BEHAVIOUR OF FIBRE REINFORCED POLYMER (FRP) STIRRUPS AS SHEAR REINFORCEMENT FOR CONCRETE STRUCTURES RYAN DAVID MORPHY

Transcription

1 BEHAVIOUR OF FIBRE REINFORCED POLYMER (FRP) STIRRUPS AS SHEAR REINFORCEMENT FOR CONCRETE STRUCTURES by RYAN DAVID MORPHY A Thesis Pesented to the Faculty of Gaduate Studies in Patial Fulfillment of the Requiements fo the Degee of MASTER OF SCIENCE Stuctual Engineeing Division Depatment of Civil and Geological Engineeing Univesity of Manitoba Winnipeg, Manitoba JUNE, 999

2 :" i' ACKNOWLEDGMENTS I would like to thank D. Sami Rizkalla fo his enthusiasm fo eseach and fo his patient guidance and encouagement though the expeimental phase and the eventual completion of this thesis. The knowledge, expeience, and helping hand of D. Emile Shehata also deseves a special thank you. The help of all of the staff and students of the R. W. McQuade Stuctues Laboatoy at the Univesity of Manitoba thoughout the constuction and testing phases was geatly appeciated and they all deseve a sincee thanks. This thesis would not have been possible wee it not fo the suppot povided by the Canadian Netwok of Centes of Excellence on Intelligent Sensing fo Innovative Stuctues (ISIS Canada). This assistance is gatefully acknowledged. ",!,I,\i'l' I Finally, the unwaveing suppot, encouagement, and assistance of my wife Ewa Mophy, my paents Dave and Joan Mophy, and my bothe Kyle has not gone unnoticed and unappeciated. To you and to all fiends and family who have lent a helping hand o passed on an encouaging wod, this thesis is dedicated. ii

3 ABSTRACT The coosion of steel einfocement in einfoced concete stuctues has lead to widespead deteioation of many buildings and bidges that ae now in need of costly epais. Due to thei location as the outemost laye of einfocement within a concete stuctue, steel stiups ae the fist to coode. Fibe einfoced polymes (FRP) povide an altenative that does not coode and has many othe benefits such as a high stengthto-weight atio. Howeve, the inheent disadvantages of FRP in cetain loading situations ae appaent with FRP stiups and must be accounted fo in design. Loading tansvesely to the fibes can cause significant losses in stiup capacity. By bending the FRP to fom an end anchoage fo the stiup within the concete, the bent potion of the ba is subjected to stesses pependicula to the diection of the fibes and will expeience stength losses. Due to the diagonal natue of shea cacks, the vetical stiups ae subjected to stesses pependicula to the fibe diection as the cack widens. This again leads to potential stength losses. This thesis pesents the expeimental pogam and esults of 3 panel specimen tests completed to detemine the losses of stiup capacity as elated to the bend effect and the inclined cack effect, detemined though multiple vaiables. These esults ae analyzed to examine the effect of each vaiable on the stength of the stiups. Based on these esults, ecommendations ae made and design guidelines ae pesented fo the use of FRP as shea einfocement. iii

4 TABLE OF CONTENTS ACKNOWLEDGEMENTS ABSTRACT TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES CHAPTER : INTRODUCTION. Geneal.2 Reseach Significance.3 Objectives and Scope CHAPTER 2: LITERATURE REVIEW 2. Geneal FRP Chaacteistics 2.. FRP Definition 2..2 FRP Reinfocements Aamid FRP Reinfocement Glass FRP Reinfocement Cabon FRP Reinfocement 2..3 Behaviou of Concete Membes Reinfoced with FRP 2.2 FRP Shea Reinfocement iv iv vii viii

5 -( 2.2. Classification and Fabication offrp Stiups Bend Effect Inclined Cack Effect CHAPTER 3: EXPERIMENTAL PROGRAM 3. Geneal 3.2 Mateials 3.2. Leadline Cabon Fibe Composite Cable C-BAR Steel Concete 3.3 Phase - Bend Specimens 3.3. Specimen Design & Fabication Testing Set-up Instumentation Test Pocedue 3.4 Phase II - Inclined Cack Effect 3.4. Specimen Design & Fabication Test Set-up Instumentation Test Pocedue v

6 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4. Geneal Phase I - Bend Capacity Leadline Stiups CFCC Stiups C-BAR Stiups Steel Stiups Effect of Bend Radius on Bend Capacity Effect of Embedment Length on Bend Capacity Effect of Stiup Anchoage on Bend Capacity Effect of Tail Length on Bend Capacity PHASE n - Capacity at Inclined Cack 4.7. Test Results Analysis of the Results "Bonded" Model Steel Specimens FRP Specimens "Tuss" Model 4.8 Related Studies CHAPTER 5: SUMMARY AND CONCLUSIONS vi

7 Table 2- Table 3- Table 3-2 Table 3-3(a) Table 3-3(b) Table 3-3( c) Table 3-4 Table 3-5 Table 3-6 Table 4- Table 4-2 (a) Table 4-2 (b) Table 4-2 (c) Table 4-3 Table 4-4 Table 4-5 LIST OF TABLES Mateial popeties of vaious FRP fibes 2-2 Mateial popeties offrp and steel stiups 3-2 Leadline specimens 3- CFCC x7 5 mm diamete specimens 3-2 CFCC single wie 5 mm diamete specimens 3-2 CFCC x7 7.5 mm diamete specimens 3-3 C-BAR specimens 3-4 Steel specimens 3-4 Inclined Cack Specimens 3-20 Bend test esults: Leadline 4-3 Bend test esults: CFCC 5mm 7-wie 4-4 Bend test esults: CFCC 5mm single wie 4-4 Bend test esults: CFCC 7.5mm 7-wie 4-5 Bend test esults: C-BAR 4-6 Bend test esults: Steel 4-6 Inclined cack specimen test esults 4-3 vii

8 l LIST OF FIGURES Figue 2- Linea stess-stain behaviou of FRP bas Figue 2-2 Pultusion pocess Figue 2-3 CFCC bas in coss-section (Tokyo Rope Manufactuing Company, 993) Figue 2-4 Vaious stiup configuations Figue 2-5 Micoscopic bent fibe (Hull, 98) Figue 2-6 Stesses at the location of the bend Figue 2-7 Bent ba test specimen of Mauyama et al (993) Figue 2-8 Bent ba test specimen ofehsani et al (995) Figue 2-9 Secant modulus used by Ehsani et al (995) Figue 2-0 Stiup bend effect test specimen of Cuie et al (993) Figue 2- Poposed equation fo bent capacity fom Nakamua & Higai (995) Figue 2-2 Bend capacity test specimen ofueda (995) Figue 2-3 Diagam of kinking behaviou Figue 2-4(a) Angled ba test specimen of Mauyama et al (989) Figue 2-4(b) Angled ba test specimen of Mauyama et al (989) Figue 2-5 Combined tension and shea test specimen of Kanematsu & Ueda (993995) Figue 3- Stiup appeaance viii

9 Figue 3-2 Figue 3-3 Figue 3-4 Figue 3-5 Figue 3-6 Figue 3-7 Figue 3-8 Figue 3-9 Figue 3-0 Figue 4- Figue 4-2 Figue 4-3 Figue 4-4 Figue 4-5 Figue 4-6 Figue 4-7 Figue 4-8 Figue 4-9 Leadline Stiup configuation 3-4 Relationship of stiups to shea cacks poducing the bend effect 3-7 Stesses on stiup at bend location 3-7 Bend specimen configuation 3-9 Bend test setup 3-6 Bend test photogaph 3-6 Inclined cack specimen configuation 3-9 Inclined cack specimen test setup 3-2 Inclined cack test photogaph 3-23 Failed Leadline bend specimens 4-8 Failed CFCC bend specimens 4-0 C-BAR "waving" impefection 4-3 C-BAR specimen failue along staight potion of stiup due to defect 4-4 Failed C-BAR specimen at bend location 4-5 Effect of bend adius on stiup capacity with equations poposed by JSCE (997) 4-8 Effect of embedment length, ld, on CFCC stiup capacity 4-20 Effect of embedment length, ld, on Leadline stiup capacity 4-2 Effect of embedment length, ld, on C-BAR stiup capacity 4-2 ix

10 ( Figue 4-0 Effect of tail length, ld*, on CFCC stiup capacity Figue 4-(a) Typical test plot of Load vs. PI gauge eadings (Leadline - 35 ) Figue 4- (b) Typical test plot of Load vs. stain gauges (Leadline - 35 ) Figue 4-2 Failed C-BAR stiups Figue 4-3 Failed Leadline stiups Figue 4-4 Load vs. Cack Width fo Leadline Figue 4-5 Load vs. Cack Width fo C-BAR Figue 4-6 (a) Load Tansfe Mechanism fo the "Bonded" Model Figue 4-6 (b) Load Tansfe Mechanism fo the "Tuss" Model Figue 4-7 Effect of inclined cack on steel stiup capacity Figue 4-8 Effect of inclined cack on Leadline stiup capacity Figue 4-9 Effect of inclined cack on C-BAR stiup capacity Figue 4-20 Poposed equations fo FRP mateial 4-40 Figue 4-2 Measued ultimate stain in Leadline and C-BAR specimens 4-4 x

11 :''If' ( :. Geneal CHAPTER : INTRODUCTION ( ( E ',: : Advanced composite mateials in the fom of Fibe Reinfoced Polymes, FRP, ae extensively used in the automobile industy, aeospace industy, defence, and ecently in a vaiety of ways in civil engineeing stuctues. Many concete stuctues ae deteioating due to coosion of the steel einfocement, and advanced composite mateials ae quickly gaining populaity fo use in the constuction of new stuctues and the epai and stengthening of existing stuctues. Stuctual enginees have a special inteest in using Fibe Reinfoced Polymes fo einfocing stuctues due to thei many favouable chaacteistics. In compaison to steel, FRPs exhibit high tensile stength, do not coode, ae electomagnetically neutal, possess a high stength-to-weight atio, high impact esistance and ae lightweight. Fibe Reinfoced Polymes povide a viable altenative to taditional steel einfocement in concete stuctues..2 Reseach Significance Stiups used fo shea einfocement in concete stuctual membes ae nomally located close to the suface of the concete than flexual einfocement. Due to the minimum

12 l CHAPTER : INTRODUCTION -2 concete cove povided, the stiups theefoe ae moe susceptible to sevee envionmental composite action with the slab cast at a late stage. Duing this peiod, coosion could be sevee. FRPs have ecently been intoduced as einfocement to ovecome the poblems elated to the coosion of steel einfocement and the consequent deteioation of concete stuctues. The eseach into the use of FRP as longitudinal einfocement to date is much effects. In some cases, stiups ae exposed above the top suface of the beam to povide, '-- l : l moe extensive than fo FRP as shea einfocement. FRP fo shea einfocement has not yet been fully exploed and the cuent available data and knowledge of behaviou is not sufficient to povide geneal design ecommendations and guidelines to enginees. Due to the diagonal oientation of shea cacks, the induced tensile foces ae typically oiented at an angle with espect to the stiups. Consequently, the tensile stength of the stiups in the diection of the fibes cannot be fully developed (Mauyama et ai., 989). Bending of FRP bas into stiups to develop sufficient anchoage could also lead to a significant eduction of the capacity (Mauyama et ai., 993; Miyata et al., 989; Cuie et ai., 994; Ehsani et al. 995)..3 Objectives and Scope The main objective of this thesis is to povide design guidelines fo the use of FRP as stiups fo the shea einfocement of concete stuctues. To accomplish this, an

13 ..' t_ ( : i ; " L CHAPTER : INTRODUCTION -3 expeimental pogam was undetaken to evaluate the effect of bending of the FRP bas and the oientation of shea cacks at an angle to the diection of the fibes. In this pogam, single stiup specimens wee designed to simulate the behaviou mechanism of stiups in concete beams. The fist of two phases in the expeimental pogam examines the effect of the bend on the stength of the stiups. In this phase the paametes include the type of mateial, ba diamete, stiup anchoage, bend adius, tail length and embedment length. The second phase of the expeimental pogam investigates the effect of the cack oientation with espect to the stiup fibe diection. The paametes fo this phase ae the mateial type and the angle between the stiup diection and an initiated cack. In both phases the esults and behaviou of FRP stiups ae compaed to steel stiups typically used as shea einfocement. Chapte 2 exploes the liteatue on FRP mateial and the eseach wok on poblems with the use of FRP as shea einfocement. Chapte 3 descibes the expeimental pogam by including desciptions of the specimens and test methodology. Chapte 4 analyzes and discusses the esults of the two phases of the expeimental pogam. Conclusions and ecommendations fom the eseach pogam ae summaized in Chapte 5.

14 : CHAPTER 2: LITERATURE REVIEW 2. Geneal FRP Chaacteistics 2.. FRP Definition Fibe einfoced polymes (FRP) ae composite mateials that consist of load-canying fibes encased in a esin matix. Fibes occupy the lagest volume faction in composite laminates and ae esponsible fo caying the load acting on the composite (Mallick, 993). The coect selection of fibe volumes, type, and oientation of the fibes influences the specific gavity, tensile stength, modulus of elasticity, compession stength, fatigue stength, failue mechanisms, electical and themal conductivity, and cetainly the cost of the composite poduct. Thee ae thee main types of fibes used in FRP mateial. These include aamid, glass, and cabon fibes. The chaacteistics of these fibes ae shown in Figue 2- based on the epesentative values given in Table 2-. In geneal, all the fibes exhibit a linea behaviou in tension up to factue in contast to steel with its clealy defined yield. Most fibes ae chaacteised by a faily high elastic modulus, low ceep and elaxation, stability at nomal ambient tempeatues and high tensile stength. The fibes ae continuous and the diamete is in the ange of 4 to 0 micometes (Mufti et ai., 99). Individual fibes ae then bundled togethe ceating a stonge oveall composite (Dolan et ai., 993).

15 CHAPTER 2: LITERATURE REVIEW 2-2 Table 2- Mateial popeties of vaious FRP fibes Mateial E-glass S-glass High-Modulus Cabon High-Stength Cabon Cabon Kevla49 Density E (GPa) at (GPa) et(%) (k2/m3) a 0-6C l " III " " " " " " " " " " " " " " " ".. " III " " "...,,"" ~ " " " " " "... " " ".. " ".. " " " " " " " " " " " " " " " " "... " " " " " " " 0.5 'MiJdSteel.5 Stain (%) Figue 2- Linea stess-stain behaviou of FRP bas

16 l l l t L CHAPTER 2: LITERATURE REVIEW 2-3 The esin matix is equied to distibute the load within the coss section though the laminate shea mechanism. It also povides a baie against an advese envionment, and potects the suface of the fibes fom mechanical abasion (Mallick, 993). The matix is not equied to withstand load in the composite and has weak mechanical chaacteistics. Matix behaviou is time-dependent, and is a function of the ate and fequency of load applied and the ambient tempeatue (Mufti, 99). The pesence of the matix enhances the esistance to buckling of the composite by suounding and inteacting with the fibes. The matix must be chosen to complement the fibes being used and to povide adequate potection and sepaation of the fibes to enhance the oveall behaviou of the composite. The themoset polyme and themoplastic matices ae the most common types used fo civil engineeing applications (Abdelahman and Rizkalla, 994) FRP Reinfocements FRP einfocement fo concete stuctues can be in the fom of ods o opes, with the advantages of (Abdelahman and Rizkalla, 994):. A tensile stength-to-weight atio in the ange of six to ten times geate than steel. 2. Excellent coosion esistance. 3. Electomagnetic neutality. 4. Low axial coefficient of themal expansion, especially fo cabon fibe einfoced polyme tendons.

17 ' ' J ''''',! CHAPTER 2: LITERATURE REVIEW A weight of one-seventh to one-fifth that of steel einfocement of equivalent diamete. The disadvantages of fibe einfoced polyme ods may include:. A cost two to ten times highe than steel. 2. Low stain at factue. 3. Low modulus of elasticity compaed to steel. 4. Low stength pependicula to the fibes. 5. Low compessive stength. 6. Some fibes susceptible to ulta violet adiation damage. 7. Some fibes absob wate leading to deteioation. Despite the bugeoning use offrp in othe fields, its application in stuctual engineeing is not as widespead. This may be due to one o a combination of the following:. Absence of codes and specifications. 2. Lack of designe confidence due to little elevant expeience and education. 3. Incomplete undestanding of mateial popeties and lack of data egading long tem behaviou. The initial costs of Fibe Reinfoced Polymes act as a deteent in cetain applications. Howeve, once maintenance and othe factos ove the life of a stuctue ae consideed in

18 CHAPTER 2: LITERATURE REVIEW 2-5 a diect compaison to steel, the cost can be justified. The use of steel incus the cost of i cathodic potection and the futue costs elated to the shote life span of steel which is geatly affected by coosion. With the longe life offrp in compaison to steel, the highe initial cost can be less significant in the long tem (Eki and Rizkalla, 994) Aamid FRP Reinfocement Aamid Fibe Reinfoced Polyme (AFRP) tendons ae poduced in a vaiety of shapes including baided, spial wound and ectangula ods. The fibes have a highe modulus of elasticity and lowe density than glass fibes but have a lowe compessive stength, and high moistue absoption (Mallick, 993). Thee types of fibes ae Kevla, Twaon, and Technoa with fibe tensile stengths of 2.65 GPa, 2.8 GPa, and 3.4 GPa espectively. Howeve, as a composite with cetain fibe volume atios the stength is educed to a ange of 0.62 to.93 GPa (Dolan, 993). Kevla 49 fibes wee patented by the DuPont Company and ae used in the making offiba od poduced by the Mitsui Constuction Company in Japan. This od has multiple bundles offibes that ae baided togethe and coated in an epoxy. In some pocedues, sand is used on the oute suface to incease the bond with the concete. Twaon fibes ae used fo Aapee tendons that ae consideed the stongest of the AFRP. Hollandsche Beton Goep of the Nethelands poduces this type of od by pultusion of Twaon fibes and an epoxy

19 CHAPTER 2: LITERATURE REVIEW 2-6 esin. Pultusion is the pocess wheeby fibes ae oughly shaped and pulled though a esin bath and though heated dies followed by a final stage of cuing, as shown in Figue 2-2 (Mufti et a\., 99 ). Technoa od is poduced by Teijin Limited of Japan using Technoa fibe and a esin combined duing the pultusion pocess. Fibe spool ~~- Pulling Device Cuing Chambe Winding spool Resin bath Figue 2-2 Pultusion pocess Glass FRP Reinfocement Glass fibe einfoced polymes ae the most economical of all fibe einfoced polymes and ae commonly used in low stess level conditions due to thei elatively low cost. GFRP, in geneal, is used fo non-pestessed einfoced membes. Disadvantages of GFRP include low elastic modulus, elatively high specific gavity (i n compaison to othe fibes), ceep uptue, and lowe fatigue esistance (Eki and Rizkalla, 994).

20 ( I' f i ) :',. CHAPTER 2: LITERATURE REVIEW 2-7 The two commonly used types of glass fibes ae the E-glass and S-glass fibes. S-glass fibes ae moe expensive than E-glass but ae moe tempeatue esistant and have a tensile stength of 3.9 GPa and a tensile modulus of 87 GPa. E-glass fibes have a lowe tensile stength of 2.3 GPa and tensile modulus of 74 GPa (Dolan, 993). The pesence of wate o sustained loads deceases the tensile stength of glass fibes (Abdelahman and Rizkalla, 994). Glass Fibe Reinfoced Polymes ae typically poduced using the puitusion pocess. The common industial name fo one type of GFRP ba is Isood poduced by Pultall Incopoated, Quebec. These bas ae manufactued with diametes of9.5, 2.7, 5.9, 9. and 25.4 mm. To impove the bond, extenal winding (to poduce a ibbed suface) is done and quatz sand gains ae attached to the suface (Abdelahman and Rizkalla, 994). Anothe type of GFRP ba is commecially known as C-BAR, poduced by Mashall Industies in Lima, Ohio, USA. C-BAR uses E-glass fibes and a polyuethane esin. It is also fabicated using the pultusion pocess and is sold in the fom of staight bas and pefabicated stiup shapes (both single legged and double legged). Both 2 and 5 mm diamete ba sizes ae available fabicated with esin idges to enhance the mechanical bonding to the concete.

21 CHAPTER 2: LITERATURE REVIEW Cabon FRP Reinfocement Cabon fibe einfoced polymes (CFRP) ae the most widely used einfocements due to many beneficial chaacteistics. They have a vey high tensile stength, high fatigue stength, a high modulus of elasticity, and yet low density, and low elaxation. They have the highest elastic moduli of all fibe einfoced polymes, anging fom 25 to 230 GPa. The Tokyo Rope Manufactuing Company Limited poduces Cabon Fibe Composite Cable (CFCC). Cabon fibe pe-pegs, which contain 2,000 filaments embedded in esin, ae manufactued by the Toho Rayon Co. Each pe-peg is twisted and then coveed by synthetic yams to ceate one single stand. A CFCC ope is typically poduced in single stands o by twisting lage numbes of stands in combinations of seven, nineteen, o thityseven wies as shown in Figue 2-3. This ope can be olled and shipped in a lage oll fom. The twisted ope can also be shaped into vaious stiup foms befoe cuing. CFCC has excellent bond chaacteistics to the concete due to the exta suface aea and high elative oughness of the ope (Abdelahman and Rizkalla, 994). Some CFCC cables have been fabicated which obtain 90% of the cabon fibe stength and 97% of the cabon fibe tensile modulus (Mochizuki et ai., 989).

22 CHAPTER 2: LITERATURE REVIEW CI=CC U (3.0-~5.0q,) CFCC lx7 ( ~) CFCC lx9 ( ) CFCC lx37 ( ») Figue 2-3 CFCC bas in coss-section (Tokyo Rope Manufactuing Company, 993) Leadline, poduced by the Mitsubishi Chemical Copoation, is anothe widely used CFRP ba. The pultusion pocess is again utilized fo the cabon fibes and an epoxy esin to poduce the Leadline ods. Leadline can be shipped in eels o coils and then cut fo use as a pestessing tendon in beams. It is also poduced in stiup shapes used fo shea einfocement consisting of fibes encased in a esin coating. An 8 mm diamete Leadline pestessing od has a guaanteed tensile stength and modulus of 970 MPa and 47 GPa especti vely. 2-9

23 a.. CHAPTER 2: LITERATURE REVIEW Behaviou of Concete Membes Reinfoced with FRP Due to the unique mateial popeties of FRP compaed to steel, beams einfoced with FRP behave diffeently than beams einfoced with taditional steel einfocement. As peviously mentioned, FRP mateials have a linealy elastic behaviou up to failue. This behaviou leads to concete beam failue eithe by cushing of the concete in the compession zone o by the bittle failue of the FRP einfocement. Due to the absence of yielding, beam behaviou and the mode of failue ae diffeent fom those associated with steel einfoced beams, and theefoe special design consideations should be used fo the design of concete stuctues einfoced o pestessed with Fibe Reinfoced Polymes. 2.2 FRP Shea Reinfocement FRP is easily molded into diffeent stiup shapes simila to taditional steel stiups. Howeve, because of its diffeent popeties, FRP cannot be diectly substituted fo steel stiups in design. Poblems ae encounteed due to the effect of the bends in the stiups and due to the effect of inclined shea cacks. These poblems must be investigated befoe design guidelines can be developed.

24 CHAPTER 2: LITERATURE REVIEW Classification and Fabication of FRP Stiups Diffeent configuations of FRP stiups ae cuently available on the maket. These ange fom individual stiups, simila to taditional steel stiups, to thee-dimensional gids of FRP mateial. Most of these foms of stiups ae deliveed to the site in the desied shape. The following listing descibes some of the cuently available poducts and the paticula method of fabication. Diffeent stiup configuations ae shown in Figue 2-4. The typical pe-bent closed loop stiups ae fabicated using a continuous filament winding pocess. The mateial is wound aound a mold into one lage stiup. Upon completion of the cuing pocess, the mold is emoved and the lage stiups ae cut into smalle stiups of the appopiate width, as descibed by Duanovic et al (997).

25 CHAPTER 2: LITERATURE REVIEW 2-2,, Pe-bent Pe-bent Elevation Side View open stiup closed stiup Pe-fomed spial Two-Dimensional Gid Thee-Dimensional Gid... I.....J I \\ I ~ Bent-up bas i A _ _ ~ I Figue 2-4 Vaious stiup configuations

26 CHAPTER 2: LITERATURE REVIEW 2-3 The open loop stiups used in this expeimental pogam wee deliveed pefabicated. Instead of being continuous on both ends simila to the closed loop stiups, they ae cut at the compession zone and have a standad hook (lapped) end. Mauyama (993) epoted on two types of bending pocesses. In the fist method, the pe-pegs (esin-impegnated continuous fibe ovings and mat) ae bent aound metal bas of the coect adius and heated fo cuing. Duing this pocess, the configuation of fibes at the bend becomes flattened and consequently the stength is significantly educed. This impefect section does not allow the full development of the stength of the FRP ba. Fo this eason anothe method has been developed to pevent this damage to the stiup section. The second method allows the coss section to emain cicula and thee is no flattening of the fibes. The pultuded FRP bas ae bent aound semi-cicula gooved metal molds with a adius equal to the ba diamete. By containing the fibes aound the bend, the coss-section is not defomed as the cuing pocess is completed. Some FRP stiups ae poduced in pe-fomed spial shapes, simila to steel einfocing. Pe-fomed spials ae made by continuously winding the FRP pe-pegs into spials accoding to the specified dimensions depending on the beam section. This continuous winding pocess ceates long sections which can be cut to the desied dimensions.

27 CHAPTER 2: LITERATURE REVIEW 2-4 FRP einfocements ae also poduced similaly to steel welded wie fabic in the fom of a two dimensional gid. One poduct cuently in poduction in both Canada and Japan is called NEFMAC. These flat o cuved gids wee studied by Clake (993). To fom the two-dimensional gid, a continuous filament winding (pin-winding) pocess is used to impegnate fibes with esin that is cued with ultaviolet light once fomed into the gid shape. This pocess is geneally favouable fo gids with small coss sections. To ceate sections of lage coss section, o to constuct thee-dimensional gids, a batch pocess is used. Hee, a peoxide cuing system is used to impegnate the fibes that ae fomed into the gid shape at oom tempeatue. Thee-dimensional gids have been used as cage einfocement fo concete beams Bend Effect Stiups ae widely used as shea einfocement in einfoced o pestessed concete beams. Howeve, to develop the full stength of these stiups, sufficient anchoage within the beam is equied FRP shea einfocement has been evaluated by investigating the behaviou of the bend. While the tensile stength of the staight section of the ba may be moe than adequate, the stength at the bend is significantly less. Mochizuki (995) notes that fo an end beaing system the stength of the bend is only 30% to 70% of the staight ba. By examining an individual fibe, shown in Figue 2-5, it is obvious that the lage defomation aound shap bends in the fibes can cause significant eduction of the oveall stength at the

28 CHAPTER 2: LITERATURE REVIEW 2-5 bent potion of the od. At these locations failues occu due to the esidual stesses induced by the bending of the fibes leading to a lowe stength. Figue 2-5 Micoscopic bent fibe (Hull, 98) A diagam showing the stesses at the bend can be seen in Figue 2-6. At the bend location the load is diected pependicula to the fibe diection. Since FRPs ae weak when loaded pependicula to the fibes, the adial stesses at the bend location will cause significant eductions of the stength. The fibe type, ba diamete, embedment length, tail length and bend adius all affect the stess level at the bend and subsequently goven the capacity of FRP stiups in elation to the stesses at the bend location. The following studies have peviously investigated the bend effect.

29 .. { CHAPTER 2: LITERATURE REVIEW 2-6 Figue 2-6 Stesses at the location of the bend Reseach into the capacity of CFRP and AFRP bent ods was conducted by Mauyama et al (993).32 They investigated the tensile stength of bent FRP ods while vaying the type of mateial, the bend adii and the concete stength. Pultuded CFRP ods, 7 -stand CFRP ods, and baided AFRP ods wee tested and compaed to steel bas. The intenal adii consideed in this investigation wee 5 mm, 5 mm, and 25 mm fo each type of od. Both high stength concete (50 MPa) and ulta high stength concete (00 MPa) wee used with the diffeent FRP ods. The bend was embedded in concete at an embedment length of 50 mm and tension was applied to the od using a 50 kn hydaulic jack, as shown in Figue 2-7.

30 CHAPTER 2: LITERATURE REVIEW 2-7 It 00 II ~ Z III! Gip ' I Loadcell i;" - I ~ ' I ~ ~ '. ~ Jack FRP Rod Stain Gauge (Loading Side) -. -' Stain G~e (Anchoing Side) -. >, " 50 Ancho.. I I I I I 70 I 80 I 00 I 400 mm F igue 2-7 Bent ba test specimen of Mauyama et al (993) The eseaches found that all CFRP and AFRP ods failed at the bend. They also discoveed that inceasing the cuvatue of the bend deceased the tensi le stength in a hypebolic elationship. As the bend adius deceased, the tensile foce tansfeed aound the bend to the anchoed potion inceased. With a highe stess at the bend, the stength is educed. The type of od as well as bending method affected the degee of incease of this tansfe. Use

31 L CHAPTER 2: LITERATURE REVIEW 2-8 of high stength concete inceased the stength of FRP at the bend. Hooked GFRP bas wee studied by Ehsani et al (995). They examined the elationship between the ba stength and the concete compessive stength which was vaied fom 28 to 56 JvPa. The othe paametes included in this study wee the atio of bend adius to ba diamete, the ba diamete (9.5, 9.0,28.6 mm), tail length beyond the hook, and staight embedment length befoe the hook. The eseach wok included thity-six specimens tested with GFRP bas encased in concete blocks. The load was applied though a special sand coated gipping system, which caused shea lag between the oute fibes and the inne fibes, especially fo bas of lage diamete. This setup is shown in Figue 2-8. Evaluation of the vaious paametes was based on the loaded end slip, failue load, mode of failue, and initial stiffness. The initial stiffness was detemined using the secant modulus of the load-slip elationship as shown in Figue 2-9. The load-slip elationship was coected by excluding the elastic extension of the ba along the unbonded potion within the concete (3 in.) and the potion outside of the concete block (4 in.) fom the measued slip.

32 CHAPTER 2: LITERATURE REVIEW L' ~ Unbonded Regi,\ Dial Gauge ~"""""""""""""MM t Figue 2-8 Bent ba test specimen of Ehsani et al (995) Secant Modulus 30,/ ~ Coected I.l C. :sa 20 '-' "C C = 5 ~ /Smax Slip (in.) Figue 2-9 Secant modulus used by Ehsani et al (995) 4" T

33 CHAPTER 2: LITERATURE REVIEW 2-20 The eseaches found that using lage diamete GFRP bas caused splitting of the concete blocks. Test esults indicated that the capacity was educed at the bend as pedicted and was affected by the bend adius and the ba diamete. The capacity anged fom 64 to 70% of the stength paallel to the fibes fo ba diametes of9.5, 9.0 and 28.6 mm using a bend adius of3dt,. These stengths deceased futhe, to 5, 6 and 8% fo ba diametes of9.5, 9.0 and 28.6 at a bend adius equal to zeo. They ecommended a minimum bend diamete of3d b fo GFRP hooks and a tail length of 2d b. As the staight embedment length of the ba inceased, the tensile stength and the initial stiffness also inceased. A total development length of 6d b was ecommended fo the 90 0 GFRP hooks. Cuie et al (993) investigated the bond development stength and failue modes offrp stiups as shea einfocement. They tested both Nylon!Cabon and Nylon! Aamid stiups with a coss sectional aea of 7 mm 2 and a fibe volume of 50%. Two stiups wee placed one above the othe and the two ends wee cast in concete blocks 25 mm x 50 mm x 50 mm. The stiups wee tested to failue in tension applied by a hydaulic jack located with a load cell in between the two concete blocks, as shown in Figue 2-0.

34 CHAPTER 2: LITERATURE REVIEW 2-2 GageSi Figue 2-0 Stiup bend effect test specimen of Cuie et al (993) Fo the Nylon/Cabon stiups, losses in stength wee obseved, as pedicted, due to the bend effect. The ultimate stess capacity was found to be 23% of a si mila staight ba, due to failues at the bend location. The eseaches concluded that the stength of the stiup was only 25 pecent of the ultimate stength of the Nylon/Cabon ba. lnceasing the bend adius of the stiup fom the 2.7 mm adius used may have pevented fai lue at the bend. The Nylon/Cabon stiups wee found to pefom pooly when compaed to the Nylon/Aamid stiups, due to thei poo capability to esist loading tansvese to the fibes.

35 L CHAPTER 2: LITERATURE REVIEW 2-22 Nakamua and Higai (995) conducted a theoetical investigation of the bend capacity of FRP stiups. They poposed an equation fo the bent ba stength shown as Equation 2-. Unde a tensile foce with no concete-frp bonding, the FRP ba will stetch Llx unifomly acoss the coss section in the staight potion of the ba subjected to the unifom axial foce. It is assumed that a cone adius of is maintained though a coss section defomation at a otation angle of ~ (ad). Using the Benoulli assumption, a hypebolic cuve epesents the stain distibution in the coss section. The coesponding stess distibution can be found by multiplying the stain by the elastic modulus E&p. Equation 2- fo the bent ba stength capacity is then obtained by integating the stess distibution ove the coss section, with cit, being the height of the ectangula coss section and b the bend adius of the section. This elationship is plotted and compaed to expeimental esults in Figue 2-. (2-)

36 l L CHAPTER 2: LITERATURE REVIEW I..--.J fv/f,uv o o ~ ~ L / I I ~... PROPOSED I id 3 Expeimental I I t I I I I I I Figue 2- Poposed equation fo bent capacity fom Nakamua & Higai (995) FRP stiups wee evaluated by Ueda et al (995) in tems of thei failue citeia and capacity. Aamid fibe stiups wee used in a configuation simulating the intesection of a closed stiup with a shea cack, atificially initiated with a O.5mm thick plastic plate, as shown in Figue 2-2. This FRP mateial had a coss sectional aea of25 mm 2, a nominal diamete of 6 mm, and a nominal stength of 2560 MPa in the diection of the fibes. The main vaiable in this study was the embedment length, o the length between the atificial cack and the bend. The lengths used wee 0 mm, 60 mm, and 0 mm. In conjunction with the testing pogam, a 2-D non-linea finite element analysis was conducted to

37 CHAPTER 2: LITERATURE REVIEW 2-24 detemine the local stesses at the location of the bend. Connecto to hinge Steel Plate Com pessive plate Concete -_--TT"'" FRPod Steel ba ~ Connecto to hinge Figue 2-2 Bend capacity test specimen ofueda (995) The esults show that the stiup capacity vaied between 4 and 00 pecent of the capacity in the diection of the fibes fo embedment lengths of 0 and 60 mm espectively. These failues occued at the location of the bend. Stain inceased fom the inside suface of the bend to the outside suface. With an embedment length of 0 mm the failue load was inceased beyond the nominal stength of the staight ba.

38 F t_ l l CHAPTER 2: LITERATURE REVIEW 2-25 FIBRA FRP bas wee investigated by Ishihaa et al (997) to detemine the ultimate capacity of FRP stiups as affected by the bend. They also conducted a finite element analysis to detemine the local stesses in the FRP bas at the bend. The test specimens ae simila to the ones used in the study conducted by Ueda et al.(995). Some of the FIBRA bas wee made fom aamid fibes, while othe bas wee made fom cabon fibes which wee twisted togethe, soaked in esin and bonded with sand on the suface. The ba diamete was 9 mm in both mateials and the nominal stengths in the diection of the fibes wee 00.3 kn fo the AFRP bas and 43.8 kn fo the CFRP bas. Thee bend adii wee used fo both mateials of 9 mm, 27 mm, and 45 mm which ae equal to one, thee and five times the ba diamete espectively. Anothe vaiable in this study was the stiup bond length within the concete, with the debonding extending fom the location of the atificial cack to whee the bend begins. Fou specimens of each mateial wee tested in conjunction with the finite element analysis. The eseaches found that as the bend adius deceases, the stiup stength also deceases. The stength eduction due to the bend vaied depending on the mateial used, as a esult of the diffeent bond chaacteistics. The bonded specimens showed a stength capacity anging fom 60 pecent to 86 pecent of the stength in the diection of the fibes fo the AFRP bas and 49 pecent to 66 pecent fo the CFRP bas. Fom the finite element analysis, the capacity of the bend can be pedicted using equation 2-2 deived fo a ectangula coss

39 CHAPTER 2: LITERATURE REVIEW 2-26 section (but which can be applied to cicula coss sections): With, Inclined Cack Effect ev =!In(l+A.) (2-2) ffuv "- lna. = ln( ~:) cit, = ba diamete b = bend adius The second majo effect which could lowe the capacity of FRP stiups used in concete membes is the inclined cack effect. This phenomenon occus when a vetical stiup cosses an inclined shea cack. As the cack widens, the stiup cossing this cack is subjected to a kinking action, as shown in the diagam in Figue 2-3. This movement loads the stiup in a diection out of line with the fibe diection and theefoe can cause a eduction in stength. l

40 CHAPTER 2: LITERATURE REVIEW 2-27 P CRACK OPENING it FRP Stiup Figue 2-3 Diagam of kinking behaviou Mauyama et al ( 989) studied the effect of the diection of the cack elative to the FRP bas. The eseaches used thee types offrp bas in thei study, namely cabon, aamid, and glass fibes in an epoxy esin matix. These bas, with nominal diametes of 5, and 6 mm, all had tensile stengths geate than 500 MPa. The FRP stiups wee cast into concete blocks, as shown in Figue 2-4, with an initiated cack of vaying angles up to 30

41 l ( c l l CHAPTER 2: LITERATURE REVIEW 2-28 degees. A hydaulic jack was used to load the specimen subjecting the stiup to shea foces simila to the foces induced in a stiup located acoss a diagonal cack. The esults showed that the tensile capacity of FRP bas was educed significantly by inceasing the angle of the applied load with espect to the fibe diection. The angle e of the FRP ba can be used to detemine the educed stength of the ba ffv with espect to the diection of the fibes and the stength in the diection of the fibes ffuv, using equation 2-3: ffv = ffuv(l-~e) 00 ) (2-3) whee k is the eduction facto and has a value ange of.9 to 2.3 fo CFRP bas,.9 fo AFRP bas,.3 fo GFRP bas and 0. fo steel. This applies fo angles of e between zeo and 30 degees. Using the povided equation, the diagonal stiup tensile stength at the maximum angle of30 degees is 30% of the stength in the diection of the fibes fo CFRP, 45% fo AFRP and 65% fo GFRP.

42 CHAPTER 2: LITERATURE REVIEW " Diagonal Cack and Stiups Angle of Embedded FRP Rods h m. > c:::: ~ ) o If o If '" Figue 2-4(a) Angled ba test specimen of Mauyama et al (989)

43 CHAPTER 2: LITERATURE REVIEW 2-30 Displacement Seismogaph Loading Fame Hdaulic Jack J. J Loadcell Figue 2-4(b) Angled ba test specimen of Mauyama et al (989) Two goups of eseaches, Kanematsu et al (993) and Ueda et al (995), studied the effects of combined tensile and shea foces on AFRP bas using specially designed concete block specimens. These bas had an 8 mm nominal diamete, a tensile stength of 280 MFa and an elastic modulus of 65 OPa. The FRP was located at the cente as shown in Figue 2-5, and the blocks wee sepaated into thee pats by stainless steel plates. Tension was applied to the FRP though the sepaation of the blocks unde loading fom the hydaulic jack. When the specified cack width was eached, the end blocks wee fixed in place to hold this width constant. An independent hydaulic jack was used to load the middle block lateally which subjected the tensioned FRP ba to shea displacement and shea foce. The cack width was vaied in the diffeent specimens. Kanematsu et al (993) epoted fou

44 CHAPTER 2: LITERATURE REVIEW 2-3 specimens and Ueda et al. (995) epoted a futhe eight specimens. In addition, 3-D linea finite element and 2-D non-linea finite element analyses wee conducted investigating the local stesses in the AFRP ba at the location of the cack. ~... P/ j("frp ~.... P/ Q/2 Q/2..., P/2 0 0 U U Figue 2-5 Combined tension and shea test specimen of Kanematsu & Ueda (993/995)

45 ,, CHAPTER 2: LITERATURE REVIEW 2-32.' i The eseaches found that the tensile stength of the AFRP ba was educed significantly at the cack location due to the combined action of the tensile and shea foces. They ecommended that in addition to consideing tensile and shea foces in the FRP ba fo the failue citeia, cack width and shea displacement should also be consideed. To detemine the stength of the stiups at the cack location using the finite element models, easonable assumptions must be made fo: (a) the FRP ba's shea modulus; (b) the debonding length aound the FRP ba at the cack location; and (c) the bond stess-slip elation of the FRP ba. Nakamua and Higai (995) completed a theoetical study of the diagonal tension stengths offrp bas. Thei model consideed a tensile foce applied to an FRP ba of length L at an angle of e with espect to the diection of the fibes. With L as the length of intesection between the FRP and the diagonal cack of angle e, the following equation was deived fo the diagonal tensile stength of a ectangula section: f - ffuv f - v coss + 6sin9 tans As well, they povided an equation fo cicula sections, - ffuv (2-4) f fv- (2-5) coss + 8sin9tanS

46 CHAPTER 2: LITERATURE REVIEW 2-33 In compaison to the wok done by Mauyama et al (989), the poposed equations wee shown to be easonable in evaluating the stength eduction due to the diagonal tensile foce. Howeve, these equations do not conside the effects of fibe type, despite the fact that the eductions have been shown to be dependent on the type of fibe used in expeimentation.

47 CHAPTER 3: EXPERIMENTAL PROGRAM 3. Geneal The two-phase expeimental pogam was designed to detemine the stength of FRP stiups as shea einfocement fo concete stuctues. The fist phase investigated the stength of FRP stiups as affected by bending FRP bas into stiup configuation. The paametes consideed in this phase wee the mateial type, adius of bend, ba diamete, anchoage conditions, tail length, and embedment length. One hunded and one specimens wee tested. In the second phase, twelve specimens wee constucted and tested to evaluate the effect of inclined cacks on the stiup capacity. The two paametes in this phase wee the mateial type and the cack angle with espect to the stiup. This chapte descibes the mateial popeties, specimen fabication, test set-up, instumentation and test pocedue used fo each phase. 3.2 Mateials Fo both expeimental phases, CFRP and GFRP stiups wee used, with steel stiups as contol specimens. The two types of CFRP stiups used wee Leadline and Cabon Fibe Composite Cable (CFCC). The GFRP stiups used wee C-BAR. The popeties of all stiups used ae shown in Table 3-, while Figue 3- shows a photogaph of these diffeent

48 CHAPTER 3: EXPERIMENTAL PROGRAM 3-2 stiups. CFRP bas, in compaison to othe commecially available FRP bas, have the highest tensile stength, the highest tensile elastic modulus, howeve they exhibit the lowest stain at failue. GFRP bas ae the most economical, have a low tensile elastic modulus but have a high tensile stain in compaison to othe FRP bas. Table 3- Mateial popeties of FRP and steel stiups Type of Ba CFRP CFCC GFRP Defonned Leadline ba U wie 7-wie C-BAR Steel ba Nominal diamete db (mm) ect. sec (5xl0mm) Nominal aea Ab Cmm 2 ) Guaanteed stength fir (MPa) Ultimate tensile stength f.u Elastic modulus E (GPa) Maximum stain e tu (%) >2.0 yield stength oased on tension tests

49 CHAPTER 3: EXPERIMENTAL PROGRAM 3-3 Figue 3- Stiup appeaance 3.2. Leadline The Mitsubishi Chemical Copoation fom Japan poduces the Leadline stinups used in thi s study. Leadline stiup bas ae ectangula with a coss-section measuing 5 x 0 mm with a I mm esin ibbed coating on the outside suface to potect the fibes fom damage. The fibe coss sectional aea is 38.5 mm 2, fom which an effective diamete, dc, of7.0 mm can be found based on Equation 3-. (3 - )

50 CHAPTER 3: EXPERIMENTAL PROGRAM 3-4 The stiups wee pe-fabicated in two configuations fo testing as shown in Figue 3-2. The inne adius of the bend, b, was eithe 20 mm o 50 mm poviding tlde atios of3.0 and 7.0 espectively. Testing of tensile specimens in the R.W. McQuade Stuctues Lab poduced an aveage tensile stess of 730 MPa and tensile stain of.26 pecent at ultimate. Figue 3-2 Leadline Stiup configuation Cabon Fibe Composite Cable Cabon Fibe Composite Cable (CFCC) is poduced by the Tokyo Rope Manufactuing Company Ltd. fom Japan. The thee ba types used in this study wee a seven-wie stand ba with a 7.5 mm diamete, a seven-wie stand ba with a 5 mm diamete, and a solid single wie ba of 5 mm diamete. The inne bend adius of the CFCC stiups vaied,

51 CHAPTER 3: EXPERIMENTAL PROGRAM Concete All specimens wee fabicated at the R. W. McQuade Stuctual Laboatoy at the Univesity of Manitoba using concete povided by Peimete Concete. The concete had a specified maximum aggegate size of 0 mm and a slump of 00 mm with a wate cement atio of 0.40 and a cement content of 330 kglm 3 Concete cylindes wee tested at the same time as the testing of the specimens, poviding an aveage compessive stength anging fom 36 to 48 MPa. Splitting tension tests poduced an aveage tensile stength anging fom 3.0 to 4.0MPa. 3.3 Phase - Bend Specimens The specimens in this phase wee specially designed to investigate the effect of the bend on the stength of the FRP stiups. The location of a stiup with espect to a shea cack in a beam is shown in Figue 3-3. Figue 3-4 shows a magnified view of the stesses at the bend location.

52 ( I "I'",I CHAPTER 3: EXPERIMENTAL PROGRAM 3-5 poviding Jde atios between 3.2 and 4.8. Tensile specimens wee cut fom stiups and tension tests wee pefomed. The esults ae given in Table <:-IJ)lIt C-BAR stiups ae poduced by Mashall Industies Composites in Lima, Ohio, USA. The bas used in this study had a nominal diamete of 2 mm and a coss sectional aea of 3 mm 2. The inside bend adius of these stiups was 50mm, which povided a t/de atio of 4.0. In multiple tests pefomed at the Univesity of Manitoba as pat of a compehensive mateials testing pogam, the ultimate tensile stess fo C-BAR was 640 MPa, as shown in Table 3-. This was lowe than the guaanteed stength of 73 MPa supplied by the manufactue Steel The defomed steel stiups used in all contol specimens wee supplied by Cowin Steel Company in Winnipeg, Manitoba. The stiups had a diamete of 6.35mm and a nominal yield stess of 600 MPa. Test esults showed a yielding stess of 660 MPa.,

53 t CHAPTER 3: EXPERIMENTAL PROGRAM 3-8 One hunded and one specimens wee constucted using CFCC, Leadline, C-BAR stiups as well as steel stiups fo contol specimens. The paametes fo this phase of the testing pogam, as shown in Figue 3-5, wee:. Mateial type - CFRP: Leadiine, CFCC 7-wie 7.5mm, CFCC 7-wie 5.0mm, and CFCC single wie GFRP: C-BAR Steel 2. Embedment length, ld - Length measued fom end of debonding to end of the stiup, which was vaied fom 25 to 350 mm. 3. Bend adius, b - Radius of the bent potion of the stiup. 4. Effective ba diamete, de =.J4 * AIf Based on the nominal ba coss-sectional aea. 5. End Anchoage Type - Type A lapped anchoage (lapped) with a tail length, l<t*, o Type B continuous anchoage. 6. Tail length Id* - Length of lapped end measued fom the end of the bent come.

54 CHAPTER 3: EXPERIMENTAL PROGRAM o lo N 600mm I 200 I ~ Concete block' :... ======= db = = Debonding tube ;;Ji':=. I mm.6d I ~;;:;;::::::;::==l = = ~ Standad Hook end PLAN P d de~4:b Figue 3-5 Bend specimen configuation Debonding Configuation 3-9

55 ' " CHAPTER 3: EXPERIMENTAL PROGRAM Specimen Design & Fabication The stiups wee placed hoizontally in the fom and concete blocks wee cast aound each end as shown in Figue 3-5. This anchoage within the concete simulates the anchoage of the stiup within a beam. Type A anchoages (lapped end) ae found in the compession zone of a concete beam, wheeas Type B anchoages (continuous end) ae found in the tensile zone. The concete blocks measued 200 x 250 x 200 mm, 300 x 300 x 50 mm, o 500 x 300 x 50 tnm, based on the stiup dimensions plus a 50 mm concete cove. All of the specimens had a 200 mm clea span between the two blocks fo the insetion of a hydaulic jack and load cell. The embedment length ld was measued fom the end of the stiup to the beginning of the debonded zone. Plastic PVC pipe was placed aound the stiup leg and sealed at the ends with plasticine to ceate this debonded length. The othe vaiables ae shown in Figue 3-5, and all specimens ae detailed in Tables 3-2, 3-3, 3-4, and 3-5.

56 CHAPTER 3: EXPERIMENTAL PROGRAM Table 3-2 Leadline specimens Mateial Bend Radius / Id b tjd e (mm) (mm) 2 (3d e ) b+db b+db 63 b+db b+db 20 b+db b+db Leadline 2 (3d e ) 50 b+db b+db b+db b+db (8d e ) b+db b+db Leadline N/A N/A N/A 300 staight 50 Stiup Anchoage Type A B A B N/A 3-

57 CHAPTER 3: EXPERIMENTAL PROGRAM 3-2 Table 3-3(a) CFCC Ix7 5 mm diamete specimens Mateial Bend Radius Id* Id Stiup Anchoage b tjd e (mm) (mm) Type 50* 80 A CFCC (9db) b+db 50* 80 B b+db *epoted fo both type A and type B specimen configuations Table 3-3(b) CFCC single wie 5 mm diamete specimens Mateial Bend Radius / Id Stiup Anchoage b tjd e (mm) (mm) Type 50* 80 A CFCC (9db) b+db 50* 80 B b+db *epoted fo both type A and type B specimen configuations

58 CHAPTER 3: EXPERIMENTAL PROGRAM 3-3 Table 3-3(c) CFCC Ix7 7.5 mm diamete specimens Mateial Bend Radius / Id Stiup Anchoage b tld e (mm) (mm) Type A b+db (30B) 00 b+db CFCC b+db A b+db b+db (20db) b+db B b+db

59 CHAPTER 3: EXPERIMENTAL PROGRAM Table 3-4 C-BAR specimens Mateial Bend Radius Id* Id C-BAR b tjd e (mm) (mm) 72 (6d a ) b+db (2 de) 00 b+db b+db Table 3-5 Steel specimens Mateial Nominal Bend Radius Id* Id Diamete db (mm) b tjde (mm) (mm) Steel (6db) b+db - 50* b+db b+db * epoted fo both type A and type B specimen configuations Stiup Anchoage Type No. of specimens tested 2 2 A B Stiup Anchoage Type 2 0 No. of specimens tested A B -- B 3 3-4

60 CHAPTER 3: EXPERIMENTAL PROGRAM Testing Set-up The testing set-up shown in Figues 3-6 and 3-7 consisted of: ) A 500 kn hydaulic jack, used to apply elative displacement between the concete blocks. 2) A 75 kip (330 kn) donut loadcell used to measue the load and tansmit data to the Data Acquisition system. 3) Steel plates and plaste bags used to distibute the applied load unifomly on each concete block suface. 4) 300 mm PI gauges (electic stain gauges mounted to a metal ac that monito the deflection of the ac as the two ends ae moved), attached to each side of the specimen, to monito the elative displacement of the blocks. 5) An elaboate potection system, constucted to avoid damage to the equipment, with olles placed along the bottom of one block to educe fiction between the block and the testing bed.

61 CHAPTER 3: EXPERIMENTAL PROGRAM 3-6 FRP Figue 3-6 Bend test setup Figue 3-7 Bend test photogaph

62 '-'.,.. I..,, f.,,io.,. ~ 'OF"--,

63 ( ( ( CHAPTER 3: EXPERIMENTAL PROGRAM Instumentation The stain in the stiup was measued in two ways. Electic stain gauges wee bonded to one banch of the stiup to measue stain. In addition, an MTS extensomete was attached to one banch of the stiup in the clea span to measue stain. The elative displacement between the two concete blocks was also measued. PI gages (300 mm) wee attached to each side of the specimen to continually ecod the movement of the specimen. The PI gauges wee affixed at the level of the stiup and the eadings fom these gauges wee ecoded by the data acquisition system The load cell, MTS extensomete, PI gauges and stain gauges wee all connected to the data acquisition system, which was configued to eceive output signals evety two seconds. The MTS extensomete was emoved at oughly 80% of the pedicted uptue load to avoid damage to the extensomete Test Pocedue The hydaulic jack and load cell wee positioned at the level of the stiup to avoid eccenticity duing the application of load. Load was applied manually though a pump and

64 l : CHAPTER 3: EXPERIMENTAL PROGRAM 3-8 was inceased at a constant ate. The load cell, extensomete, stain gauge, and PI gauge eadings wee monitoed on the compute fo the duation of the test while a file was being ecoded fo late analysis. 3.4 Phase n - Inclined Cack Effect Due to the diagonal natue of shea cacks, these cacks intesect the stiups at an angle and as the cack width is inceased, the stiups ae subjected to tansvese loading in addition to axial load. To detemine the losses associated with the inclined cack effect the following paametes wee examined: Mateial type - CFRP Leadline, GFRP C-BAR, steel Angle - 25, 35, 45, 53, Specimen Design & Fabication The specimen configuation is shown in Figue 3-8. The stiups wee placed as shown, at specified angles of25, 35,45, 53 o 60 degees. Table 3-6 shows a summai)' of the diffeent specimens. Steel einfocement was placed in each end of the specimen to contol cacks duing testing. The steel cage was constucted using 0 mm defomed steel bas and 6 mm smooth steel bas. The specimen was pe-cacked at the cente by the placement of sheet

65 CHAPTER 3: EXPERIMENTAL PROGRAM 3-9 metal between the two sides. A small aea measuing 75 x 235 mm was left in the cente section (whee the stiups cossed) to allow fo natual cack development and tansfe of the load fom the concete to the stiups. Two indentations wee ceated on eithe side of the specimen at the cente to allow oom fo a hydaulic jack and loadcell. These indentations wee angled towads the cente of the specimen to educe the amount of unnecessay concete in the specimen nea the citical cente cack aea and to futhe contol the specific cack location Ii!=-IH II ii". I -----~. : i I 600 i / FRP A -=--=- Sheet metal,- I, / i ~f l 240 Figue 3-8 Inclined cack specimen configuation

66 CHAPTER 3: EXPERIMENTAL PROGRAM 3-20 Table 3-6 Inclined Cack Specimens, ( l Test Set-up Mateial Nominal Effective Stiup Diamete Diamete Angle db de (mm) (mm) (deg) Leadline ect. 0x5 mm C-BAR Steel A lage plywood boad was used as a testing suface with steel plates on one end and aluminum olles on the othe end to allow fo the fee elative movement. The equipment used in this test was simila to that used fo the bend specimens: ) 500 kn hydaulic jacks placed on eithe side of the specimen in the designated indentations to apply elative displacement between both ends.

67 CHAPTER 3: EXPERIMENTAL PROGRAM 3-2 2) A 75 kip (330 kn) Donut loadcell on each side to measue the applied load. 3) Steel plates and plaste bags placed against concete load-beaing sufaces to unifomly distibute the applied load. A typical test set up is shovm in Figue 3-9. Figue 3-9 Inclined cack specimen test setul)

68 -.-

69 l ( l ( CHAPTER 3: EXPERIMENTAL PROGRAM Instumentation The instumentation fo these tests consisted of two pats. Fist, on the top and the bottom stiup in the specimen, electical stain gauges wee attached at the cack location to monito the stain in the stiup when it became load beaing due to cacking of the concete. Along the top of the specimen, a 00 nun PI gauge was attached to monito cack width and to identify the initiation of load tansfe fom the concete to the FRP stiups Test Pocedue The hydaulic jacks wee connected to independent pumps to allow full contol ove the load application on both sides. The load was applied at a constant ate fom both pumps and continually monitoed. Ifnecessay, adjustments wee made manually to the pumping ate to evenly load both sides. Load was applied until the failue of one of the stiups occued All infonnation gatheed by the attached instuments was displayed and was simultaneously ecoded by the data acquisition system. A photogaph of the entie testing aea, with a test in pogess, is shown in Figue 3-0. The esults of this goup of tests ae given in Chapte 4.

70 CHAPTER 3: EXPERIMENTAL PROGRAM 3-23 Figue 3-0 Inclined cack test photogaph

71

72 " ' l ( ( CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4. Geneal A total of 3 specimens wee fabicated and tested at the R.W. McQuade Stuctues Laboatoy at the Univesity of Manitoba to examine the behaviou and detemine the stength of FRP stiups as shea einfocement fo concete stuctues. The testing pogam consisted of 0 specimens designed to investigate the bend effect, and 2 specimens tested to evaluate the effect of inclined cacks located at an angle to the stiups. The paametes included in the bend effect study wee the mateial, bend adius, ba diamete, embedment length, anchoage conditions, and tail length. The inclined cack study included two paametes: the mateial type and the stiup angle with espect to the cack. The esults of both of these studies ae pesented and discussed in this chapte.

73 ( ( CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION Phase I - Bend Capacity The fist phase of the expeimental pogam was undetaken to evaluate the effect of the bend on the stength of FRP stiups. A summay of the 0 specimens tested is shown in Tables 4-, 4-2, 4-3, and 4-4. These tables indicate the paametes fo each specimen, failue stesses and mode of failue. The symbols used fo the diffeent modes of failue in the tables ae descibed as follows: R-S: Ruptue of the FRP stiups (o yielding of the steel stiups) along the staight potion of the stiup between the two concete blocks R-B: Ruptue of the FRP stiups (o yielding of the steel stiups) at the stiup bend location R-D: Ruptue of the stiups at the end of the debonded length inside the concete block S: Slippage of the bonded pat of the stiup S-RB: Slippage of the bonded pat of the stiup followed by uptue of the stiup at the bend potion R-BD: Ruptue of the fibes at both the bend and at the end of the debonded length -

74 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-3 l Table 4- Bend test esults: Leadline Mateial Bend Radius / Id Stiup Failue Anchoage b tld e (mm) (mm) Type 2 (3dJ b+db b+db 63 b+db A stess ffy f,jffw Failue Mode S-RB R-O S-R S-RB b+db 20 b+db R-B S-RB S-R B b+db Leadline 2 (3d e ) 50 b+db b+db b+db A R-B R-O R-B R-B R-B R-B R-B R-B R-8 l ( b+db (8dJ b+db R-8 R-8 R-S R-8 R-S R-8 R b+db CFRP N/A N/A N/A 300 N/A staight R-8 R-B R-8 S S R-S: Ruptue of stiups along staight potion between concete blocks R-B: Ruptue of stiups at the bend R-D: Ruptue of stiups at the end of the debonded length inside the concete S: Slippage of the bonded potion of the stiup S-RB: Slippage of bonded potion of stiup followed by uptue at the bend R-BD: Ruptue of some fibes at the bend and othes at the end of the debonded length

75 l l ( l L CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION Table 4-2 (a) Bend test esults: CFCC 5mm 7-wie Mateial Bend Radius / Id Stiup Failue fjffuv Anchoage stess b Jd e (mm) (mm) Type ffy SO A 975. CFCC (9d b ) b+db mm (7-wie) SO B b+db S2 *epoted fo both type A and type B specimen configuations Table 4-2 (b) Bend test esults: CFCC 5mm single wie Mateial Bend Radius Id Id Stiup Failue ff.)ffuv Anchoage stess b Jd e (mm) (mm) Type ffv SO CFCC (9db) b+db A 9S mm 50* (single) SO B b+db S *epoted fo both type A and type B specimen configuations 4-4 Failue Mode R-S R-S R-B R-S R-S R-B Failue Mode R-S R-S R-B R-S R-S R-B

76 E L ( l CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION Table 4-2 (c) Bend test esults: CFCC 7.5mm 7-wie Mateial Bend Radius / Id Stiup Failue f,./fuv Anchoage stess b tjd e (mm) (mm) Type fy A b+db (3db) b+db CFCC mm (7-wie) b+db A b+db b+db (20db) b+db B b+db Failue Mode R-B R-B R-B S-RB S-RB R-B S-RB S-RB R-B R-S R-B R-B R-B R-O R-B R-S R-S R-B R-S R-S R-B

77 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-6 l L Table 4-3 Bend test esults: C-BAR Mateial Bend Radius Id* Id stiup No. of Failue Anchoage specimens stess b tld e (mm) (mm) Type tested ftv 72 (Sd e ) b+db A (2 de) C-BAR b+db 6 345* B b+db 0 347* * aveage value ** failue by splitting of the concete block, subsequent specimens potected against this type of failue Table 4-4 Bend test esults: Steel 400 f,.}ffln Failue Mode 0.62 R-B 0.56 R-S 0.78 R-S 0.73 R-S 0.59 R-S 0.63 R-D 0.63 R-S 0.58 R-D 0.48 R-B 0.57 R-D 0.72 R-D 0.42 R-D 0.82 Splitting** 0.79 R-BD 0.7 R-O 0.49 R-B l Mateial Nominal Bend Radius Il Id Stiup No. of Failue Diamete Anchoage specimens stess db (mm) b tld e (mm) (mm) Type tested ttv A 757 Steel (6d b ) b+db * B b+db b+d b B * epoted fo both type A and type B specimen configuations f,.}fw Failue Mode.28 R-S.26 R-S 0.99 R-B.28 R-S.2 R-8.26 R-B

78 , L l C CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION Leadline Stiups In most of the Leadline stiup specimens, failue occued in the bent potion of the stiup. The failue was initiated though bond loss between the inne fibes and the oute esin coating of the ba. Figue 4- shows typical Leadline specimens afte failue measued stength in the stiups (98% of the guaanteed 760 MPa) was achieved though the use of a lage embedment length of 350 mm o 50de. Fo compaison, the measued tensile stength of staight bas tested at the Univesity of Manitoba was 730 MPa. The expeimental esults indicated that the stength of the Leadline stiups was affected by the embedment length within the concete block as elated to the bend adius. At the smallest embedment length ofb+de, the capacity of the stiups was limited to 35 pecent and 44 pecent of the guaanteed stength fo t/de atios of 3.0 and 7.0 espectively. The capacity fo the maximum embedment lengths of 50 mm (2.5de) and 350 mm (50de) wee 78 pecent and 98 pecent of the guaanteed stength fo the same two atios of 3.0 and 7.0 espectively. These esults clealy show the dependence of high stength on sufficient embedment within the concete. following the splitting of the concete block to inspect the embedded stiup. The highest L

79 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-8 f: :,;A." FAILURE LOCATION Figue 4- Failed Leadline bend specimens

80

81 L CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-9 Othe factos affecting the capacity of these stiups wee the anchoage type and the tail length, ld*. Fo the Type A anchoage used fo stiups with a t/de atio of 3.0, the stiup capacity deceased as the tail length was deceased. Howeve, the esults also suggest that with a sufficient tail length in the Type A anchoage, the pefomance is simila to the Type B anchoage of the same stiup. Thee was no significant vaiation in the stiup capacity by vaying the tail length fo stiups with a t/de atio of CFCC Stiups The Cabon Fibe Composite Cable, CFCC, stiups exhibited significant losses due to the bend effect. As the embedment length was deceased, the measued ultimate stength was significantly lowe than the guaanteed stength in the diection of the fibes. A maximum embedment length of 50 mm was sufficient to achieve the full capacity of the mateial in the diection of the fibes esulting in failue within the staight potion of the stiup. Typical failues of the specimens, afte splitting the concete blocks following testing, ae shown in Figue 4-2.

82 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-0 Figue 4-2 Failed CFCC bend specimens

83 < <,,,,..~

84 L \ CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4- F o the smalle diamete specimens, stiup stength anged fom 5 to 2 pecent of the guaanteed stength in the diection of the fibes, with the vaiation due to changes in the embedment length. The stengths of the 5 mm 7 -wie stiups wee ten pecent highe than the values fo 7.5 mm 7-wie stiups, with all othe paametes equal. This diffeence may be attibuted to shea lag acoss the coss section of the 7.5 mm stiup. By examining the esults of the 5 mm diamete stiup tests it can be shown that an embedment length of only 80 mm (6de) is equied to each the guaanteed stength in the diection of the fibes. The testing pogam fo the 7.5 mm 7-wie stiups included vaiation of all paametes. As the tjde atio was inceased, the measued capacity at the smallest embedment length of b+de (5de) was also inceased. With a tail length of 45 mm (7de), the bend capacity was 43 and 62 pecent of the stength in the diection of the fibes fo tlde atios of 3.2 and 4.8 espectively. These esults clealy indicate that the bend adius geatly affects the bend capacity of CFCC stiups. The tail length, ld, also affected the ultimate capacities fo these stiups. It can be seen in Table 4-2( c) that as the tail length is inceased, the measued ultimate bend capacity inceases. The capacity was 42 and 0 pecent of the guaanteed stength in the diection of the fibes fo tail lengths of 20 (3de) and 50 (20de) espectively. Fom these

85 c ' CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-2 esults, using a tail length of 90 mm (5de), the diffeences between Type A and Type B anchoages ae negligible. With a shote tail length, the Type A anchoage specimens have a lowe capacity than the Type B specimens. Beyond 5de, the lapped end has sufficient length to be equivalent to a continuous anchoage C-BAR Stiups C-BAR stiups exhibited less desiable behaviou due to pematue mateial failue. At vaious locations along the stiup, it was obseved that the fibes wee not paallel to the diection of the stiup. This impefection o "waving" of the fibes caused pematue failue of the stiups. Figue 4-3 shows photos of a typical C-BAR stiup both befoe and afte failue, whee the waving effect is clealy shown. This impefection subjects the fibes to non-axial stesses and theefoe causes a significant eduction of the ultimate capacity of the ba at this location. This type of impefection was found in most of the stiups but the location of the impefection and the seveity of the fibe misalignment vaied. Using this type of mateial lead to inconsistent esults fom the diffeent paametes consideed to evaluate the capacity of the stiups.

86 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-3 Figue 4-3 C-BAR "waving" impefection

87 .' J ~,,Jf.,, ";;" ", f:, -<. " ~'..~-.k-.,,.,t..,,'"... ;' ~,~, t~ '~."'" " ~,. :, "'. >,I.,,,,\., '.,...., ".. > ~ ~.:;(, >t' ";-.:. " 'I;',..,,T (, '. " '" t\~,~ ~'\\, ", (, \,,.. 't.'!" q " ''#,, -t'~.~; 'f'.ii~ ~\.. "., ~ :..,; ff. -~-....,, ",'i- "'~i.. '_n ' ", ~. ". ~' ':~:.'t ) ~.,.; ~~ ~)t(! ". \~'it.,, ~,,- " :. f~ -,,,,.' i ";"{"". ". If I ~,.:

88 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-4 Despite the inconsistent esults, the effect of a vaied embedment length was clealy noted. When ful ly debonded, the stiups had a capacity of 48 pecent of the guaanteed value. In tension tests pefomed at the Univesity of Manitoba, the measued tensile stength was only 640 MPa as compaed to the guaanteed stength of 73 MPa povided by the manufactue. The geatest stength achieved in the bend specimens was only 82 pecent of the guaanteed val ue. This loss of capacity fom 82% to 48% shows the dependence of C-BAR stiup stength on the embedment length. Unfotunately few othe conclusions could be made due to the impefections ofthe mateial used. A C-BAR stiup with failue along the staight potion initiated by a defect is shown in Figue 4-4 while the failue of a C-BAR specimen at the bend can be seen in Figue 4-5. Figue 4-4 C-BAR specimen failue along staight potion of stiup due to defect

89 --,,. :. " I,;' «

90 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-5 FAll.,URE LOCATION Figue 4-5 Failed C-BAR specimen at bend location

91 !,.. tij":',\. ~,.~., " ) '",./J. " "..;:. ~ ~P:.. ',.. '. - "., ",.,.'... ~, ~':".", "i' ī

92 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION Steel Stiups The steel stiup specimens failed due to yielding of the stiups along the staight potion fo specimens with a lage embedment length, and yielding at the bend fo specimens with the minimum embedment length. The failue stess in all specimens was equal to the guaanteed stength of a staight ba due to the anisotopic behaviou of steel. While losses due to the bend effect wee not noted and should not be pesent in this mateial, slippage could occu in specimens with shot tail lengths. 4.3 Effect of Bend Radius on Bend Capacity F o most mateials used in this pogam, test esults indicated that the ultimate capacity was highe fo the stiups with the lage bend adius. This behaviou is a esult of the inne fibes at the bend of small adius stiups showing moe sevee bending and buckling in compaison to stiups with a lage bend adius. Results showed that the vaiation of the bend adius had a moe ponounced effect on CFCC specimens then Leadline specimens.

93 ( CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-7 The elationship between the measued stess at failue to the guaanteed stength paallel to the fibes, ffv/ffuv, as affected by the bend adius to ba diamete atio, tlde, is shown in Figue 4-6. Equations 4- and 4-2, ecommended by the JSCE eseach committee (997) fo aveage and consevative design, ae shown on the same figue. whee,...!ft.. = 0.09!.h. + OJ ~.0 ffuv de.lh.. = O.OS!h. + OJ ~.0 ffuv de (4-) (4-2) ffuv = ultimate stength of the FRP mateial in the diection of the fibes b = bend adius de = the effective ba diamete

94 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-8 f ".4, ,.2 I. CFCC x CBAR I :. ~.. ' If 0.8 f" = l =. ~. i..... f,,,, \ d, ~. II!... g.... -it "'... ' :,::... --;.~ >-.. ~... '-::::\'.! = 0, f uv d e O.,0::: f " ' b -.. ~ _ _----~------_----~----~ o Figue 4-6 Effect of bend adius on stiup capacity with equations poposed by JSCE (997) Figue 4-6 indicates that the JSCE equations ae below the aveage of the measued data fo CFCC and C BAR stiups tested in this pogam. Howeve, the equation oveestimates the stength fo the Leadline stiups with a lage bend adius, Test esults wee used to establish design ecommendations fo the bend adius of FRP stiups. In ode to achieve a stiup capacity geate tban 50 pecent of the guaanteed

95 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-9 stength in the diection of the fibes a bend adius to ba diamete atio tjdeof 4.0 should be used fo CFCC and C-BAR stiups and 7.0 fo the Leadline stiups. Theefoe, fo the detailing of CFCC and C-BAR stiups the bend adius b shall not be less than 4de o 50 tnm, whicheve is geate. Fo Leadline stiups the bend adius b, shall not be less than 7de o 50 mm. 4.4 Effect of Embedment Length on Bend Capacity Test esults indicated that a eduction of the embedded length of the stiup exposes the bend zone to much highe stesses causing a eduction in stength. If the embedment length is not sufficient to develop the full stength of the ba, losses will occu due to the educed stength at the bend. The effect of the embedment length fo CFCC, Leadline, and C-BAR stiups is shown in Figues 4-7, 4-8, and 4-9 espectively. The elation is povided in tems of the measued capacity to the guaanteed stength paallel to the fibes ffv/ffuv, and the embedment length as a function of the effective diamete, ld/de. The esults show that the eduction in the stength occus below a cetain value of the IJde atio fo each type of stiup. The ld/de value was 20 fo the CFCC Type A stiups and 6 fo CFCC Type B stiups. Using the value of ld/de of 5.0 epesenting the end of the bend zone, the measued aveage stength fo CFCC stiups using Types A and B

96 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-20 stiups wee 50 and 74 pecent of the stength in the diection of the fibes, espectively, as shown in Figue 4-7. Figue 4-8 indicates that the limiting value of ld/dc fo the Leadline stiups to achieve the guaanteed stength ffu" paallel to the diection of the fibes is 42. Using a small development length educes the stength to 40 pecent of the guaanteed stength paallel to the fibes, as shown in Figue 4-8. Equations 4-3, 4-4, and 4-5 given in the two figues can be used to pedict the stength capacity of the stiup ffv fo intemediate development lengths:,-', , Type 8 _ J:nc=060+~ ffu\', 40d, c,,,, T I' ~.,... I I ~ Anchoage Typo A. lapped end debonding l o ~ o 0 20 ~.h 40 ~ d, AAchoago Typo B contlnuo!js end debonding Figue 4-7 Effect of embedment length, (d, on CFCC stiup capacity

97 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-2.4, ,.2 ~ f ~ ~ = ,, _ ffln 70d e I L d Figue 4-8 Effect of embedment length, 0, on Leadline stiup capacity Anchoage Type A -lapped end debonding Anchoage Type B - continuous end debonding f, f~ (0.60 maximum) ; " \ " " f tv Id --:: I f "". 20d. i 0 o Figue 4-9 Effect of embedment length, 0, on C-BAR stiup capacity

98 "~ -!.'. ~ i..:~'.o.",, -'iii -;.'''- IJ.. ~... "'-,,,:...,.',, ' ~'. ", ',' \,... } ~"!',.: ",' ' '; '.-~.. '. oj,

99 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-22 CFRPCFCC 0.52 ~ ffv = 0.35+l ~.00 ffuv 30d e Type A Anchoage (4-3) 0.73 ~ ffv =0.60+l ~.00 Type B Anchoage (4-4) ffuv 40d e CFRP Leadline 0.47 ~ ffv = l ~.00 ffuv 70de (4-5) The same analysis was attempted to pedict the behaviou of the C-BAR stiups. Howeve, due to the "waving" effect fom the impefections intoduced in the poduction of the bas, none of the specimens eached the full guaanteed stength in the diection of the fibes. The stength of C-BAR specimens eached only 80 pecent of the stength in the diection of the fibes as shown in Figue 4-9. Due to this constaint, the failue citeion poposed based on test esults is valid up to a maximum of 80 pecent as given in equation 4-6.

100 t CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-23 GFRPC-BAR 0.49 ~ ffv = _l_d_ ~ 0.80 ffuv 20d e 4.5 EtTect of Stiup Anchoage on Bend Capacity (4-6) Two anchoage conditions wee simulated in the expeimental pogam. Type A stiups simulate a standad hook anchoage, typically located in the compession zone of a einfoced concete beam. Type B stiups simulate a continuous anchoage typically located in the tension zone of the beam. It was obseved that the tail length, ld*, beyond the bend location fo Type A specimens geatly influenced the stength of the anchoage. Test esults indicate that the capacity of the Type A anchoage is less than the Type B anchoage up to a tansition value of the tail length (which is dependent on the ba type and diamete). This eduction in capacity is due to slipping at the bend fo stiups with a shot tail length leading to highe stesses at the bend. Fo the 5 mm diamete efee specimens using only one ld* length of 45 mm (9de), the Type A anchoage is notably weake in compaison to the Type B anchoage. Fo the 7.5 mm diamete efee specimens, as ld* eached 2de, the stength of the Type A anchoage was equivalent to

101 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-24 the stength of the Type B anchoage. Fo ld* less than 2de, the stength of the Type A anchoage was less than fo the Type B anchoage F o the two bend adii used, in the Leadline specimens, the Type B anchoage had a highe capacity than Type A until the tail length eached a value of 8de. Theefoe it can be concluded that the Type B, o continuous anchoage, has a geate capacity than the Type A, standad hook, anchoage unless a sufficient tail length, ld*, of 2de fo CFCC, and 8de fo Leadline, is povided. 4.6 Effect of Tail Length on Bend Capacity Based on the esults obtained with the two anchoage types, the tail length ld*, was vaied in elation to the ba diamete in this investigation. In geneal, it was noted that as the tail length deceased, the ultimate capacity of the stiups also deceased, due to a loss of anchoage. Without sufficient end anchoage, the bond along the embedment length becomes the only paamete affecting the capacity of the stiup. Fo the CFCC stiups with 7.5 mm diamete, tail lengths of 3, 6, 9, 2, and 20 de wee tested. Figue 4-0 indicates that as the tail length inceases, the capacity as epesented by the atio fvlffuv, also inceases. At a tail length of 5 de, the guaanteed stess value

102 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-25 was achieved fo all embedment lengths. Below a tail length of 5 de the stength is lowe than guaanteed to a minimum of 42 pecent of the guaanteed stength in the diection of the fibes at a tail length of 3de. Using the minimum embedment length of Tb+dc, equation 4-7 is poposed to pedict the capacity of the stiup as a function of the tail length as follows: CFCC fv = _l_d_ ::; loa (4-7) ffuv 7dc.2, , f f\' f fu\' e e e Jl _ Id-b+de. d-6de e ld=24de !..L d, Figue 4-0 Effect of tail length, Id*' on CFCC stiup capacity

103 '""%,',:,',',:":.':,,.,:,, :"~, I. Z: i. : igl : ~: i!!. _~,I (.,~ I ~!L::! '~:i -:,; I.: U ( ) I I ):J : :. ~ ; ~'j'(.-f,;~ i" j f.f:~.:-; - : '" " ~ i) ~ '... I ~ '".<'.', "-" -..,~" ~_... _, o.'! b / '. :, '.,. D.) \::..,.':.j ~! (I,- -J.i.: J:' : f., '.,... ' w_~~_~ _.:....':... " :;l,~",:""::,, ;~,

104 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-26 The obseved elationship between fiv/ffuvand Id*/de was not as well-defined fo Leadline specimens. The tail lengths of Leadline specimens wee vaied at 3, 6, 9, 2, and 8de. Fo the small bend adius to diamete atio, t/de of 3.0, behaviou was simila to what was obseved fo the CFCC specimens. Howeve the stength was much lowe, anging fom 35 to 44 pecent of the guaanteed stength in the diection of the fibes. The capacity of stiups with a lage bend adius to diamete atio, t/de of 7.0 anged fom 44 to 53 pecent of the guaanteed stength in the diection of the fibes. A minimum tail length of 70 mm (lode) is ecommended to develop bend capacities of 40 and 50 pecent of the guaanteed stength in the diection of the fibes fo t/de atios of 3.0 and 7.0 espectively. Due to the mateial impefections in the C-BAR stiups, esults fom the study of tail length ae inconclusive. Two tail lengths wee used of 6de, and 2de. It was obseved that the vaiation of tail length did not significantly affect the stength of the FRP stiups. In geneal, fo the detailing of all FRP stiups, the tail length, ld*, shall not be less than 6de o 70 mm, whicheve is geate.

105 ( CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION PHASE II - Capacity at Inclined Cack The second phase of the expeimental pogam was undetaken to evaluate the effect of cack inclination on the capacity of FRP stiups. A plot of the applied load vesus the cack opening displacement based on PI gauge eadings duing a typical test fo Leadline an angle of 35 degees is shown in Figue 4-(a). A plot showing the load vesus the measued stiup stain fom the same test is shown in Figue 4- (b). Both figues show the tansfe of load fom the concete to the stiups as the concete cacks acoss the section. This behaviou simulates the load-tansfe mechanism in concete beams. Failed C-BAR and Leadline stiups at the cack location afte failue ae shown in Figues 4-2 and 4-3 espectively.

106 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION ~ ~ ~ z.:.: "0.. o..j ~~ ~~~~~~::~------~ ~~ ~ ~ ~ O ~ ~ o Displacement (mm) Figue 4- (a) Typical test plot of Load vs. PI gauge eadings (Leadline - 35 ) ~ _ ~ 2 ::. "0.. o..j ~~ ~~S~G~B~o~tt~o~m~~~S~G~T~o P~ O '-----~------~ ~ ~ o Stain (ms) Figue 4- (b) Typical test plot of Load vs. stain gauges (Leadline - 35 )

107 ,, it'. ~.. -'.,, " h~ ",,. ~..'..., '.,..,: ",, i ;. -I; '; j ', """,

108 CHAPTER 4: EXPERIMENTAL RESULTS AND DISCUSSION 4-29 Figue 4-2 Failed C-BAR stiups Boken ba with sepaation of esin coating _ oo _ Figue 4-3 Failed Leadline stiups