTION USING GLULAMAfter the tests with the preliminary connection, the joint was redesigned, and calledthe \"prewelded connection\". For the description of the development of the joint see section4.3. Four specimens with pre-welded connection were prepared : two with a single #20re-bar, and two with double #20 re-bars. (side by side). All four specimens failed by tensionfailure in the re-bar. See photograph 4.5.Photograph 4.5 Pre-welded joints after failure.The following table presents tests results.SPEC # Ld(mm)ALPHA(deg)Fmax(kN)Dmax(mm)STRESS(Mpa)CON-1 380 30 214.8 12.5 595CON-4 380 30 158.4 8.3 438Average 188.6 10.4 517Table 4.6-A. Pull out tests results. Pre-welded joints. Single #20 bar.41SPEC # Ld(mm)ALPHA(deg)Fmax(1(N)Dmax(mm)STRESS(Mpa)CON-2 2 x 380 30 334.4 18.8 463CON-3 2 x 380 30 354.9 10.7 492Average 344.7 14.8 478Table 4.6-B. Pull out tests results. Pre-welded joints. Double #20 bar.SUMMARY1. Tension failure in the re-bars was observed for both single and double rod joints.2. 26% difference in ultimate force values was recorded for the single bar specimens. Thereason is the variability of steel material used for reinforcing rods, and possible weldingdamage. No sign of wood failure was noticed, in any of the specimens. For double barspecimens the difference between the ultimate forces was only 6%.3. The double bar joints failed at 90% of twice the force of single specimens.FORCE DISPLACEMENT RELATIONS The following two graphs present the force displacement relations for single anddouble bar joints. See Figures 4.7 and 4.8. The force displacement relations showed twotypes of behavior : linear and nonlinear. For the smaller forces the displacement relationremained linear. After the re-bar started to yield the displacement was growing faster thanthe force and a yielding plateau was observed.For double bar specimen linear behaviour and yield plateau can be easily separated.See Figure 4.8. Single bar specimen relation is less clear and shows 30% shorter ductilityrange than that of the double joint.10 mm ductility rangeLinear behavio25^5^7.5^10^125^15^17.5DISPLACEMENT (mm)4217.5 2025^5^7.5^10^125^15DISPLACEMENT (mm)Figure 4.7 Force displacement relation for specimen CON-1.Figure 4.8 Force displacement relation for specimen CON-2.CONCLUSIONS FROM TESTING PRE-WELDED GLULAM JOINTS1. In the prewelded connections, the compression perpendicular to grain stresses had beendesigned not to exceed the bearing capacity. They were as follows :5.9 Mpa for single bar specimens.5.3 MPa for double bar specimens.43As a consequence no compression perpendicular to grain failures were observed.2. Tension failures in the re-bars were observed for double re-bar specimens. The two barsfailed simultaneously.3. Manufacturing techniques for the pre-welded connections were developed.4.9 CONCLUSIONS FROM PULL-OUT TESTS.1. The pull-out tests provided information that enables us to design structural timber con-nections which will fail by tension failure in steel without affecting the wood.2. Safe embedment lengths to get re-bar tension failure are as follows :- 250 mm for #10 re-bar- 300 mm for #20 re-bar3. Specimens with double bars have the same failure mode - tension failure of the re-bar,as single bar joints.4. 30 degree hole inclination to grain results in optimum force values and enhances the shearcapacity of the cross-section.5. A manufacturing technique for pre-welded joints was developed for single and doublebars. The joints can be prefabricated in the glulam plant and erected similar to steel structureby using steel straps and bolts.6. The force displacements relations of specimens which failed by tension failure in there-bar, are similar to normal steel tension tests.7. The glue bond failures occur at stresses above yield and close to ultimate tensile resistanceof the re-bars.44CHAPTER 5BEAM TEST5.1 TEST OBJECTIVESThe subsequent step in the development of the pre-welded glued-in rod connectionwas manufacturing and testing of a beam splice connection, with double bars top and bottom.See Figure 5.1.Figure 5.1 Beam splice connection using glued-in rods.TEST OBJECTIVES1. Develop a manufacturing method for a beam splice connection.2. Check the behaviour of a \"real\" connection - joining two glulam elements.3. Compare the maximum tensile force in the re-bars, at tension failure of the re-bars, withpull out test results.1500 mm 1500 mm300 mmi600 mm1300 mmII1 1111376 mm170 mmLVDT's BOTTOM!, BOTTOM2ON EACH SIDE455.2 BEAM SPLICE CONNECTIONA splice connection using double #20 re-bars, top and bottom, was placed in the middleof the span of a 3.6 m long glulam beam. The cross-section of the beam was 175 x 380 mm.The embedment length of the re-bars was increased from 380 mm ( pull out tests ) to 470mm This was done to develop a manufacturing technique for gluing longer re-bars. Longerre-bars increase the shear capacity of the cross-section, which may be useful for otherapplications of glued-in rod joints. 30 degree hole inclination was used. Holes were inclinedout of plane, so the bars could cross in the middle of the cross-section, without piercing eachother.5.3 TEST SETUP AND TESTING METHODThe beam was loaded in bending by two point loads (0.6 m apart) in the middle of thespan.Figure 5.2 Beam test setup and instrumentation.46The following measurements were recorded :1. Loading force from the load cell.2. Mid span deflection.3. The movement of the gap between the two glulam pieces by 4 LVDT's placed at thecorners of the cross-section.Instrumentation of the specimen is presented on Photograph 5.1, and Figure 5.2.Photograph 5.1 Instrumentation of the beam test specimen.10^20^30^40^50^60^70MID SPAN DEFLECTION (mm)47A load controlled load application was used. The beam was loaded in five loadincrements with maximum forces of : 50, 75, 100, 125, and 122 - failure. After each loadingspecimen was unloaded.5.4 BENDING MOMENT - MID SPAN DEFLECTION ANALYSISRelation between the bending moment in the middle of the span and mid spandeflection, for all five loadings is presented on Figure 5.3.Figure 5.3 Mid span deflection and bending moment relation for all loadings of the beamtest specimen.Large permanent deflections were noticed after each loading cycle. Before the lastloading permanent deformation reached 44 mm.Maximum bending moment value was 92 kNm which was 95% of the glulam resistance,calculated assuming maximum bending stress for 24f-EX D-Fir glulam - 30.6 MPa.The specimen failed by tension failure in the re-bars, at the bottom part of the joint.See Photograph 5.2. The maximum force in the re-bars was 237 kN, which was 70 % of theaverage force for the double # 20 re-bar pull-out specimen. The maximum stress in there-bars was 344 Mpa.Photograph 5.2 Re-bar tension failure in the beam specimen.5.5 GAP MOVEMENT ANALYSISGap movement analysis is based on the readings of four LVDT's measuring theopening and closing of the gap between two glulam pieces. Figure 5.4 presents the gapmovement during loading #4, where the stress in the re-bars reached 344 Mpa.The permanent bottom gap opening after this loading reached 6 mm. On the top sideof the joint, permanent gap closing was 1 mm.48490^10 20 30 40 50 60 70 80 90 100BENDING MOMENT (kNm)Figure 5.4 Gap movement and bending moment relation. Beam test. Loading #4.5.6 CONCLUSIONS1. The beam splice joint using glued-in rods failed by tension failure in the re-bars, whatindicates, that the joint connecting two glulam elements can fail by the steel failure.2. Force in the re-bars during failure was lower than the average from the pull-out tests.- beam test F = 237 kN, stress in re-bars 344 MPa- pull out tests F = 345 kN, stress in re-bars 478 MPaThe reason for the difference in the maximum force values may be different bearingconditions. During pull-out tests the specimens rested against the rigid steel plate, whileduring the beam tests two glulam parts rested on each other.3. Very large deflections were noticed during the beam test. A solid beam of the same crosssection and span would deflect - 12.5 mm. The glued-in rod joint deflected 76 mm. Therewere several reasons :A. Oversize holes for the studs. (2 mm oversize)B. Compression perpendicular to grain failure at the bottom of the joint.4. The inside steel plate area should be designed to resist compression perpendicular tograin.5. Beam splice joint was very labor intensive, relative to the information obtained, because50two joints have to be manufactured.6. The test setup was not effective since load could not be reversed.7. It was decided to change the test setup. New tests should be able to load the glued-inrod joints under cyclic loading - tension and compression, in order to provide better infor-mation for the earthquake behaviour of the joints.51CHAPTER 6RACKING TESTS6.1 TEST OBJECTIVESThe racking test was designed to be a more efficient test method than a beam test, asdiscussed in the previous chapter.TEST OBJECTIVES 1. Design a glued-in rod column foundation joint.2. Develop a manufacturing method of making a glued-in rod joint, connecting a glulamcolumn to a steel foundation plate.3. Check the behaviour of glued-in rod joints under cyclic loading. Similar to a columnloaded with wind or earthquake loads.4. Check the possibility of tension failure in re-bars, for joints equipped with 2, 3, and 4re-bars.5. Evaluate the change of capacity of a joint with the increase of bearing area.6. Check the moment capacity of multi bar glued-in rod joints and compare the ultimateforces with pull-out and beam tests.7. Examine vertical and horizontal displacements of the joint area when subjected to cyclicloading.8. Find the moment-rotation behaviour.6.2 RACKING TEST SETUPRacking test setup ( See Photograph 6.1 ) consists of two parts :1.A loading jack with a load cell bolted to a steel frame, fixed to the concrete floor by bolts.2. Foundation plate, fixed to the floor by bolts.A stiffened bracket was manufactured, and attached to each side of a test specimen.One leg of the bracket was attached to the foundation plate. The other was attached to the52glulam via the glued-in re-bar. See Figure 6.1.Photograph 6.1 Racking tests setup.6.3 TEST READINGS AND INSTRUMENTATION OF THE SPECIMENThe following measurements were recorded during racking tests :1. The loading force from the load cell.2. The horizontal displacement of the column at the height of the force.3. The vertical displacement, at the glulam face, was measured by four LVDT's situated atthe corners of the glulam cross-section.GLULAM 170x490 mmFORCE^DISPLACEMENTVERTICALDISPLACEMENT534. The horizontal displacements of the outside connector was measured by two horizontalLVDT's.From the above measurements the movement of the lower end of glulam face can bedetermined. See Figure 6.1.Figure 6.1 Instrumentation of the racking tests setup.6.4 EVOLUTION OF THE RACKING TEST SPECIMENEach racking test specimen was equipped with two glued-in rod joints of the same size.Specimen 1, which was the undamaged part of the beam test, had a joint with two #20re-bars.Specimen 2 had three #20 bars - See Figure 6.2Specimen 3 had four #20 re-bars - See Figure 6.3GLULAM170x490mmRE\u00E2\u0080\u0094BARS54TOP VIEWBOTTOM VIEWFigure 6.2 Specimen 2.TOP VIEW^______\u00E2\u0080\u009E___^,..,.-,.. -..^---. -,^GLULAM...... ...\u00E2\u0080\u009E^-,.. --..--,^--. ,--... ^ ---. --.,^170x490mm-,.. ,^--... ...--. --...^, ....... --.-N., ,..)^-.....>i7 i 7-- --^, .---- --^.- --, -- --.- --^..--- --^, --RE\u00E2\u0080\u0094BARS----^,..- ...--- --^..-- --_. _., \u00E2\u0080\u0098....\u00E2\u0080\u0094 \u00E2\u0080\u0094i... .--BOTTOM VIEWFigure 6.3 Specimen 3.Specimen 4 had four #20 re-bars similar to Specimen 3 but the distance between therods parallel to grain, was increased and so was the length of the inside plate. The length55of the steel plate was increased from 250 mm to 400 mm. This was necessary because of thehigh compression perpendicular to grain stresses developing in the joint. See Photograph6.2Photograph 6.2 Specimen 4 during manufacturi.g.6.5 LOADINGThe same loading pattern was used for all racking tests. The specimens were loadedby a horizontal force 1.5 m above the foundation plate. The force was applied in twodirections. The first load cycle was 10 kN in both directions. The next loadings were applied40z..w aciaacr. 2040056with increment equal to 10 kN. The load control pattern was used.The specimens were loaded with four full load cycles at each load level. After eachload level, the specimen was unloaded and the data was saved.Figure 6.4 presents the relation between loading force and time.Figure 6.4 Loading force and time relation during racking tests.6.6 MOMENT DISPLACEMENT DIAGRAMSThe relations between the bending moment and the displacement illustrate the overallbehaviour of the joint. The following sections present this relation in graphical form forselected loadings. The general character of the graph is discussed. The analysis of maximumbending moments and displacement at failure is presented. The diagrams present the averageof four cycles at the stated load levels.6.6.1 SPECIMEN 1. (2 -#20 re-bars)A linear relation between the bending moment and the displacement was present onlyduring the first load level. (Maximum bending moment 15 kNm) Then the loops wereprogressively pinched up to the final load, where there was almost no moment resistancebut a displacement of 40 mm. See Figure 6.5. Specimen 1 failed by tension failure in there-bars, at a bending moment of 92 kNm and a vertical displacement of 62 mm. See Pho-tograph 6.3.The maximum force in the re-bars was 237 kN, which resulted in a stress of 395 MPa.57The ultimate value was similar to the one obtained during the beam test.Figure 6.5 Bending moment and displacement relation for Specimen 1.RESULTS: 1. Specimen 1 showed a very weak behaviour under cyclic load. The reason might be thatthis joint was loaded previously during the beam test.2. The final failure mode was tension failure in the re-bars. The two re-bars failed in tensionat the same load.6.6.2 SPECIMEN 2. (3-#20 re-bars)Specimen 2 maintained a linear relation between bending moment and displacementup to a moment of 125 kNm. See Figure 6.6. Then up to 175 kNm a slightly pinched behaviourwas observed.The first failure occurred at 200 kNm, when the single bottom bar failed. The momentdropped to 150 kNm, but subsequently increased to 225 kNm, at which point, the next twore-bars failed. The load was then reversed. On the other side of the joint one re-bar brokeat 175 kNm, and the moment held steadily at 150 kNm, up to the end of stroke in the actuator,due to yielding of re-bars.58250200150100W50O\u00E2\u0080\u00A2 02(-7\u00E2\u0080\u00A2 -50Oz -100Lu-150-200 -60^-40^-20^0^20^40^60^80DIPLACEMENT (mm)Figure 6.6 Bending moment and displacement relation for Specimen 2.RESULTS: 1. Specimen 2 showed linear behaviour up to the moment of 125 kNm, which was a bigimprovement relative to Specimen 1.2. The failure mode was tension failure in the re-bars, on both sides of the joint. The singlere-bar, failed first, followed by the simultaneous failure in the two bars in the second row.(Close to the foundation plate)3. When the glued in rod joint consists of two rows of re-bars, the row closest to the end ofthe specimen fails first.6.6.3 SPECIMEN 3. (4-#20 re-bars)Specimen 3, despite having one bar more in the joint (4 #20 re-bars) behaved linearlyup to the same load level as Specimen 2.(M=125 kNm) See Figure 6.7. The pinching ofthe loops was more pronounce when the moment reached 150 kNm. During the next loadingthe two outside bars failed at 150 kNm.After the load reversal, the specimen failed abruptly by tension failure in all fourre-bars. See Photograph 6.4. The maximum bending moment was 200 kNm.-250^-8059^250 ^200-1 50 ^1 0050-60^-40^-20^0^20^40^60^80DISPLACEMENT (mm)Figure 6.7 Bending moment and displacement relation for Specimen 3.RESULTS: 1. Specimen 3, despite of the increase of re-bar area failed under similar load as Specimen2. The reason might be, that the bearing area was to small for 4 bar joint, and wood crushingtherefore governed the moment capacity.2. Four re-bars in two rows failed at the same instant.6.6.4 SPECIMEN 4 (4420 re -bars, plate length increased to 400 mm)With the increase of bearing area, the moment displacement relation of Specimen 4become different from previous specimens. See Figure 6.8. Linear character of the graphterminated at the bending moment of 50 kNm. Then the loops are more and more pinched.The same trend continues up to the bending moment of 245 kNm.Specimen 4 failed by shear failure in the studs during the next loading at the bendingmoment value of 200 kNm. Then the load was reversed and the bending moment reached245 kNm without failure, at which point the stroke limit was reached in the actuator.00 -805z -100eb-150-200-250-802 re-bars fail60200 ^^50 .4 ce....iiiii-50 ^/?OD ^... ^/ .. ^i^-80^-60^-40^-20^0^20^4DISPLACEMENT (mm)Figure 6.8 Bending moment and displacement relation for Specimen 4.RESULTS: 1. With the increase of bearing area, moment capacity of the joint increased, because oflower compression perpendicular to grain stresses.2. The character of the moment displacement relation changed showing bigger area insidethe loops, which indicates more energy absorption.3. It is evident from Figure 6.8 depicting the behaviour of Specimen 4 that the increasedbearing area resulted in a joint with an acceptable behaviour. The loops are regular and lesspinched, which is desired for the earthquake resistant structures. This was the objective ofthe development.tu0zz8061Photograph 6.3 Specimen 1 after failure. (above)Photograph 6.4 Specimen 3 after failure. (below)626.7 ULTIMATE MOMENTS, AND FORCES IN THE RE-BARS OBTAINEDFROM TESTS AND A SIMPLIFIED MODEL.The following graph presents forces in the racking test specimen.Figure 6.9 Forces in racking test specimen.The calculation of ultimate bending moments assuming tension failure in the re-bars.Assumed maximum yield stress in re-bars is : Fv = 400 MPa, alpha = 30 degrees.Shear force V2 is assumed small in comparison to Fm, and it was neglected, for the calculationof Fr.Frd = Fv x ArebFmd = Fr / cos(30)Mrd = Fm x 1where :Frd - force in re-bars (model)Fmd - force in the steel plate (model)Mrd - bending moment at the base (model)The following tables present design and test values obtained from analysis - Table 6.1and recorded during testing - Table 6.263Specimen Frd(kN)Fmd(kN)Mrd(kNm)1 240 276 932 360 414 1843 480 478 2144 480 478 214Table 6.1 Design yield forces and bending moments in racking test specimens.Specimen Frt(kN)Fmt(kN)Mrt(kNm)stress(MPa)Frt/Frd1 233 268 90 388 0.972 341 392 175 378 0.943 390 448 200 325 0.814 478 549 245 398 1.00Table 6.2 Test results of ultimate forces and bending moments in racking test specimens.where :Frt - force in the re-bars (test)Fmt - force in the steel plate (test)Mrt - bending moment at base (test)RESULTS: 1. Comparison of the design and the measured internal forces shows the effectiveness ofglued in rod joints tested during racking tests. The most effective was Specimen 4. Specimen2 was close, but the design of Specimen 3 was not balanced. This indicates the earlier drawnconclusion that not enough bearing area was provided in the joint of Specimen 3.2. When glued in rod joints are properly designed, a ratio of stress obtained during tests atfailure to steel yielding stress is around 1.06.8 ANALYSIS OF BEARING STRESSES OBTAINED FROM ANALYSISAND MEASURED DURING TESTS.The following table presents the maximum bearing force, assuming fc = 5.9 MPaconstant on the whole area. Fct is the maximum bearing force obtained during testing.64Specimen Ab(mm ^ 2)Fc(kN)Fct(kN)Fbt/Fb%1 30500 180 155 862 38000 224 227 1013 38000 224 260 1164 60800 359 318 89Table 6.3 Maximum bearing forces from tests and analysisRESULTS: 1. The bearing stresses analysis indicates, that the design of Specimen 1 and 4 was balanced.Specimens 2 and 3 were over stressed.6.9 ANALYSIS OF BOTTOM END MOVEMENTS UNDER LOADThe movements of the bottom end of the specimen were monitored by four verticaland two horizontal LVDT's. The readings of these instruments were used to compare thebehaviour of the two sides of the joint. They were also used to find the components of theoverall displacement at the height of the force.The vertical LVDT's provided data used to establish joints rotation. The horizontalLVDT's readings were used to obtain the sliding shear deformation. For the location ofLVDT's see Photograph 6.565Photograph 6.5 Location of LVDT's during racking tests.6.9.1 VERTICAL MOVEMENTSDuring the analysis of the vertical movements of Specimens 2 and 3, different beha-viours of two sides of the joint were observed. See Figure 6.10, which illustrates the behaviourof Specimen 3.so 80 100EE0.5(/)0U-1.5-100 -80 -60 -40 -20^0^20 40BENDING MOMENT (kNm)66Figure 6.10 Vertical displacement of the joint area versus bending moment for Specimen 3.During the first half of the load cycle, side A of the connection, is under compressionand side B under tension. See Figure 6.11. The re-bar is bent outside under compressionand deformed inside, crushing the wood on the tensile side. This should not happed inproperly designed joint. When the loads are reversed, the re-bar which was bent outside isbeing straightened by the tensile load.On the other side of the joint, the crushed-in re-bar is moved even more inside thewood under compression. Therefore side A shows the linear steel type behaviour, while onside B, because of the constant crushing under both directions of load, the non linearbehaviour is noticed.This explains why on side A lines are closer to each other during loading. On side Bthe loops are more apart because of the combined steel and wood crushing behaviour. Whenthe bearing area was increased for Specimen 4 both sides represented \"steel type\" behaviour.67 LINEARBEHAVIOR\u00E2\u0080\u0094 STEELADEFORMATIONB44^NONLINEAR(..) BEHAVIORrzo \u00E2\u0080\u0094 WOOD4.. CRUSHINGDEFORMATIONFigure 6.11 Movements of the steel plates during loading.6.9.2 HORIZONTAL MOVEMENTSThe same phenomenon of different behaviour of two sides was noticed while analyzingreadings of horizontal LVDT's. See Figure 6.12, presenting the data obtained duringSpecimen 3 experiments.^0.8^-g 0.6-E^-Wj OA--U^20, -0LI^0 -cr)0-a2-Z^-NN -0.4rr^-OI -0.6--0.8-160^-80^-40^0^40^80^120BENDING MOMENT (kNm)Figure 6.12 Horizontal movements of the joint area, Specimen 3.-120 160686.10 DISPLACEMENT COMPONENTSThe displacements components were established in order to separate the componentsof the overall deformation at the height of the force. The percentage of glulam deformationsand hinge rotation as well as sliding shear deformation is presented for the racking testspecimens in graphical form in Figure 6.13.The total horizontal displacement d to ,a , = d flex + d rt + d sishwas due to the elastic and in-elastic deformations of the components namely :* bending and shear deformation of glulam - d 11 \u00E2\u0080\u009E,,** hinge rotation - drt*** sliding shear deformation - d sishd flex d rt d sishd flex = d bend + dsh^d rt^ d slshFigure 6.13 Displacement components of the racking test specimens.All the displacement were estimated using simple models.- was calculated according to the static rules, as a cantilever beam. (calculated)d rt^- was calculated using the readings of the vertical LVDT's, as a rigid body.(measured)d sIsh^^- was calculated using the readings of the horizontal LVDT's, as a translation.(measured)6.10.1 The flexure deformationThe flexure deformation has two components : bending and shear.d flex = d bend + dsh69The flexural component can be calculated from the formula:Pxh3d bend \u00E2\u0080\u00943XEXI.where :^P - horizontal forceh - distance from the force to the joint.I - moment of inertia of the cross-section about the major axis ( mm ^ 4)E - modulus of elasticity of timber ( MPa )The shear component of the deformation can be estimated bykxPxhd\u00E2\u0080\u009E= AxGwhere :^k - shear distribution factor - 1.5G - shear modulus ( assumed E/15 - 867 MPa)A - cross-section area of the specimen ( mm ^ 2 )6.10.2 Deformations due to hinge rotation.Vertical LVDT's were placed at the corners of the joint region to measure the verticaldisplacements dl and d2. See Figure 6.14 drt /-----^/ ^-\\L/ 7---- - -I ^j / /^---^/ / )^/ ^I^/ // // / /& ^ / I/ /,...^ /,, u //I ------^jo^// !d 2h//// dd rt^ \u00E2\u0080\u0094^ eFigure 6.14 Deformations due to hinge rotation.70Deformation due to hinge rotation can be calculated from the following formula :d rt = h x tan0where :(j) d 1 + d 2-^ (dog)ewhere : dl, d2 - readings from vertical LVDT's (mm)e - distance between LVDT's (mm)6.10.3 Deformations due to sliding shear.The horizontal deformation of the specimen relative to the foundation plate wasmeasured using LVDT's placed horizontally at two sides of the joint. Deformation due tothe sliding shear was taken as the average of two LVDT readings. The horizontal LVDt'swere placed during the last two experiments: Specimen 3 and Specimen 4.d sash-.....\u00E2\u0080\u0094..--J\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2.---11.\u00E2\u0096\u00A0g1d slshFigure 6.15 Deformations due to sliding shear.g 1 + g 2d sish 2Where : gi, g2 - readings from horizontal LVDT's (mm)6.10.4 Total deformationsAll the components mentioned above can be now summed up to find the overallhorizontal displacement.d total= d f lex + d rt + d sishSPECIMEN 110 20 30 40 50 60 70FORCE (kN)SPECIMEN 280 90 100 11010 15 20 25 30 35FORCE (kN)FORCE (kN)10 20 30 40 50 60 70 80 90 100FORCE (kN)71The following graphs present both calculated and measured displacement for all fourracking test specimens.-4.- CALCULATED DISPL -.- MESSURED DISPL -0- MESURED --4-- CALCULATEDSPECIMEN 3 SPECIMEN 4--1.- MESURED --.- CALCULATED -\u00E2\u0080\u00A2- MEASURED --o- CALCULATEDFigure 6.16 Calculated and measured displacement of racking tests specimens.The deflection shown on Figure 6.16 as \"calculated\" is the summation of the calculatedstatic deflection and measured : rigid body rotation and sliding shear.The next four graphs present how these components were changing with the increasingloads, for all racking test specimens.zLU2908070LU 60500_ 4030-J201000O 10^20^30^40FORCE (kN)NAKX.1:Kv`IlLeh\u00E2\u0080\u00A2 NLN. , 0#\u00E2\u0096\u00A0,,AK\u00E2\u0080\u009ENy\ 45-,Xv5,2,6\ tow, %wowE72flexure^,V hinge rotationFigure 6.17 Horizontal displacement components - Specimen 1.EE 45I-- 40lL\u00E2\u0080\u00A2 35O 3\u00C2\u00B05 25a.co 20E-I 15I\u00E2\u0080\u0094 10O 50 10 20 30 40 50 60 70 80 90 100 110FORCE (kN)flexure^ZSS hinge rotationFigure 6.18 Horizontal displacement components - Specimen 2.EE 25-7-Lu25 15-'a.a 10 - '-J00 10 20 30 40 50 60 70 80 90 100FORCE (kN)73flexure^sliding shear ES hinge rotationFigure 6.19 Horizontal displacement components - Specimen 3.EE 601\u00E2\u0080\u0094z 502 405o_ 30200 1000 10 20 30 40 50 60 70 80 90100110120130140150160FORCE (kN)flexure^gm sliding shear ES hinge rotationFigure 6.20 Horizontal displacement components - Specimen 4.74SUMMARY1. The largest component comes from the rigid body rotations, which is about 70% of theoverall displacement.2. The second largest was glulam deflection around 30%.3. Sliding shear component was less than 10%. The source of sliding shear componentmight be the slipping of bolts in the concrete floor.6.11 CONCLUSIONS FROM RACKING TESTS1. Proper bearing area has to be provided to obtain the full capacity of the glued-in rodjoint.2. Properly designed joint fails under a bending moment equal to Mr obtained from analysis,when the ultimate stress in re-bars is assumed Fv = 400 MPa.3. Detailed analysis of the joint area showed that when the bearing area is too small, thetwo sides of the joint behave differently under load. This phenomenon is not significant inthe design process.4. Tension failure in the re-bars is present in multi bar joints. Maximum steel area that failedat the same instant was 1200 mm \" 2 - 4-#20 re-bars.5. Analysis of the deformation components showed that 70% of the deformation of theglued in rod joint comes from joint rotation. The joint rotation may come from movementsof studs in oversized holes, steel plates deformations, and small re-bar movements.75CHAPTER 7KNEE JOINT TESTS7.1 TEST OBJECTIVESThe last type of tests described in this report is the knee joint test. Four full scalespecimens, of the same geometry as designed in Appendix A, were tested under cyclic loading.TEST OBJECTIVES: 1. Develop a technique for making a glued in rod knee joint.2. Design a joint that fails by tension failure in re-bars on the exterior and interior of theknee joint connection.3. Compare the ultimate bending moments (positive and negative) from testing with thosefrom analysis.4. Determine the rotational stiffness of the knee joint.5. Establish the behavior of welded and bolted glued in knee joint.6. Investigate the behaviour of glued in rod joints under reversed loading, assuming negativemoment (wind, earthquake) around 20% of positive moment. (gravity load)7.2 SPECIMENS - types of joints, and manufacturing stepsFive specimens were tested during the knee joint test experiments, which are describedin this thesis. Specimen 1 was a preliminary one, to develop the manufacturing techniqueand to check the likelihood of tension failure occurring in the re-bars. It was built frommaterial leftover from the racking tests.The glulam for specimens 2,3,4, and 5 was prepared specially for the purpose of kneejoint tests. The geometry of the knee joint was the same as one designed in chapter 3, wherethe preliminary frame analysis was presented.The full size specimens consisted of the upper half of the column, and the lower halfof the rafter connected by the glued in rod joint. The exterior connection had four # 2076re-bars on each end, while the interior was equipped with two # 20 re-bars glued into eachpart of glulam.In Specimen 2 the interior steel plates and exterior strap were connected by welding.Specimens 3,4 and 5 were of the same geometry but bolts were used for connectingthe two parts.MANUFACTURE OF THE CONNECTIONThe glulam for the knee joint tests was made in Penticton by Structurlam Ltd. Theinitial step, to prepare the glued in rod joint, was to cut out the required gaps in the glulamand drill holes for the reinforcing bars. The geometry of the openings is presented on Figure7.2.The next step was to make the steel parts of the joint. The plates and straps were cut,the holes for the bolts ( Specimen #3 to #5 ) drilled and taped. The straps had the samecross section as the inside plates, and were fixed to each other by welding. See Figure 7.1.The reinforcing bars were cut to the desired lengths and bent to fit inside the holes sothe flat part of the re-bar was touching the plates. The re-bars were then inserted in theholes and tackwelded to the plates, into the correct position. The assembly was then pulledout and the final welding was performed away from glulam. This was done to avoid burningof the glulam, and prevent damage to epoxy glue.After the welding the assembly was glued to glulam by epoxy glue. One side, exterioror interior was glued at a time. After one day the specimen was flipped around and the steelplates were glued on the other side. The curing time used before testing was seven days.After the glue had cured, the column and the rafter were fixed together by straps,which were bolted or welded to the plates. On the interior part of the connection, the strapwas reinforced by two steel plates forming stiffeners. See Photograph 7.2EXTERIOR l1IEXTERIOR PLATE^II\\\ \\ \INiiElIOR PLATE\\///////^/ ^1^/ I/ 1/INTERIOR STRAPINTERIORzEXTERIOR STRAP77Figure 7.1 Knee joint using glued-in re-bars.U78Figure 7.2 Manufacturing steps of the knee joint. PREPARATION OF TIMBERSTEP 1GROOVES FOR THE EMBEDDED REBARSARE ROUTED TO A DEPTH OF 35mmFROM THE SURFACESTEP 2THE AREA TO BE COVERED BY THESTEEL PLATES IS ROUTEDTO A DEPTH OF 12.7mm TO ACHIEVEA FLUSH OUTER SURFACESTEP 325.4mm HOLES OF LENGTH 460mmARE THEN DRILLED AT 30 DEGREESPARALLEL TO GRAIN FOR THE REBARSAND POINT A & B ARE CHISELEDFOR EASE OF REBAR INSERTIONPREPARATION OF STEEL PARTSSTEP I^ STEP 2RE-BARS ARE CUT AND BENT INSIDE PLATES ARE CUT FROMFLAT BAR - 12x152 mm. HOLESARE DRILLED AND TAPEDSTEP 3RE-BARS ARE INSERTED IN THE HOLESINSIDE PLATE IS PLACED ON TOPAND TACKWELDED TO THE RE-BARSSTEP 4STEEL INSERTS ARE REMOVEDFROM GLULAM.WELDING OF RE-BARSTO STEEL PLATES(AWAY FROM GLULAM)GLUEINGSTEP IEPDXY IS MIXED AND POUREDINSIDE THE HOLESSTEP 2STEEL ASSEMBLIESARE INSERTED INTO THE HOLESLONGITUDINALSUPPORTACTUATORTRANSVERSESUPPORT797.3 TEST SETUPThe test setup for the knee joint tests consists of three major parts :1. Actuator with transverse support.2. Longitudinal support.3. Brace preventing vertical deformations.The specimec was connected to the jack, which was anchored to the concrete floor,and support via 50 mm diameter steel pins. See Figure 7.2. These pins were resting on 20mm thick steel plates fixed top and bottom on one side to the actuator and on the other sideto the longitudinal support.The force was then transferred to the wood by steel plates attached by glulam rivets.Two steel plates were nailed at either sides of the specimen. The setup is presented onFigure 7.3 and Photograph 7.1.BRACEFigure 7.3 Knee joint test setup.80Photograph 7.1 Knee joint test setup.7.4 INSTRUMENTATION OF THE SPECIMENThe data acquisition system was used to read 8 channels of data. The loading forcewas read from the load cell. The change of the distance between the pins was recorded bya electronic displacement gage.The opening and closing of the gap between the glulam members was monitored bytwo LVDT's , one situated on the exterior and one on the interior part of the joint.The movement of the steel plates, relative to the glulam, was recorded by 4 LVDT's.Instrumentation of the internal part of the specimen is presented on Photograph 7.2.81Photograph 7.2 Instrumentation of the knee joint test specimen.7.5 TESTING PROCEDUREAll specimens were tested under cyclic loading. Four cycles of each load step wereapplied. Load steps of 10 kN were used for Specimen 1, and 20 kN for other specimens.The specimen was unloaded, and the data was saved on disks, before the next load step wasapplied.The specimens were loaded with the bigger force under compression then undertension. The tensile force was about 20% of the compressive loading, because of the con-ditions in the design, performed in Chapter 2. See Figure 7.4. The first failure was designedto be the tension failure of the exterior re-bars, when the specimen was loaded incompression.COMPRESSIONTIMETENSION40 130 ^20100102082Figure 7.4 Loading history during knee joint tests.7.6 ANALYSIS OF TEST DATAThe observations from the tests would be presented for all the joints in the followingorder.1. Overall behaviour of the joints - displacement between pins, bending moments.2. Rotation of the joint.3. Movements in the steel plates.4. Conclusions.7.6.1 OVERALL BEHAVIOR OF JOINTS7.6.1.1 SPECIMEN 1Specimen 1 was a preliminary knee joint built from glulam left over from the rackingtests. The glulam cross-section was 175 x 495 mm. The joint was equipped with four #20re-bars on the exterior and two #20 re-bars on the interior part of the connection. Theinterior and exterior straps were connected by bolts, to the column and the rafter. ( boltdiameter = 16 mm ).83The following analysis is based on the relationship between bending moment at thejoint and the change of the displacement between the pins. Specimen 1 was loaded with 16different load increments, each having four cycles of tension and compression. The averageof four cycles is presented on the graph.The load was applied using a load control pattern. The loads were increased in 10 kNintervals.The following bending moment convention was used. Positive bending moment isoccurring when the outside part of the connection is under tension emulating the frameloaded with gravity loads. The specimen was loaded in compression. Negative bendingmoment is present when the inside part of the joint is under tension analogous to theearthquake or wind loads. The specimen was loaded in tension.The moment displacement relation was linear only during the first loading, when thebending moment reached 10 kNm under compression and 5 kNm under tension. Duringthe next, higher loadings a much stiffer behavior was observed under compression than undertension. The reason was the development of a hinge at the weld between the interior straps.This was an oversight, it should have had a stiffener.During the second to last loading wood crushing started at the peak loads (98% of theultimate load) on the exterior part of the joint. Specimen 1 reached its ultimate strength,when extensive wood crushing causing large decrease in the distance between the pins ; theend of stroke was reached in the actuator. The maximum positive bending moment was 148kNm, which was 90% of design glulam resistance, assuming a bending stress in the glulamof 30.6 kNm. (Stress in glulam during testing - 27.5 MPa)The specimen was then loaded in tension to failure. The failure was located at theweld connecting the re-bars and the steel plate on the interior part of the joint. It was causedby the mentioned development of the hinge at the weld joining the straps, since stiffenerhad not been used. See Photograph 7.3. The maximum negative bending moment was 49kNm. The relation between the bending moment at joint and the displacement betweenpins is presented on Figure 7.5. Displacement between pins is caused by joint deformation1 _-70 -60 -50 -40 -30 -2 -10 0^10 2084and members deformations.The specimen after failure in shown on Photograph 7.3. Note the wood bearing failureon the exterior part of the connection and the weld failure on the interior part of the joint.10050-tsz.. -50-i-z^ -100 -=i1_ -150-11\u00E2\u0080\u0094 -200 -zIII\u00E2\u0080\u00A2 -250--O -3002-350-400-80wood crushing failure30DISPLACEMENT BETWEEN PINS (mm)Figure 7.5 Displacement and bending moment relation for Specimen 1.85Photograph 7.3 Specimen 1 after failure.This was as mentioned a preliminary test, and should not be considered in the overall analysisof the joints.7.6.1.2 SPECIMEN 2Beginning with specimen 2 the geometry of the glulam was specially designed for thepurpose of testing knee joints, to satisfy the conditions described in Chapter 2.Specimen 2 had four #20 re-bars on the external side, and two #20 re-bars on theinternal part of the joint, in each of the members. The straps were welded to the insideplates. Specimen 2 was the only one where this welding was used. The purpose of this wasto check the behavior of the welded strap, and compare it with that of bolted straps.The interior part of the joint was equipped with two steel plate stiffeners, which werein the same planes as the re-bars. (See Photograph 7.2.) This was designed to prevent thedevelopment of a hinge at weld connecting the outside plates, which was observed, duringSpecimen 1 experiments.100Ei---500-50zO -100-150-2002 -25002 -300-350-40086The relations between the bending moment at the joint and the change of the distancebetween two pins were very linear for all the loadings except the final one. See Figure 7.6.The linear character was lost under loads creating bending moments grater then 275 kNm.The maximum bending moment at failure was 302 kNm. The specimen failed by re-bartension failure on the exterior part of the joint. The two outside re-bars failed in tension,and then the front part of the steel plate crushed into the wood. See Photograph 7.4 and7.5.inside re-bar failure ^----i----0.14\u00E2\u0080\u0094.0\u00E2\u0080\u00A21501111111gre-^^- , 1/-^i o^utside re-bar failure^0 ^-10^0^10^20^3130^-i0^-60^-60^-40^-30^-20 1DISPLACEMENT BETWEEN PINS (mm)Figure 7.6 Displacement and bending moment relation for Specimen 2.Specimen # 2Bending moment - positive (kNm) 302Bending moment - negative (kNm) 80Glulam capacity - Mr (kNm) 314Table 7.1 Ultimate bending moments at joint for Specimen 2.87Photograph 7.4 Specimen 1 after failure. Re-bar failure - phase 1. (above)Photograph 7.5 Specimen 1 after failure. Wood crushing failure - phase 2. (below)887.6.1.3 SPECIMENS 3,4 and 5Specimens 3,4 and 5 were identical. They were equipped with four #20 re-bars on theexternal side of the joint and two #20 re-bars on the internal side. The strap and the steelplates were connected by bolts. ( Eight 16 mm bolts were used on the exterior side and fourbolts on the interior side, in each part of the connection ) The connections had the stiffenersimilar to that of Specimen 2.The relations between the bending moment in the joint and the displacement betweenpins was very similar for all three specimens. See Figures 7.7,7.8 and 7.9. For the loadingsresulting in the moments less than 300 kNm, the relation is almost linear and identical forall four cycles of each loading. The failure was tension failure in the re-bars, for all specimensexcept Specimen 5 which failed under positive moment by wood crushing. In the otherspecimens two exterior re-bars failed in tension. Then the front part of the steel plate crushedinto wood. After the failure under positive bending moment the specimens were loaded tofailure under negative moment. In all three specimens the failure was tension failure in there-bars. The failure modes are presented on Photographs 7.6,7.7 an 7.8.The ultimate bending moments at joint for Specimens 3,4 and 5 are presented in Table 7.2.Specimen 4 failed in tension under a smaller moment than the other two because of a largedisplacement after failure on the outside of the joint, which bent the re-bars on the insidepart of the connection.Specimen # 3 4 5 AVERAGEBending moment - positive (kNm) 349.2 346.4 358.0 351.2Bending moment - negative (kNm) 71.7 41.0 68.0 60.0Glulam capacity - Mr (kNm) 314 314 314 314Table 7.2. Ultimate bending moments at joint for Specimens 3,4 and 5.8910050-E--50-7.3 -100 -- -150-1\u00E2\u0080\u0094 -200-z1-1-I\u00E2\u0080\u00A2 -250-O -300--350- outside re\u00E2\u0080\u0094bar failure-70 -60 -50 -40 -30 -20 -10 0^10 20 30DISPLACEMENT BETWEEN PINS (mm)Figure 7.7 Displacement and bending moment relation for Specimen 3.outside re\u00E2\u0080\u0094bar failure-70 -60 -50 -40 -30 -20 -10 0^10 20DISPLACEMENT BETWEEN PINS (mm)Figure 7.8 Displacement and bending moment relation for Specimen 4.^100^50-0-z- -50-s5) -100-- -150F- -200-zLLI -250-O\u00E2\u0080\u00A2 -300--350-^-400 ^-80Inside re\u00E2\u0080\u0094bar failure30-70I^I^ 1^I^I^I-60 -50 -40 -30 -20 -10^0100^50 ^0 ^-50}--^-z -100 -0^-150 ^I\u00E2\u0080\u0094 -200 ^--250 ^2w0 -300 --350 -^-400^-80DISPLACEMENT BETWEEN PINS (mm)Figure 7.9 Displacement and bending moment relation for Specimen 5.Photograph 7.6 Specimen 3 after failure.90Inside re\u00E2\u0080\u0094bar failurewood crushing failure1 '0^20 3091Photograph 7.7 Specimen 4 after failure. (above)Photograph 7.8 Specimen 5 after failure. (below)927.6.1.4 SUMMARY1. Stiffeners, at the interior of the joint, are needed to eliminate the formation of a hingein the strap on the interior part of the joint.2. The welded joint failed at a bending moment 50 kNm less than the average of the boltedjoints.The reason may be that under the peak loads the two welded plates transferred thewhole force to the back row of re-bars. Specimens 3,4 and 5, where the plates where bolted,failed at slightly higher loads because the interior and the exterior plates yielded, and thebolts moved in the oversized holes so more load was transferred to the front row of re-bars.The bolted specimens allowed more play in the joint region and were more flexible.The indication of this can be shown on the moment displacement graphs presented inthis section. Under failure the displacement between the pins for the welded specimen was33 mm, while for the bolted ones it was between 41 and 43 mm.3. The maximum positive bending moments for bolted specimens were close to each other.(Ultimate negative moment was measured after failure at the exterior of the joint)4. Bolted joint moment resistance, measured during tests, was 10% higher than nominalglulam capacity.937.6.2 THE ROTATION OF THE JOINTThe rotation of the joint was measured by two LVDT's : one situated at the exteriorcorner of the joint, and the other near the interior corner. The LVDT's were attached tothe side of the glulam members, with one part on the column and the other on the rafter.The instrumentation of Specimen 5, typical for all tests, is presented on Photograph 7.9.The bending moment at the knee, obtained from the analysis, under specified loadswas 150 kNm. Since the deflections are checked under specified loads the gap openingscaused by this moment are presented. See Table 7.3.The relations between the bending moment in the joint and the gap openings at thelocations of LVDT's are presented on graphs 7.8 to 7.11. The solid line on the graph rep-resents the movement of the interior placed LVDT. The dotted line shows the exterior gapopening.Photograph 7.9 Instrumentation of the specimen to measure the joint rotation.947.6.2.1 SPECIMEN 2 (welded connection)There was very little movement at the interior part of the joint. At ultimate loadings,the inside displacement was less than 1 mm The exterior Lvdt's showed much moremovements. Under specified loads the gap opened 3.5 mm, Andunder ultimate loads theopening reached 12 mm at the exterior side. This deformation was caused by compressionperpendicular to grain on the exterior part of the joint, as well as stretching of the straps.The relation between the bending moment at the joint and the gap opening is almost linearfor all loadings, with exception of the final one. See Figure 7.10.2015-E-Co 10-z0-5^-400 -350 -300 -250 -200 -150 -100 -50MOMENT AT JOINT (kNm)50 100Figure 7.10 Gap opening in Specimen 2. (dotted line - exterior, solid line interior)7.6.2.2 SPECIMEN 3,4 and 5 (bolted connection)The movements on the interior of the joint were larger for the specimens equippedwith bolted connections. The biggest increase can be observed on Figure 7.10 showing thebehavior of Specimen 4. However under specified loads the interior movements are lessthen 3 mm for all three specimens.95The exterior gap opening for the bolted specimens, under specified loads, was in thesame range as for the welded joint. It varied from 1.7 mm in Specimen 3, through 2 4 mmin Specimen 5, up to 3 mm in Specimen 4. The relation between the gap opening and thebending moment at joint is similar to that of Specimen 2, linear for all the loadings excludingthe failure ones. See Figures 7.11, 7.12, and 7.13.20,\u00E2\u0080\u00A2,.^...........................^.......^....^.............^\u00E2\u0080\u00A2^........ ....^..^....-........^...................^\u00E2\u0080\u00A2.......^.-5^-400 -350 -300 -250 -200 -150 -100 -50^0MOMENT AT JOINT (kNm)Figure 7.11 Gap opening in Specimen 3. (dotted line - exterior, solid line interior)50 100150 10zwa_02096\u00E2\u0080\u00A2 ......^\u00E2\u0080\u00A2................^.^.........^..........-5-400 -350 -300 -250 -200 -150 -100 -50^0^50 100MOMENT AT JOINT (kNm)Figure 7.12 Gap opening in Specimen 4. (dotted line - exterior, solid line interior)20...... ......^.. ........-5^-400 -350 -300 -250 -200 -150 -100 -50^0MOMENT AT JOINT (kNm)Figure 7.13 Gap opening in Specimen 5. (dotted line - exterior, solid line interior)50 10097The following table summarizes the resultsSpec GAP OPENING - (mm)# At specified load At ultimate loadexternal internal total external internal total2 2 1 3 13 1 143 2.5 1 3.5 12 1 134 3 1.5 4.5 12 3 155 1.5 1 2.5 10 2 12Table 7.3 Gap opening results.7.6.2.3 SUMMARY1. Gap opening analysis showed the difference between welded and bolted joints. For thewelded connection, the inside crushing of wood was less than 1 mm up to failure load. Forthe bolted joints, the inside crushing of wood reached 3 mm in Specimen 4. On the outside,gap opening was similar for welded and bolted joints. Under failure the joint opened around15 mm for all four specimens. This translates to a deflection of the real structure of 36 mmand 150 mm for the service load and ultimate load respectively, caused by rotation alone.2. For the design and specified load range, the relations between the gap opening and thebending moment at joint are very linear and consistent for all four cycles of each loading.987.6.3 MOVEMENT IN THE STEEL PLATESThe performance of the glued-in rod connection depends on the bond between there-bars and the wood. One of the methods to monitor this behavior is to check how muchthe steel plates move relatively to glulam surface under loading.During this series of tests, movement of all four steel plates was recorded. FourLVDT's were fixed to the wood with its moving parts rested on magnets attached to theplates. The following analysis uses the relation between the displacement of all four platesand bending moment at joint for selected loadings. (each loop represents different loading)7.6.3.1 SPECIMEN 2Very little movement was observed on the interior part of the joint for the weldedconnection of Specimen 2. The displacement under designed loads was less than 1 mm. Onthe exterior side of the joint a non-symmetrical behavior was observed.A much bigger movement was noticed in the steel plate fixed to the rafter than theone attached to the column, what is indicated by the location of failure which took place inthe outside bars of the rafter plate. The other plate moved less than 1 mm even under peakloads. For the relation between the steel plate movements and the bending moment at kneesee Figure 7.14. Solid lines on the graph represent the movements of the interior plates,while the dotted lines represent the movements of the exterior plates........ . '\u00E2\u0080\u00A2,---350 -300 -250 -200 -150 -100 -50^0^60 ^100MOMENT AT JOINT (kNm)8Ez- 6 -w2\u00E2\u0080\u00A2 4-o_cn2-w0_ 0- Ji-cn -2-40099Figure 7.14 Steel plate movements of Specimen 2. (dotted line - exterior, solid line interior)7.6.3.2 SPECIMEN 3,4 and 5The steel plates displaced much more in the bolted specimens, than in the welded joint.This can be observed on the interior part of the joint where the movement of the platesunder peak loads reached 3 mm ( Specimen 3 ), comparing with less than 1 mm for the weldedconnection.On the exterior part the similar behavior as for the welded joint was observed. Thesteel plate attached to the rafter moved much more than the one attached to the beam,where displacement and moment relation remained linear during all loadings including thepeak ones. For the relations of the steel plate movement and the bending moment at jointsee Figures 7.15, 7.16.10EEF- 7.5-zw2w^5-5a_ 2.5-O0_Lu-1 -2.5-w1 00---- : --------5-400 -350 -300 -250 -200 -150 -100 -50^0^50 100MOMENT AT JOINT (kNm)Figure 7.15 Steel plate movements of Specimen 4. (dotted line - exterior, solid line interior)10 ^EE7.5-zw2Lu^50- 2.5 -Cl)Lu 0n_Lu--I -2.5Lu-Co -5-400 -350 -300 -250 -200 -150 -100 -50^ 50 100MOMENT AT JOINT (kNm)Figure 7.16 Steel plate movements of Specimen 5. (dotted line - exterior, solid line interior)1017.6.3.3 SUMMARY1. The analysis of the steel plate movement showed the linear relationship and very smalldeformations during specified and design loads. The re-bar tension failure can be trackedmore precise looking at the movement of the steel plate attached to the rafter, where there-bars failed.2. The steel plate displacement indicates very stiff bond between glulam and embeddedre-bars, which is a desired requirement for the glued-in rod joint.7.7 CONCLUSIONS FROM THE KNEE JOINT TESTS1. The knee joint tests demonstrated a balanced designed glued-in rod knee joint. Thebending moment resistance of the connection was 10% higher than the nominal glulamcross-section capacity.2. Knee joints failed by tension failure in the re-bars in the interior and exterior of theconnection except the preliminary joint #1 and Specimen 5 which failed by wood crushingunder positive moment. The ultimate resistance of Specimen 5, despite different failuremode was in the same range as for the specimens which failed under re-bar tension failure.3. The positive bending moment resistance - 350 kNm, was very consistent for all boltedspecimens and 50 kNm higher than the resistance of the welded joint - 300 kNm.4. The preliminary test showed the need of using a stiffener on the inside part of the con-nection.5. The gap opening analysis showed that under specified loads the opening of the gap onthe exterior of the joint was less than 3.6 mm, which translates to 51 mm deflection in thereal structure. For the ultimate load the deflection caused by the joint rotation will be around215 mm6. The steel plate movement analysis showed the existence of a strong bond between there-bars and the glulam, which allowed less than 1 mm movement of the steel plates.102CHAPTER 8DESIGN OF GLUED-IN RE-BAR JOINTS8.1 INTRODUCTIONBased on the test results and an engineering analysis a proposed, approximate designmethod for glued-in rod joints is presented. Two configurations are considered :1.A symmetrical joint such as would be used in a column to foundation connection, or in asplice joint for a beam, where the moments are resisted by the forces in the steel rods.2. An unsymmetrical knee joint as in a portal frame.8.2 DESIGN STEPSA design of a glued-in rod joint may consist of the following steps :A) STRENGTH LIMIT STATE. STEP 1.^Moment resistance - Mr of the joint. Knowing the external bending momentMf from analysis determine the area of re-bars needed to carry the moment,according to the geometry of the joint.STEP 2.^Bearing capacity of the steel plate. Force perpendicular to grain has to betransferred to glulam by compression perpendicular to grain.STEP 3.^Steel plate resistance. A. Tension resistance of the steel plates: inside plate and straps.B. Bending resistance of the steel plates.C. Tension and bending combined.STEP 5.^Bolts or studs.Calculate the number of bolts and check them for shear.STEP 6.^Welds :^Strap weld.Re-bar to steel plate weld.Fu Fh,103STEP 7.^Shear resistance. Interior stiffener, design for shear.B.) SERVICEABILITY LIMIT STATE. STEP 8.^Check deflections under specified loads.DESIGN ASSUMPTIONS 1.Failure of glued-in rod joint takes place in re-bars - failure mode is bar tension failure.2.Bending moment is carried by tension in re-bars and compression by re-bars in conjunctionwith compression in the glulam.3.Shear resistance is provided by outside straps and inside stiffener.4.Shear force V2 (See Figure 8.1) is assumed small in comparison to Fr and it was neglected.The testing was performed for the angle of re-bar inclination to grain equal to 30degrees. In the design steps a general angle alpha is used.8.3 SPLICE JOINT AND FOUNDATION JOINT8.3.1 STEP 1. - CALCULATE BENDING RESISTANCE OF THE JOINT.During Step 1 of the design process the re-bar area needed to transfer the moment isdetermined.The geometry of the joint is presented on Figure 8.1 .Figure 8.1 Geometry of the splice or foundation joint.Knowing the geometry of the glulam cross-section and external forces : bending104moment and shear force, the number of re-bars needed to transfer the moment can becalculated.To calculate lever arm I we add steel plate thickness to the height of the glulam member.^l = d + t^ (1)where:d - depth of the member.t - thickness of the strap.Force Fm acting in the strap can be calculated from the following formula.MfFM 1Knowing Fm axial force in the re-bars Fr and shear force in the re-bars V1 can beobtained from the following formulas, neglecting V2. (See Figure 8.1):F R =F A,, cos(a)V, = F M x sin(a)^ (4)To determine the required re-bar area, force Fr should be dived by the characteris-tic yielding stress in the re-bars.F RA r:qh =-^vwhere : Fv is a yielding stress, according to CAN/CSA-S16.1-M89After determining the required re-bar area the number of re-bars can be estab-lished. Weldable re-bars must be specified.Shear resistance of the re-bars can now be compared with shear force V1 :17 ,=4)x A r eb X F.where :Fs = 0.66 FvAreb - area of the re-bars.(2)(3)(5)(6)105The moment capacity of the joint can be calculated:M \u00E2\u0080\u0094 OF ,A rebIr^cos(a) (7)where : phi = 0.67 (number used in design of testing specimen)8.3.2 STEP 2 - DETERMINE BEARING CAPACITY OF THE STEEL PLATE.Calculate the dimensions of steel plate needed to carry compression perpendicular tograin.1. Find compression perpendicular to grain :Compression force can be calculated as follows (See Figure 8.1):1 /F c = F ,tan(a) ^1cos(a)For the design purposes the worst case should be checked so the compression perpendicularto grain should be checked for the following force:F c = F m tan(a)^ (9)where : Fm = Mf/12. Assume steel plate dimensions 11, b.3. Find the bearing resistance assuming the stress distribution according to Figure 8.2F \u00E2\u0080\u009E=0.5x(0.51 1 x f cp xb)^(10)Where fcp is obtained from CAN/CSA-089.4. Check if Frc > Fc - if not increase length 11 and go back to 3.1 1fcp11/2(8)Figure 8.2 Assumed compression perpendicular to grain distribution.1068.3.3 STEP 3 - CHECK COMBINED TENSION AND BENDINGRESISTANCE OF STEEL PLATES.A. CHECK STEEL PLATES FOR TENSION.The inside steel plates and straps must be checked for tension. The minimum nettcross-sections should be calculated according to the following formula :Ar^FMnett 0.854) (10)Where : steel properties according to CAN/CSA-S16.1-M89In this research program 12 mm thick steel plate was used.B. CHECK STEEL PLATES FOR BENDING.The strap and the interior steel plate must be checked for the bending moment Mlsee Figure 8.3.Assumption : wood bearing is neglected.MI =F R x R +17 ,x,^ (1 1 )Where :x = /sin(a)x v = / cos(a)M11 1Figure 8.3 Forces acting on the steel plate.Bending resistance Mr should be calculated according to CAN/CSA-S16.1-M89,107considering that, the cross-section consists of two steel plates, re-bars and welds.C. CHECK STEEL PLATES FOR TENSION AND BENDING COMBINED.The steel strap should be checked for tension and bending. Tensile resistance shouldbe calculated using total cross-section ( no holes for bolts should be placed over the firstre-bar row).Resistance to bending and tension should be checked according to the followingformula :FA, + M I <10T r Mr8.3.4 STEP 4 - CHECK BOLTS (STUDS) FOR SHEARIn this step the number of studs needed to carry the force from the strap to the insideplate is calculated. The inside plate has tapped holes, National Fine thread is recommended.The studs should be checked for shear for force - Fm.Shear resistance of the studs, and bearing resistance of the steel plates should bechecked according to CAN/CSA-S16.1-M89.In this testing program 5/8 NF high strength alloy studs were used.8.3.5 STEP 5 - CHECK WELDS.RE-BAR TO STEEL PLATE WELDThe weld between steel plates and re-bars should be calculated for the force Fmaccording to CAN/CSA-S16.1-M89.8.3.6 STEP 6Not applicable.8.3.7 STEP 7Not applicable.1088.4 KNEE JOINT8.4.1 CALCULATION OF THE BENDING MOMENT RESISTANCE OF THEJOINT.1. CALCULATION OF RE-BAR AREA AT THE OUTSIDE PART OF THE JOINT.Figure 8.4 shows a typical geometry of a knee joint. The cut bisects the angle betweencolumn and rafter in to equal parts. The force Fr (tension in re-bars) is assumed to act halfway between the two rows of re-bars, if two rows are used.1Figure 8.4 Forces in the knee joint.Knowing the positive bending moment - Mf, at the knee from gravity load, force Fmcan be determined from the formula :M,F A4 =-7Lwhere : 1 - lever arm according to Figure 8.4Knowing the angle beta the following components can be calculated :FM = F A, x cos(13)F IL=F,xsin(f3)^ (14)(12)(13)The force Fmp travels through the steel plate and then is transferred to the re-bars.109Knowing Fmp axial force in the re-bars and shear force in the re-bars can be obtained fromthe following formulas \u00E2\u0080\u00A2F R =FZxcos(a)V, =^x sin(a)The required area of re-bars can be now found :FRA ir.\" =gb F,where : Fv according to CAN/CSA-S16.1-M89After determining the required area of the re-bars, the number of re-bars is established.Only weldable re-bars must be specified.Shear resistance of the re-bars can now be compared with shear force V1 :V r =4)x A reb xF,^ (6)where :Fs = 0.66 FvAreb - area of the re-bars.Now the positive moment capacity of the joint can be calculated:M = OF\u00E2\u0080\u009EA,b/r cos(a)sin (13)(19)where : phi = 0.67 (as for the steel connections to assure that the connection will not breakbefore the member.)(15)(16)(17)1102. CALCULATION OF THE RE-BAR AREA ON THE INSIDE PART OF THE JOINT.The geometry of the inside area of joint is presented on Figure 8.5.Figure 8.5 Forces on the inside part of the knee joint.Knowing the maximum negative bending moment value from wind or earthquake load,force Fm can be calculated from the following formula.M 1where : 1 - lever arm according to Figure 8.5Knowing the angle gamma the following components can be calculated :Pm = F ,\u00E2\u0080\u009Excos(y)^ (18)FZ=F m xsin(y)^ ( 1 9)The force Fmp travels through the steel plate and then is transferred to the re-bars.Knowing Fmp axial force in the re-bars Fr and shear force in the re-bars V1 can beobtained from the following formulas (See Figure 8.5):F R =FMxcos(ac)^ (3)V, = FM x sin(a) (4)111To determine the required re-bar area, force Fr should be dived by the characteris-tic yielding stress in the re-bars.F ,A req =F 0where : Fv is a yielding stress, according to CAN/CSA-S16.1-M89After determining the required re-bar area the number of re-bars can be estab-lished. Weldable re-bars must be specified.Shear resistance of the re-bars can now be compared with shear force V1 :Vr =4)x Ar.bxFswhere :Fs = 0.66 FvAreb - area of the re-bars.The moment capacity of the joint can be calculated:Mr - cos(a)where : phi = 0.67 (number used in design of testing specimen)8.4.2 STEP 2 - DETERMINE BEARING CAPACITY OF THE STEEL PLATE.Calculate the dimensions of steel plate needed to carry compression perpendicular tograin. Compression perpendicular to grain has to be checked for at two ends of the steelplate :A. At the end of the plate where the re-bars are welded. Compression check same as forthe foundation joint.B. At the peak of the knee joint. Compression perpendicular to grain has to be checkedfor the force Fmc (See Figure 8.4)OF\u00E2\u0080\u009E A reb 1( 5)(6)(7)112A. Check compression perpendicular to grain at the end of the plate where the re-bars arewelded: Compression force can be calculated as follows (See Figure 8.4):V IF c = FL' tan(a) ^ (8)cos(a)For the design purposes the worst case should be checked so the compression perpendicularto grain should be checked for the following force:F,-- F IZtan(a)^ (9)1. Assume steel plate dimensions 11, b.2. Find the bearing resistance assuming the stress distribution according to Figure 8.6F\u00E2\u0080\u009E---- 0.5x (0.5/, x i cp x b)^ (10)3. Check if Frc > Fc - if not increase length 11 and go back to 3.Where fcp is obtained from CAN/CSA-089.1 1i \u00E2\u0080\u00947E1Z]::] I cp11/2Figure 8.6 Assumed compression perpendicular to grain distribution.B. Check compression perpendicular to grain at the peak of the joint. Compression perpendicular to grain at the peak of the joint has to be equal to forceFmc calculated from the following formula :FL=F m x cos((3)^ (13)1. Assume steel plate dimensions 11, b.2. Find the bearing resistance assuming the stress distribution according to Figure 8.7F\u00E2\u0080\u009E= 0.5x (0.5/, xf cp xb)^ (10)1133. Check if Frc > Fmc - if not increase length 11 and go back to 3.Where fcp is obtained from CANICSA-089.1 1Figure 8.7 Assumed compression perpendicular to grain distribution.8.4.3 STEP 3 - CHECK TENSION AND BENDING RESISTANCE OF STEELPLATES.A. CHECK STEEL PLATES FOR TENSION.The inside steel plates and straps must be checked for tension. The minimum nettcross-sections should be calculated according to the following formula :FPA,:t -^In0.850F\u00E2\u0080\u009E ( 10)Where : steel properties according to CAN/CSA-S16.1-M89In this research program 12 mm thick steel plate was used.B. CHECK STEEL PLATES FOR BENDING.The strap and the interior steel plate must be checked for :A. Bending moment M1B. Bending moment M2.A. Check for bending moment Ml. For the location of bending moment M1 see Figure 8.8Assumption : wood bearing is neglected.M, = F ,,, x R A-17 , xWhere :x r = /, sin(a)x t, = /,cos(a)114M111Figure 8.8 Forces acting on the steel plate.Bending resistance Mr should be calculated according to CAN/CSA-S16.1-M89,considering that, the cross-section consists of two steel plates, re-bars and welds.B. Check for bending moment M2. For the location of bending moment M2 see Figure 8.9Assumption : wood bearing is neglected.M2 = F cmx /1^ (12)M2 ,--)1 1Figure 8.9 Forces acting on the steel plate.Bending resistance Mr should be calculated according to CAN/CSA-S16.1-M89,considering that, the cross-section consists of two steel plates, re-bars and welds.115C. CHECK STEEL PLATES FOR TENSION AND BENDING COMBINED.The steel strap should be checked for tension and bending. Tensile resistance shouldbe calculated using total cross-section ( no holes for bolts should be placed over the firstre-bar row).Resistance to bending and tension should be checked according to the following formula :F fi M1 2\u00E2\u0080\u0094+^<1.0T r Mr8.4.4 STEP 4 - CHECK BOLTS FOR SHEARIn this step the number of bolts needed to carry the force from the strap to the steelplate is calculated. The steel plate must have tapped holes. National Fine thread is rec-ommended. Bolts should be checked for shear for force - Fmp.Bolts for shear and steel plate bearing should be checked according to CAN/CSA-S16.1-M89.In this testing program 5/8 NF high strength bolts were used.8.4.5 STEP 5 - CHECK WELDS.WELDING OF RE-BAR TO STEEL PLATE.Weld between steel plates and re-bars should be calculated for force Fmp accordingto CAN/CSA-S16.1-M89.WELDING OF STRAPCheck weld thickness for strap and inside stiffener.Weld should be checked for force Fm according to CAN/CSA-S16.1-M898.4.6 STEP 6 - DESIGN INTERIOR STIFFENER - SHEAR RESISTANCEThe shear resistance of the knee joint is provided by the interior stiffener.Interior stiffener should be checked for shear since the shear resistance provided by thestraps is limited. For the interior stiffener geometry see Figure 8.9.116Figure 8.10 Interior stiffener.Shear force obtained from analysis - Vf should be less then shear resistance VrOA FsF,= 0.66FyThe shear area required can be found as follows :A req \u00E2\u0080\u0094V00.66F,So the height of the stiffener is :A rRq1 : egh = 2twhere t - thickness of stiffener ( use 2 stiffeners for two rows of re-bars )In this testing program t =12 mm lsh =95 mm.The interior stiffener plates should allow direct transfer of the force from one memberto the other for negative moments. Therefore the alignment of the re-bars, should coincidewith the outside edge of the stiffeners, and the plane of the stiffeners. See Figure 8.10(27)(28)(30)1178.4.7 STEP 7 - CHECK DEFLECTIONS UNDER SPECIFIED LOADS.The deflection check of a structure equipped with glued-in rod connections can beperformed using all frame analysis computer programs, which are able to simulate joints asrotational springs. Glued-in rod joints are modelled as rotational springs with known stiff-ness. The average rotational stiffness of the knee joints tested was c= 46900 kNm/rad. Thestiffness was calculated from joint opening data, the rotational change of two members andthe readings under specified load were used.118CONCLUSIONS1. A stiffer connection method has been developed that will enable the industry to designand build statically indetermined structural systems. This in turn will permit the industryto expand its' markets to include the structures which they could not build previously dueto a lack of connection methods.This connection is based upon the use of glued-in reinforcing bars inserted at a 30degree angle to the length of the glulam member, thus engaging the whole cross-sectionto participate in the transfer of the forces.2. The required minimum embedment length was established but it is recommended thatthe reinforcing bars be inserted so some of them cross in order to improve the shear car-tying capacity of the glulam in addition to forming the anchor for transferring tension andcompression forces.3. A simple manufacturing method was derived to produce test specimens but a moresophisticated production methods should be developed for commercial manufacturing.4. The type of structural connections used in this experiment includes connections for :beam splice, column foundation, knee joints for frames.5. Suggested design methods for these connections are outlined in the thesis.6. The appearance of the connection was of primary concern. By attaching the reinfor-119cing bars to the underside of a dapped in steel plate the appearance was found to beacceptable.7. The connection is designed so the failure takes place as ductile steel failures withoutaffecting the timber. The timber connection thus has a reliability equal to that of a steeljoint.FUTURE RESEARCH The tests described in this thesis have been limited to short term tests. Howeversince the failure takes place as yielding in the reinforcing bars it could be argued that theexperience with steel connections can be transferred directly.It would never the less be prudent to conduct additional tests aimed at investigatethe effect of varying temperature and moisture content might have on the interfacebetween the glue and the wood.Glued-in rods with the rods placed parallel to grain have been used in practice inEurope for more than 25 years without apparent trouble, but despite this it would be pru-dent to conduct additional experiments.The creep behaviour of the joint would also be useful information.120SUMMARY OF TEST DATATest type ConnectiontypeNumber ofbarsbar dia.(mm)FailuremodeStressin barsat failure(MPa)Max.forcein steelplates(kN)Max.bendingmoment(kNm)Pull-out Preliminary 1x20 Re-bartensionfailure495 171Pull-out Pre-welded 1x20 Re-bartensionfailure517 189Pull-out Pre-welded 2x20 Re-bartensionfailure478 345BeamtestPre-welded 2x20 Re-bartensionfailure344 237 92Rackingtests:Spec-1 Pre-welded 2x20 Re-bartensionfailure336 231 90Spec-2 Pre-welded 3x20 Re-bartensionfailure378 341 175Spec-3 Pre-welded 4x20 Re-bartensionfailure325 390 200Spec-4 Pre-welded 4x20 Shearstudfailure398 478 245121Test type ConnectiontypeNumber ofbarsbar dia.FailuremodeStressin barsat failure(MPa)Max.forcein steelplates(kN)Max.bendingmoment(kNm)Kneejointtests:Spec-1 Pre-welded 4x20 Compr.perp.failure276 381 148Spec-2 Pre-welded 4x20 Re-bartensionfailure342 472 302Spec-3 Pre-welded 4x20 Re-bartensionfailure395 545 349Spec-4 Pre-welded 4x20 Re-bartensionfailure392 541 346Spec-5 Pre-welded 4x20 Compr.perp.failure405 559 358122BibliographyBaird, JA, Ozelton, E.C.,1990Timber designers' manual, Collins, London, U.K.Buchanan, A.H., Moss, P.J., Townsend, P.K., 1990Reinforced bars epoxy bonded in glue laminated timber, Proceedings of the 1990 Inter-national Conference on Timber Engineering, Tokyo, Japan.Canadian Institute of Steel Construction, 1992Handbook of steel construction, Canadian Institute of Steel Construction, Willowdale,Ontario, Canada.Canadian Wood Council, 1990Wood designers manual, Canadian Wood Council, Ottawa, Ontario, Canada.Crews, K.J., 1991Joints and connections: Timber engineered structures, Australian Forest Industries Jour-nal, November 1991.Fairweather, R.H., Buchanan, A.H., 1992Beam column connections for multi-storey timber buildings, Research report, Depart-ment of Civil Engineering, University of Canterbury, Christchurch, New Zealand.Foschi, R.O., Folz, B.R., Yao, F.Z., 1990Reliability based design of wood structures, Report No. 34, Department of Civil Engi-neering, University of British Columbia, Vancouver, Canada.123Keenan, F.J.,1986Limit states design of wood structures, Morrison Hershfield Ltd, Toronto, Ontario, Can-ada.Madsen, B., 1992Structural behaviour of timber, Timber Engineering Ltd., Vancouver, BC,Canada.McIntosh, K.A., 1989From theory to reality - 30 years in glulam manufacture, Proceedingsof the second pacific timber engineering conference, Auckland, New Zealand.Riberholt, H.,1980Steel bolts glued into glulam, Proceedings, Meeting of IUFRO Wood EngineeringGroup, Oxford, U.K.Riberholt, H.,1986Glued bolts in glulam, Department of Structural Engineering, Technical University ofDenmark, Series R, Number 210.Riberholt, H.,1988Glued bolts in glulam. Proposals for CIB code, Proceedings, CIB Meeting 21, Parksville,BC, Canada.Townsend, P.K., Buchanan, A.H., 1990Steel dowels epoxy bonded in glue laminated timber, Research report, Department ofCivil Engineering, University of Canterbury, Christchurch, New Zealand.124Turkowskij, S., 1991Prefabricated joints of timber structures on inclined glued-in bars, Proceedings of the1991 International Timber Engineering Conference, Volume 3, Trada, London, U.K.Turkowskij, S., Lukyanow, E.I., Pogoreltsew, A.A., 1991Use of glued-in bars for reinforcement of wood structures, Proceedings of the 1991 Inter-national Timber Engineering Conference, Volume 3, Trada, London, U.K.125APPENDIX A.FRAME ANALYSIS1.1 DESIGN OF TWO HINGE PORTAL FRAME.A two hinge glulam frame of the geometry shown on Figure A.1 was designed for anindustrial building in Vancouver. The following sections present the loads, and the designof decking and purlins, as well as design of the frame.12.0 PiFigure Al. Two hinge glulam frame.1.2 SNOW LOADThe snow load was calculated for Vancouver ( Granville and 41 ).The specified loading, S, due to snow accumulation on a roof isS = S s (C,C,\u00E2\u0080\u009EC s C a )+ S rSs - ground snow load in kPa^Ss = 2.5 kPaSr - associated rain load in kPa^Sr = 0.3 kPaCb - basic roof snow load factor^Cb = 0.8Cw - wind exposure factor^Cw = 1.0Cs - slope factor^ Cs = 1.0 when roof slope alpha < 15 degalpha = 8 degCa - accumulation factor^Ca = 1.0126S = 2.5 x 0.8 x 1.0x 1.0x 1.0 + 0.3S = 2.3 kPaSpacing range - 3.0 to 6.0 mSmax is the snow load for the maximum spacing - 6.0 m.k N )S max = 2.3 x 6 = 13.8(-InThe code requires checking of two loading modes.1.Full load on entire span.2.Full load on half of the span, and half load on the other part.Smax2^kN- 6.9 \u00E2\u0080\u0094 = 6.9 m mm1.3 WIND LOADEXTERNAL WIND PRESSUREThe specified external pressure or suction due to wind is:p= qXC.XC g XC pp - the specified external pressure acting statically and in the direction normal to the surface.q - the reference velocity pressure ( probability of being exceeded in any one year in 30 forthe design of structural membersq= 0.55 kPaCe - exposure factor^ Ce = 0.9 for height < 6 mCg - the gust effect factorCp - the external pressure coefficientCp Cg - used together for low buildingsDirection of pressureRoof slopedeg1front2roof3roof4back8 0.84 -1.3 -0.78 -0.65Table 2.1 External wind pressures.127INTERNAL WIND PRESSUREThe following three cases have been checked:CASE 1. Openings in wall at which wind is blowing.CASE 2. Openings in wall opposite the wall wind is blowing.CASE 3. Wind parallel to ridge.Table 2.2 presents internal wind pressure coefficients.Direction of pressureCASE # 1front2roof3roof4back1 -0.7 -0.7 -0.7 -0.72 0.5 0.5 0.5 0.53 0.7 0.7 0.7 0.7Table 2.2 Internal wind pressures.TOTAL WIND PRESSURESThe following table presents the total wind pressures.Direction of pressureTotal wind pressures (kPa)CASE # 1front2roof3roof4back1 0.07 -0.99 -0.73 -0.672 0.66 -0.64 -0.39 -0.073 0.76 -0.64 -0.39 -0.02Table 2.3 Total wind pressures.Figure 2.6 illustrates the directions of wind pressures.Figure 2.6 Directions of wind pressures. (Positive values)1281.4 EARTHQUAKE LOADThe minimum lateral seismic force, V, shall be calculated in accordance with the fol-lowing formula.v eV =(-1-7)UWhere :U = 0.6R - force modification factorR = 2.0 ( moment resisting wood space frame with ductile connections )In the present code it is assumed that the commercially available types of fasteners areused. The force modification factor R should reach 3, depending on the actual stiffness, forthe glued-in rod joints. The actual stiffness is discussed in the chapters describing exper-imental results. ( Chapter 8 ).Presently used connection methods do not provide rigid joints. Assuming rigidity andfailure in steel for glued in rod joint this structure should be treated as a steel structure.The equivalent lateral seismic force, representing elastic response, Ve :V.--uxsx/xFxu,v - zonal velocity ratiov = 2.0 for VancouverZa = Zv = 4.0Za/Zv = 1.0s - seismic response factorT - fundamental period0.09 x h\u00E2\u0080\u009ET =^VD:Ds = 12.0 m hn = 3.6 m0.09 x 3.6T -^ _ 0.093sV12129From Table 4.1.9.A ( National Building Code ) S = 3.0I - seismic importance factor I = 1.5F - foundation factor - assume F = 1.5W - the weight of the structure1.1-) = U.) roof + (I) purlins^U.) frame + tl) snowUl roof = 0.27 x 6.0 x 12.0= 19.4kNtu p,,ns = 0.24 x 6.0x 7 = 10.1 kNtoiran = (2 x 3.6+ 12.0)x 0.63 = 12.1kNti)\u00E2\u0080\u009E\u00E2\u0080\u009E.= 2.3x 6x 12x 0.25= 41.4kNw = 83.0kNV,= 0.2x 3x 1.0x 1.5x 83.0= 74.7kNV = LIY) x 0.6 = 22.4kNForce V = 22.4 kN is applied at node 2 during finite element analysis1.5 DESIGN OF MEMBERS1.5.1 MATERIALSThe following materials were used for design :1. Commercial grade decking.2. Purlins - glulam, D.Fir-L, 20f-E.3. Frame - glulam, D.Fir-L, 20f-E.1.5.2 DECKING DESIGN- Roof deck- Specified Dead Load =^- wood decking self weight^= 0.13 kPa- felt^ = 0.14 kPaTotal load = 0.27 kPa130- Specified Live Load =^- snow load^= 2.3 kPa- Purlin spacing =^2.0 m- Dry service condition- untreated- deflection limitation of L/180 based on total specified load- use controlled random pattern western red cedar- commercial gradeFactored loading =^Wf = 1.25x0.27 + 1.5x2.3 = 3.8 kPaTotal specified loading =^W = 0.27 + 23 = 2.6 kPaFrom Decking Selection Tables ( Wood Design Manual ) select 38 mm thicknessWfr = 11.0 kPa > 3.8 kPa O.K.Wdr = 3.95 kPa > 2.6 kPa O.K.1.5.3 PURLIN DESIGNPurlin spacing =^2.0 mPurlin span = 6.0 mSpecified Dead Load =^0.27 kPa ( decking + felt )Specified Live Load =^2.3 kPa ( snow )Standard load durationDry service conditionUntreatedCALCULATIONTributary area =^2.0 x 6.0 = 12.0 m ^ 2Factored load = 1.25x0.27 + 1.5x2.3 = 3.8 kPaSpecified load =^0.27 + 23 = 2.6 kPaSelf weight = 5.3x0.13x0.342 = 0.24 kN/m - try 130x342 glulam131Wf = 3.8x2.0 + 0.24x1.25 = 7.9 kN/mW = 2.6x2 + 0.24 = 5.4 kN/mW1 = 2.3x2.0 = 4.6 kN/mM =w I L 2 7.9 x 6.02 - 35.6kNmI 8 -^8tu f x L 7.9 x 6.0Vf -2 -^ 23.7 kN2For L/180 deflection limit based on total load:^E. s l ,,,,,= 18053to8L 3^Sx 5.4 x 600034 Nun 2 = 180 x^ - 2734 x 109Nrnm2384For L/360 deflection limit based on live load:E sl Reqd= 3605w1.3 Nrnrn 2 = 360 x 5 x 4.6 x 60003- 4656 X 10 9 Nmm 2^384^ 384From beam selection tables for glulamuse : cross-section 130 x 342 mm, D.Fir-L, 20f-EMr = 58.4 kNmVr = 53.4 kNEsI = 5370* 10 ^ 9 Nmm ^ 21.5.4 TOTAL LOADSSeveral load combinations were checked during the analysis. See 1.11.4.2.According to the CAN/CSA-0861-M89 Clause 4.2.4.1 the factored load combination shallbe taken as:aDD + yip(a L L + a Q Q+ aTT)where:D - dead load due to load of membersL - live load due to snow and rainQ - live load due to wind and earthquake132T - loads due to contractions or expansion caused by temperature changesLOAD FACTORSa,= 1.25, v 0.85E .cases.of. stress . reversalaL= 1 . 5a Q = 1.51 or winda = 1.0 f or earthquake= 1.25= 0.7 when .two. L ,Q . act1.5.5 LOADING ON THE FRAMEAll load values presented in this section are unfactored.1. ROOF DEAD LOADWs = 0.27 + 0.1 = 0.37 kPaWsp = 2 x 0.37 = 0.74 kN/mVp = 0.74 x 6 = 4.4 kN - reaction from 2 purlins2. SNOW LOAD Ws= 2.3 kPaWsp = 2.3 x 2 = 4.6 IN/mVp = 4.6 x 6 = 27.6 kN - reaction from 2 purlins3. WIND LOAD Wind loads :CASE 1:P1 = 0.42 kN/m P2 = -5.94 kN/m P3 = -4.38 kN/m P4 = -4.02 kN/mCASE 2:P1 = 3.96 kN/m P2 = -3.84 kN/m P3 = -2.34 kN/m P4 = -0.42 kN/mCASE 3:P1 = 4.56 kN/m P2 = -3.84 INN P3 = -2.34 kN/m P4 = -0.12 kN/m1334. EARTHQUAKE LOADV = 22.4 ( earthquake load applied at node 2 )1.5.5.1 LOAD CASES1.Frame self weight2. Roof dead load3. Full snow load4. Earthquake load5. Half snow load6. Wind load - case 17. Wind load - case 28. Wind load - case 31.5.5.2 LOAD COMBINATIONS1. (Roof dead load + frame self weight) x 1.25 + (full snow load) x 1.52. (Roof dead load + frame self weight) x 0.85 + (wind load) x 1.53. (Roof dead load + frame self weight) x 0.85 + (earthquake load) x 1.04. (Roof dead load + frame self weight) x 0.85 + ((wind load - case 1) x 1.5 + (half snowload) x 1.5) x 0.71.6 MATERIAL PROPERTIESModulus of elasticity for 20 f-EX glulam according to the code is E = 12400 MPa.Shear modulus G = 852 MPa. Glulam density = 5.3 kN/m ^ 3.1.7 ANALYSIS RESULTSThe following table presents the internal forces obtained during analysis at knee jointfor all load combinations.134Load combination Bending momentkNmShear forcekNAxial forcekN1. Full snow 219.5 111.4 147.92. Wind 51.6 26.5 18.53. Earthquake -22.6 3.3 22.94. Wind + half snow 76.6 24.4 25.3Table 2.4 Internal forces obtained during analysis.The maximum bending moment occurred during full snow load - positive momentresulting in tension on outside of the frame, and during earthquake load - negative bendingmoment resulting in tension on the inside part of the knee joint area. Negative momentfrom earthquake load was around 10% of the positive moment from full snow load. Maximumshear and axial forces were found during full snow load.1.7.1 DESIGN OF CROSS -SECTIONTry glulam 175 x 646 mm 20f-EX, D.Fir-L.The following design was done according to CAN/CSA-086.1-M891.7.1.1 BENDING MOMENT RESISTANCEThe factored bending moment resistance shall be taken as :M r = 4F b SK L K Xwhere :(0=0.9Fb=J b(KDKHKSbKT)I( D =1;K H =1;K sb =1;K T =1;K L =1;K x =1b=25.6Mpa175x 646 2S= = 12.2x 10 6 nn 36135M r = 0.9 x 25.6 x 12.2x 10 6 x 1 x 1 = 280.4kNm > M f =219.5kNm.O.K.1.7.1.2 SHEAR RESISTANCEThe factored shear resistance shall be taken as :2A17,=OF\u00E2\u0080\u009E-3 K,where :(1) = 0.9F f u(KDKNKst,KT)f = 2.0M paK^1;K H =1:K sv = 1;K T =1K N =1A= 175x 646 = 11.3x 10 4 mm 2= 0.9 x 2.0x 11.3x 10 4 + 1 = 135.7kN > V f =111.4kN.O.K.1.7.1.3 COMPRESSIVE RESISTANCEThe factored compressive resistance parallel to grain shall be taken as:P r = OF c AK,where :0=0.9F c = f c (K D K HKscKT)f c =20.4MpaKD= 1;K H = 1:K se = 1;K T =1A= 175x 646 = 11.3x 10 4 mm 23.6^\u00E2\u0080\u0094 20.60.1750.76 x 0.87x 12400x 1 x 1\u00E2\u0080\u0094 20.020.41361^0.87x12400x1x1K -0.622 x 20.6 2^20.413 ,.= 0.9 x 20.4 x 11.3x 10 4 x 0.62 = 1293.3kN > P f = 146.9kN .0.K.1.7.1.4 RESISTANCE TO COMBINED BENDING MOMENT AND AXIAL LOADMembers subject to combined bending and axial loads shall be designed to satisfy thefollowing interaction equationP f A4,-+ ----- < 180Pr Mr147.9^+ 219.5 - 0.11 + 0.78 = 0.89 < 1.0.0.K.1293.3 280.4Use 175 x 646 mm glulam 20f-EX D.Fir-L"@en . "Thesis/Dissertation"@en . "1993-11"@en . "10.14288/1.0050496"@en . "eng"@en . "Civil Engineering"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Glued-in re-bar connection"@en . "Text"@en . "http://hdl.handle.net/2429/1972"@en .