Вы находитесь на странице: 1из 9


Title no. 107-S25


Carbon Fiber-Reinforced Polymer Confined Reactive Powder Concrete ColumnsExperimental Investigation

by Adnan R. Malik and Stephen J. Foster
An experimental investigation was conducted to investigate the behavior of ultra-high-strength reactive powder concrete (RPC) columns confined by carbon fiber-reinforced polymers (CFRPs) and subjected to concentric and eccentric loadings. Seventeen columns were cast with the concrete mixture containing either no fibers, with a concrete strength of approximately 140 MPa (20.3 ksi), or 2% (by volume) of straight steel fibers, with a concrete strength of approximately 165 MPa (23.9 ksi). The column specimens contained no conventional steel reinforcement, either in the longitudinal or transverse direction with the tensile forces carried by the CFRP. Experimental data for strength, lateral and axial deformation, and the failure mode were obtained for each test. For the concentrically loaded specimens, failure occurred at or close to the peak loading with little or no residual capacity. The transverse strains measured at the fracture of the CFRP for confined columns were found to be significantly lower than the ultimate tensile strength reported by the manufacturer or obtained from the standard tensile coupon tests. For the eccentrically loaded columns, the final failure was sudden and explosive but only after the peak load was passed and at the point of tearing of the CFRP wrapping. There was no evidence, however, that the use of CFRP in the hoop direction significantly increased the strength for the eccentrically loaded columns.
Keywords: carbon fiber-reinforced polymer; columns; confinement; reactive powder concrete; wrapping.

INTRODUCTION The introduction of fiber-reinforced polymers (FRPs) in civil engineering structures has progressed at a rapid rate in recent years. These high-performance materials that consist of high-strength fibers embedded in a polymer matrix have unique properties, making them extremely attractive for structural applications. FRPs are noncorrosive, have high strength-to-weight ratios, possess good fatigue behavior, and allow easy handling and installation. Moreover, as the fiber types and fiber volumes can be combined in numerous ways, their overall mechanical properties can be tailored to provide optimum solutions to a wide range of structural applications. One area where the use of fiber-reinforced composites has attracted considerable interest is in the strengthening of concrete columns. The confinement of concrete columns is a well-established technique for improving both the compressive behavior and flexural response. With FRP wraps, the confining layer is very thin and is applied directly onto the surface of the column. The composite wraps are flexible and can be handled and cut with little effort. Alternatively, FRP may be used as a permanent formwork and used with other high-performance materials, such as reactive powder concrete (RPC), to produce a high-performance composite material. As construction and material costs escalate, demand has increased for stronger materials that occupy less space, provided the small initial capital cost of the structure is offset by the more significant economic benefit of increased rental ACI Structural Journal/May-June 2010

space. Reactive powder concrete is an ultra-high-strength, low porosity cement-based composite with high ductility. Unlike conventional concrete, RPC containing a significant quantity of steel fibers exhibits high ductility and energy absorption characteristics.1,2 Conventional concrete is a heterogeneous material with components from fine cement to coarse aggregates each having different strengths and moduli of elasticity. Under a system of forces, all of these component materials deform at different rates. Reactive powder concrete is composed of particles with similar elastic moduli and is graded for dense compaction, thereby reducing the differential tensile strain and enormously increasing the ultimate load carrying capacity of the material. Interest in ultra-high-strength cement-based materials is not solely because of their increased strength. They possess other high-performance properties, such as low permeability, limited shrinkage, increased corrosion and abrasion resistance, and increased durability.3,4 These are all valuable characteristics used in the construction industry for concrete structures. The technology offers the possibility to build structural elements without passive reinforcement (for example, conventional steel ties in columns) in structural elements and combines innovation, lightness, and high durability. Reactive powder concrete may also be used from the standpoint of weight reduction, for its architectural aspects, or for its high resistance to blast and impact loadings. The objective of this research was to examine the behavior of carbon fiber-reinforced polymer (CFRP) confined fiber and non-fiber-reinforced RPC circular columns and to determine any increase in the magnitude in strength and ductility provided by the CFRP, without conventional reinforcement. Seventeen columns were cast and tested under concentric and eccentric loading in this study and the results are reported herein. The test results form an important data set for future development of design models for this unique combination of materials. RESEARCH SIGNIFICANCE With the advancement of high-performance construction materials, such as those reported herein, new structural applications come to the fore, for example, the potential of using CFRP as a permanent formwork shell combined with an ultra-high-performance RPC core in bridge piers and foundations. Research5-9 has shown that the strength and ductility of a concrete column can be significantly improved by including a high-performance fiber composite shell
ACI Structural Journal, V. 107, No. 3, May-June 2010. MS No. S-2007-321.R3 received April 28, 2009, and reviewed under Institute publication policies. Copyright 2010, American Concrete Institute. All rights reserved, including the making of copies unless permission is obtained from the copyright proprietors. Pertinent discussion including authors closure, if any, will be published in the MarchApril 2011 ACI Structural Journal if the discussion is received by November 1, 2010.


Adnan R. Malik is a Structural Design Engineer with Cardno Stanwill Consulting Engineers, Newcastle, Australia. He received his PhD in 2007 from the University of New South Wales, Sydney, Australia. His research interests include the use of ultrahigh-strength reactive powder concrete in structural elements, nonlinear finite element analysis of concrete structures, and the use of fiber-reinforced polymers in structural applications. ACI member Stephen J. Foster is a Professor in the School of Civil and Environmental Engineering at the University of New South Wales. He received his PhD from the University of New South Wales in 1993. His research interests include the structural use of high-strength concretes, fiber-reinforced concrete, design of high-strength concrete columns, nonlinear finite element analysis of concrete structures, and strutand-tie modeling.

was a control column with no wrap. Details for the columns tested are presented in Table 1. RPC mixture design and material properties The RPC was mixed using locally available materials: 920 kg/m3 (57.4 lb/ft3) of Type 1 general portland cement, 920 kg/m3 (57.4 lb/ft3) of Sydney sand, 221 kg/m3 (13.8 lb/ft3) of undensified silica fume, and 157 kg/m3 (9.8 lb/ft3) of steel fibers for the FR-RPC mixture. The steel fibers used were high-strength straight steel fibers 13 mm (0.5 in.) long, 0.2 mm (0.008 in.) in diameter, and had an ultimate tensile strength of 1800 MPa (261 ksi). The water-binder ratio (w/b) was 0.17. All of the dry constituents of the RPC were batched by an electronic balance. The dry constituents were then mixed in a pan concrete mixer for approximately 10 minutes. Water and a high-range water-reducing admixture were then added gradually until the materials were uniformly mixed. The fibers were introduced last and dispersed uniformly using a sieve and were mixed for an additional 10 minutes. The control specimen properties are presented in Table 2, where f is the volumetric ratio of fibers, Eo is the modulus of elasticity, is Poissons ratio, fcm is the mean compressive cylinder strength, fdp is the double punch tensile strength calculated by the Chen and Yuan12 equation, fsp is the split cylinder tensile strength, fcf is the flexural tension strength, and Gf is the fracture energy. A typical stress-strain curve for a 200 x 100 mm (8 x 4 in.) diameter FR-RPC cylinder is presented in Fig. 1. Fabrication of specimens The columns were cast vertically in circular steel molds with two columns cast from each batch. All of the stainless steel molds were cleaned and oiled to allow for smooth stripping. To prevent fiber segregation, the RPC was compacted using external vibrators that were clamped to the external surfaces of each mold. Approximately 2 hours after casting, all of the columns and control specimens were covered with wet hessian and plastic sheet. The FR-RPC specimens remained covered for 24 hours, whereas the RPC specimens without fibers were kept under cover for 48 hours. The samples were covered for different lengths of time because non-fiber-reinforced RPC is an extremely brittle Table 2RPC material properties
Mixture f , % 1 2 3 4 5 6 7 8 9 2 2 2 2 2 2 0 0 0 fcm, MPa (ksi) 163 (23.6) 165 (23.9) 168 (24.4) 172 (24.9) 168 (24.4) 165 (23.9) 139 (20.2) 143 (20.7) 145 (21.0) 0.13 0.1 0.1 0.13 0.15 0.15 0.13 0.13 Eo, GPa (ksi) 41.7 (6048) 42.0 (6092) 42.6 (6179) 44.2 (6411) 42.6 (6179) 44.2 (6411) fdp, MPa (ksi) 8.2 (1.2) 7.7 (1.1) 7.9 (1.1) 7.6 (1.1) 8.7 (1.3) 7.7 (1.1) fsp, MPa (ksi) 8.0 (1.2) fcf, MPa (ksi) 26.2 (3.8) 35.4 (5.1) 17.7 (2.6) 26.2 (3.8) Gf, N/mm (lb/ft) 20.7 (1418) 32.7 (2240) 14.1 (966) 18.4 (1261)

around the column. For RPC confined in a steel tube, the ultimate strength and ductility of the material also increases when subjected to a compressive load.10 To date, most of the experimental, analytical, and numerical studies conducted on FRP confined columns are based on conventionalstrength concrete, with limited research for high-strength concrete (HSC).11 No experimental data sets are available for the development of models for ultra-high-performance concrete confined concentrically or eccentrically by CFRP. This is addressed herein. EXPERIMENTAL PROGRAM Test specimens In this study, 17 RPC columns were tested consisting of 10 CFRP confined steel fiber-reinforced RPC (FR-RPC) columns, six CFRP confined RPC columns with no steel fiber, and one FR-RPC column with no wrapping. The columns were circular; 152 mm (6 in.) plus or minus 0.2 mm (0.8 in.) in diameter and 1050 mm (41.3 in.) high. No longitudinal reinforcement was used in any of the columns. The columns tested were identified by load eccentricity, wrap type, existence of steel fibers in the concrete, and an identification number for the specimen if the test was repeated. For example, Specimen FC351 was cast with FR-RPC, was tested at an initial eccentricity of 35 mm (1.4 in.), and was wrapped with a Type 1 layout carbon fiber polymer sheet. Specimen PC022 was cast with RPC with no fibers, was tested under concentric loading (zero millimeters eccentricity), was wrapped with a Type 2 carbon fiber sheet, and was a repeat test. Column FC0 Table 1Column details
Column FC60-1 FC35-1 FC20-1 FC10-1 FC0-1-1 FC0-1-2 FC60-2 FC35-2 FC20-2 FC10-2 PC60-1 PC35-1 PC20-1 PC10-1 PC0-1-1 PC0-1-2 FC0 Mixture (Table 2) 1 2 1 2 3 3 5 5 6 4 8 8 7 7 9 9 6 Wrap type (Table 4) 1 1 1 1 1 1 2 2 2 2 1 1 1 1 1 1 Load eccentricity, mm (in.) 60 (2.4) 35 (1.4) 20 (0.8) 10 (0.4) 0 0 60 (2.4) 35 (1.4) 20 (0.8) 10 (0.4) 60 (2.4) 35 (1.4) 20 (0.8) 10 (0.4) 0 0 0

24.7 21.1 (3.60) (1439) 33.3 (4.8) 6.1 (0.88) 28.8 (1973) 0.03 (2.1) 0.02 (1.4) 0.03 (2.1)

40.5 3.1 (5874) (0.45)

8.6 6.4 3.3 (0.48) (1.25) (0.93)

41.0 3.0 8.6 6.0 (5947) (0.44) (1.25) (0.87)


ACI Structural Journal/May-June 2010

material and any mishandling might have cracked or damaged the whole sample. Therefore, non-fiber-reinforced RPCs were kept under cover for 48 hours to attain enough strength. All specimens were then stripped and cured for 72 hours at 80C (176F) in a hot water bath. After 3 days, the specimens were removed from the hot water bath and air cured until the day of capping. The top surfaces of all of the circular columns were then capped with approximately 5 to 10 mm (0.2 to 0.4 in.) thick nonfiber RPC paste to provide a smooth horizontal loading surface and were air cured for 5 to 7 days. All of the columns and the control specimens were then placed for an additional 72 hours at 80C (176F) in a hot water bath. After 3 days, the specimens were removed from the hot water bath and were air cured for at least a week before wrapping. FRP material properties and wrapping Two types of CFRP were used. The resin used for the bonding of the CFRP was a two-part epoxy adhesive. The mechanical properties of the CFRP and epoxy, as given by the manufacturer, are presented in Table 3. The CFRP was tested to obtain the ultimate strength and elastic modulus in accordance with ASTM D303913 and ISO 10406-2:2008.14 The mean ultimate tensile strength of the three CF120 specimens was 3420 MPa (496 ksi), with the strain corresponding to a failure load of 0.0147. The elastic modulus was 233 GPa (33,794 ksi). Before the CFRP was applied to the column, the concrete surface was cleaned using sandpaper. The two-component primer was mixed and applied to the concrete surface using medium-sized rollers. The two-part epoxy adhesive was then mixed and applied over the tacky primer. Next, the CFRP was placed on the concrete surface and gently pressed and rolled into the resin. After placing and wrapping the fiber sheet, a second coat of saturant was applied to the FRP sheet using a medium-sized roller. Additional fiber plies were installed by resaturating the surface with resin and repeating the steps previously described. For the circumferential wraps, an overlap length of 200 mm (8 in.) was used for the joining of the fiber sheets. The wrap details are presented in Table 4 and shown in Fig. 2. In Fig. 2(a), the strain gauge locations are marked as SG. The wrapped columns were cured in an ambient condition for at least 7 days before testing. Table 3Mechanical properties of CFRP and adhesive
Property Tensile strength, MPa (ksi) Ultimate strain Thickness, mm (in.) Fiber weight, g/m2 (lb/ft2) CF120 3800 (551) 0.0155 200 (0.041) CF530 2650 (384) 0.0041 400 (0.082) Saturant >50 (7.2)

Instrumentation and test setup For the eccentrically loaded columns, lateral deflection measurements were taken at the column midheight and 150 mm (6 in.) above and below midheight using three linear variable differential transducers (LVDTs). Axial deformation was recorded using two LVDTs over a gauge length of 900 mm (35.4 in.). Lateral deflection measurements (when recorded) for concentrically loaded columns were taken at column midheight using three LVDTs placed 120 degrees apart on the circumference. Axial deformation was recorded using three LVDTs over a gauge length of 900 mm (35.4 in.) that were located equally between the lateral LVDTs. For each column, four electrical strain gauges were placed at column midheight in the longitudinal direction and four in the circumferential direction to measure axial and hoop strains, respectively. The instrumentation details are shown in Fig. 3(a).

Fig. 1Axial compressive stress versus axial (compressive) strain and circumferential (tensile) strain for RPC cylinder.

Modulus of elasticity, GPa (ksi) 240 (34,809) 640 (92,824) >3 (435) 0.117 (0.0046) 0.19 (0.0075)

Table 4Details of CFRP wrapping

Wrap Type 1 Wrap Type 2 Laminate Carbon fiber Carbon fiber structure Sheet Wrap direction Sheet Wrap direction Layer 1 Layer 2 Layer 3 Layer 4 CF120 CF120 CF120 CF120 Longitudinal Longitudinal Circumferential Circumferential CF530 CF530 CF120 CF120 Longitudinal Longitudinal Circumferential Circumferential

Fig. 2CFRP wrapping: (a) overlap details; and (b) exploded view for circumferential wrapping. 265

ACI Structural Journal/May-June 2010

The columns were tested using a specially designed loading apparatus (similar in design to that adopted by Fam et al.15) that applied a coupled axial load and bending moment. The apparatus consisted of 155 mm (6.1 in.) diameter steel end caps placed over the ends of the column. The RPC specimens were tested in a 5000 kN (1124 kips) capacity closed-loop servo control system testing machine with special end assemblages attached to the top and bottom platens of the machine. The top and bottom base plates contained knife edge supports to transfer the eccentric load to the specimens. The test setup is shown in Fig. 3(b). The same loading setup was used for concentric loading tests, with the axis of the column set exactly in line with the axes of the knife edges (zero millimeters eccentricity). EXPERIMENTAL RESULTS AND DISCUSSION The peak loads Pu, moments at peak load Mu, corresponding lateral displacements mid at midheight of the specimen and the location of failure zone for the columns tested are given in Table 5. Also presented are the loads on the descending branch at which the CFRP ruptured, Fu, and the corresponding midheight displacements. After each test, the end steel cap was removed to examine the top of the specimen. No local damage was observed in the concrete underneath the steel cap.

Table 5Peak loads and corresponding moments

Moment Failure Peak load at Pu, * P , kN kNm location, u Column mm (in.) (kip) (kip-ft) FC60-1 FC35-1 FC20-1 FC10-1 FC0-1-1 FC0-1-2 FC60-2 FC35-2 FC20-2 FC10-2 PC60-1 PC35-1 PC20-1 PC10-1 PC0-1-1 PC0-1-2 FC0 +330 (13) +50 (2) +200 (8) +170 (7) +150 (6) +50 (2) +330 (13) 50 (2) 5 (0.2) +230 (9) +328 (13) +238 (9.5) +20 (0.8) +150 (6) +160 (6.3) +200 (8) 403 (90.6) 714 (160.5) 1357 (305) 2221 (499) 2971 (668) 2993 (673) 317 (713) 833 (187) 1367 (307) 1912 (430) 334 (75.1) 773 (174) 1287 (289) 1756 (395) 2571 (578) 2495 (561) 2510 (564) 35.5 (26.2) 38.9 (28.7) 46.4 (34.3) 42.0 (31) 23.3 (17.2) 43.0 (31.7) 47.0 (34.7) 39.4 (29) 29.8 (22) 43.4 (32) 45.2 (33.3) 38.6 (28.5) mid at Pu, mm (in.) 28.2 (1.1) 19.4 (0.9) 14.2 (0.56) 8.9 (0.74) 13.5 (0.53) 16.6 (0.65) 14.4 (0.57) 10.6 (0.42) 29.2 (1.15) 21.2 (0.83) 15.1 (0.6) 12.0 (0.47) Failure load Fu, kN (kip) 373 (83.3) 532 (119.6) 366 (82.3) 835 (187.7) 2971 (668) 2993 (673) 183 (41.1) 562 (126.3) 493 (110.8) 117 (26.3) 327 (73.5) 521 (117.1) 818 (183.9) 813 (182.7) 2571 (578) 2495 (561) 2510 (564) mid at Fu, mm (in.) 32.4 (1.3) 35.6 (1.4) 53.4 (2.1) 39.6 (1.6) 20.5 (0.8) 23.0 (0.9) 30.2 (1.2) 47.2 (1.8) 32.5 (1.3) 37.2 (1.5) 30.0 (1.2) 39.6 (1.6)

Fig. 3Testing arrangements: (a) instrumentation; and (b) test setup. 266

+ is above and is below column midheight. Test stopped before fiber rupture.

ACI Structural Journal/May-June 2010

Fig. 4Typical failure of eccentrically loaded CFRP confined FR-RPC columnSpecimen PC10-1: (a) failure at column midheight; (b) side view of ruptured CFRP; (c) hoop tension failure of circumferential fibers; and (d) tensile failure of longitudinal fibers and splitting of circumferential fibers. Eccentrically loaded columns For the tests conducted on CFRP confined columns under eccentric loading, several snapping sounds were heard near the peak load as the CFRP fibers began to stretch and/or rupture; however, there were no visible signs of any impending failure on the surface of the FRP wrapping. For columns tested under initial load eccentricities of 10, 20, and 35 mm (0.4, 0.8, and 1.4 in.), specimen failure occurred when the longitudinal FRP wrapping ruptured in tension and the circumferential wrapping split vertically on the tensile side. The final collapse of the specimen was induced by hoop tensile failure of the CFRP circumferential wrapping on the compressive side. Figure 4 shows details of typical failure of eccentrically loaded CFRP confined RPC columns. In the eccentrically loaded specimens, however, the fracture of the CFRP occurred well beyond the peak loading. In Fig. 5, the circumferential strain, measured by the gauge located on the compression side of the specimens, is plotted against the axial strain for fiber and non-fiber-reinforced RPC columns confined with CFRP Type 1 wrapping. The figure shows that the compressive concrete was under a considerable confining pressure toward the end of the test, particularly for the columns with the smaller 10 and 20 mm (0.4 and 0.8 in.) initial loading eccentricities. At the peak load, however, the degree of confinement supplied by the CFRP was small. The circumferential-axial strain relationship at the maximum axial compressive zone is distinctly nonlinear, the transition point indicating the onset of significant confinement. ACI Structural Journal/May-June 2010

Fig. 5Axial compressive strain versus circumferential tensile strain for columns: (a) with steel fibers; and (b) without steel fibers. For eccentrically loaded non-fiber-reinforced RPC columns, considerable straining beyond the peak load was recorded for columns with smaller 10 and 20 mm (0.4 and 0.8 in.) initial loading eccentricities before the failure of the specimens. This response, however, was not observed for non-fiber-reinforced RPC columns tested under higher 35 and 60 mm (1.4 and 2.4 in.) initial loading eccentricities. Because of the brittle nature of the failure in tension for RPC columns without fibers, once the column cracked on the tensile side with increasing lateral displacement, there was a sudden uncontrolled increase in deflection and the CFRP ruptured suddenly, particularly for columns with high initial eccentricities of 35 and 60 mm (1.4 and 2.4 in.). For confined FR-RPC columns, the ductile failure mechanism in tension for FR-RPC coupled with the stiffness of the CFRP wrap led to an increase in ductility. The peak load achieved for RPC columns without fibers was lower compared with FR-RPC columns because of their lower compressive and tensile strength. Figure 6 plots the graph between the axial load and circumferential tensile strain on the compressive side for fiber and non-fiber-reinforced RPC columns confined with Type 1 CFRP wrap. Figure 6 shows that, for the eccentrically loaded FR-RPC columns, there was considerable straining 267

Fig. 8Axial load versus midheight lateral displacement for RPC columns with Type 1 and Type 2 CFRP wrapping. plateau in the moment-curvature relationship for the specimens with initial eccentricities of 20, 35, and 60 mm (0.8, 1.4, and 2.4 in.), indicating that considerable ductility was achieved using CFRP wrapping. Specimen FC10, with an initial eccentricity of 10 mm (0.4 in.), also shows good ductility for a member that is essentially axially loaded. The strain gauges were damaged during the test for specimens confined with Type 2 wraps before the fracture of the CFRP and post-peak curvature data was not available for those specimens. Also observed in Fig. 7 is the effect of slip between the CFRP shell and the RPC core in Specimens FC10 and FC20. In Specimen FC10, a change of strain in the compressive gauge of 720 microstrain was recorded at a moment of 44.2 kNm (32.5 kip-ft). Similarly, between moments of 39 and 44 kNm (28.7 and 32.4 kip-ft), a reduction in strain of 410 microstrain was recorded in the compressive gauge attached to the CFRP. Figure 8 compares the load versus midheight lateral displacement graph for eccentrically loaded FR-RPC columns wrapped with Type 1 and Type 2 wraps. No major difference is observed between the load versus midheight deflection response for confined RPC columns wrapped with CFRP wrap Types 1 and 2. Using the higher stiffness CFRP wrap (Type 2) did not lead to a stiffer initial load-deflection curve. It is likely that increasing the number of wrapped layers with a higher modulus of elasticity in the columns longitudinal direction will lead to a stiffer load-deflection response. Further investigations, however, are required to verify this. Figure 9 compares the load versus midheight lateral displacement, measured at the midheight, for the CFRP confined RPC columns with and without steel fibers. The figure shows that both types of wrapped columns behaved in a similar way with final failure occurring well beyond the peak load. It is worth noting that the failure of the non-fiberreinforced RPC column without wrap is expected to be extremely brittle. Thus, the CFRP wrapping was useful in controlling the failure of the specimens even without steel fibers in the RPC mixture and provided for some ductility. The peak axial loads and corresponding moments for the CFRP confined RPC columns are plotted in Fig. 10, together with the axial force bending moment interaction diagram. The interaction diagram was obtained using an elasticACI Structural Journal/May-June 2010

Fig. 6Axial load versus circumferential tensile strain for CFRP confined RPC columns: (a) with steel fibers; and (b) without steel fibers.

Fig. 7Moment-curvature diagram for FR-RPC confined columns with Type 1 wrapping. (Note: 1 mm = 0.0394 in.) beyond the peak load before the final failure of the columns by fiber rupture. The moment-curvature diagram for the eccentrically loaded FR-RPC columns confined using Type 1 CFRP wrapping is presented in Fig. 7, with the moment calculated as the axial load multiplied by the sum of the load eccentricity at the ends and the midheight displacement. Curvatures were calculated from strain gauges placed longitudinally on the exterior of the CFRP and located at midheight on the tensile and compressive faces. The figure shows an extended flat 268

Fig. 10Peak axial loads and corresponding moments for columns tested. Fig. 9Axial load-versus-midheight lateral displacement for RPC columns with and without steel fibers. plastic stress-strain model with a yield stress of 0.9fcm, an elastic modulus of 42.5 GPa (6164 ksi), and a compressive failure strain of 0.005, determined experimentally from compressive cylinder tests. The tensile strength of the concrete is taken as zero. The figure shows that the peak load is greater than the model calculation. Concentrically loaded specimens The concentrically loaded columns failed in a sudden and explosive manner and the descending branch could not be captured using the ram displacement control. For the CFRP confined specimens tested, failure was characterized by CFRP failing in hoop tension followed immediately by explosive rupture of the RPC. The typical failure of the concentrically loaded RPC columns is shown in Fig. 11. The axial load and axial strain continued to increase (Fig. 12) until the FRP shell failed in tension. Figure 12 shows that the confinement provided to Specimen FC0-1-2 increased the failure load by 19% over that of the unconfined Specimen FC0. The tests revealed that the ultimate circumferential tensile failure strains recorded (Fig. 12) were significantly lower than the ultimate tensile failure strain reported by the manufacturer or determined from the standard tensile coupon test, and lower than what is typical for normal-strength concrete FRP wrapped columns.16 This reduction in the failure strain of the CFRP composite can be attributed to several causes that include misalignment or damage to jacket fibers during handling and layup, the cumulative probability of weakness in the FRP material because FRP wraps are much larger than tensile coupons, the radius of curvature of the CFRP wraps on columns as opposed to flat tensile coupons, and the presence of voids or protrusions and the misalignment of fibers. Although confined and still able to resist loads near failure, the concrete is internally cracked resulting in nonhomogeneous deformations that might lead to local stress concentrations in the CFRP. It is well established that mechanical properties and failure modes differ significantly for normal-strength concrete and HSC. The use of HSC may also lower the ultimate FRP failure strain for FRP confined HSC concrete columns than for FRP confined normal-strength concrete. It is difficult to draw any conclusions regarding this statement, however, unless a large data set for FRP confined HSC concrete is available. Until evidence exists to the contrary, it is ACI Structural Journal/May-June 2010

Fig. 11Failure of concentrically loaded columns of Specimen PC0-1-1. recommended that in design models used to predict the strength of CFRP confined RPC columns, the hoop strain in the CFRP not be taken as greater than 0.004 for determining the confining pressure applied to the section at the strength limit condition. It is noted that this strain is significantly lower than that adopted in ACI 44016 for the design of FRP wrapped conventional strength concrete columns. In Table 6, the peak loads measured in the concentric tests are compared to values obtained using various models 269

Table 6Experimental and predicted loads

Column Experiment Lam and Teng17 (A) (A/Experiment) Saafi et al.

FC0-1-1, kN (kip) 2971 (668) 3180 (715) 1.07 3459 (778) 1.16 3365 (757) 1.13 3848 (865) 1.3 4006 (901) 1.35 3007 (676) 1.01 3330 (749) 1.12 3130 (704) 1.05

FC0-1-2, kN (kip) 2993 (673) 3188 (718) 1.07 3468 (780) 1.16 3373 (758) 1.13 3857 (867) 1.29 4015 (903) 1.34 3014 (678) 1.01 3338 (750) 1.12 3138 (706) 1.05

PC0-1-1, kN (kip) 2571 (578) 2802 (630) 1.09 3077 (692) 1.2 2999 (674) 1.17 3459 (778) 1.35 3614 (813) 1.4 2649 (596) 1.03 2975 (669) 1.16 2762 (621) 1.07

PC0-1-2, kN (kip) 2495 (561) 2813 (632) 1.13 3066 (689) 1.23 2988 (672) 1.2 3447 (775) 1.38 3602 (810) 1.44 2639 (593) 1.06 2968 (667) 1.19 2754 (619) 1.10

Average of model/ experiment 1.09 1.19 1.16 1.33 1.38 1.03 1.15 1.07


(B/Experiment) Samaan et al.19 (C) (C/Experiment) Toutanji20 (D) Model predictions (D/Experiment) Mander et al.21 (E) (E/Experiment) Youssef et al.22 (F) (F/Experiment) Berthet et al.11 (G) (G/Experiment) ACI 440.2R16 (H) (H/Experiment)

Table 7Experimental and predicted failure strains

Column Experiment Lam and Teng17 (A) (A/Experiment) Samaan et al.19 (B) (B/Experiment) Toutanji20 (C) Model predictions (C/Experiment) Mander et al.21 (D) (D/Experiment) Youssef et al.22 (E) (E/Experiment) ACI 440.2R16 (F) (F/Experiment) Average of FC0-1-1 and PC0-1-1 and model/ FC0-1-2 PC0-1-2 experiment 0.004 0.0158 3.95 0.036 9 0.018 4.5 0.016 4 0.0058 1.45 0.01 2.50 0.0039 0.0167 4.28 0.0345 8.8 0.02 5.12 0.0178 4.56 0.0063 1.61 0.01 2.56 4.1 8.9 4.8 4.3 1.53 2.53

Fig. 12Axial load versus axial and transverse strain. experimental failure strain recorded for the CFRP confined concentrically loaded RPC columns. All of the predicted failure strains are significantly greater than the actual experimental failure strains for the columns tested, with the exception of predicted failure strains by the Youssef et al.22 model. It is worth noting that because the tests did not show a bilinear stress-strain response typical of FRP confined conventional-strength concrete columns with similar levels of relative confinement, due to the lower confinement pressure applied, a comparison of the ultimate failure strain with the experimental FRP failure strains is less significant. CONCLUSIONS In this study, 17 RPC columns were tested with 16 confined using CFRP. The column specimens contained no conventional steel reinforcement, either in the longitudinal or transverse direction with longitudinal tensile forces carried by CFRP. For the concentrically loaded specimens, failure occurred at or close to the peak loading with little or no residual capacity. Comparing the strength of the CFRP confined specimen with the unconfined specimen, the CFRP increased the strength by 19%. The transverse strains measured at the fracture of the CFRP for confined columns were significantly ACI Structural Journal/May-June 2010

reported in the literature11,16-22 for determining the strength of CFRP confined conventional and HSC columns. In calculating the predicted loads, the strength of in-place concrete is taken as fcp = k3 fcm, where k3 is the in-place strength factor and is taken as k3 = 0.9. It may be noted that all of the predicted loads are higher than the actual experimental capacity of the columns tested. The Youssef et al.22 model predicts the experimental results most closely whereas all models overpredict the strain in the CFRP at the point of failure. It appears that the confinement effectiveness by CFRP is significantly less for ultra-HSC compared with conventional-strength concrete. This may be attributed to the lower dilation of the higher-strength concrete compared with conventional-strength concrete and the higher energy release at failure causing fractures in the circumferentially wrapped CFRP. Due to the limited data available on CFRP confined ultra-HSC, however, further research is required. Table 7 presents the predicted ultimate confined compressive failure strains by various models and compares it with the 270

lower than the ultimate tensile strength reported by the manufacturer or obtained from the standard tensile coupon tests. It is therefore recommended that in design models used to predict the strength of CFRP confined RPC columns, a factor be placed on the rupture strain of the CFRP compared with the rupture strain obtained from standard FRP flat coupon tensile test data. It appears that the FRP confinement effectiveness decreases in concentrically loaded FRP confined RPC specimens because of the lower dilation of RPC under axial load. The formulas developed to predict the peak strength of concentrically loaded FRP confined conventional-strength concrete should be used with caution for CFRP confined RPC specimens under concentric loading. Most of the examined formulae available in the literature overestimate the strength of concentrically loaded RPC columns confined using CFRP wrapping. This said, the Youssef et al.22 model was able to predict the experimental results closely. For the eccentrically loaded specimens, the CFRP was shown to be effective in controlling the failure of the specimens with considerable straining occurring beyond the peak loading. There was no evidence, however, that the use of CFRP in the hoop direction significantly increased the strength of the columns. As hoop strains increased beyond the peak loading, the stresses in the wrapping induced failure of the hoop CFRP and, eventually, a final explosive collapse of the specimen, resulting in a total loss of any residual strength. ACKNOWLEDGMENTS
This study is funded via Australian Research Council (ARC) discovery grant DP0211516. The support of the ARC is acknowledged with thanks. The authors are also thankful to MBT (Australia) Pty. Ltd. for the supply of the CFRP laminates and resins and their support toward the project.

d Eo Fu fcf fcm fcp fdp fsp Gf k3 Mu Pu mid f = = = = = = = = = = = = = = = diameter of circular column modulus of elasticity of concrete load on descending branch at which CFRP ruptured flexural tensile strength mean compressive cylinder strength strength of in-place concrete double punch tensile strength split cylinder tensile strength fracture energy concrete in-place strength factor moment at peak load peak load midheight lateral displacement Poissons ratio of concrete volumetric ratio of fibers

1. Richard, P., Reactive Powder Concrete: A New Ultra-High-Strength Cementitious Material, 4th International Symposium on Utilization of High Strength/High Performance Concrete, Press de l'Ecole Naturale des Ponts et Chausses, Paris, France, 1996, pp. 1343-1349. 2. Richard, P., and Cheyrezy, M. H., Reactive Powder Concretes with High Ductility and 200-800 MPa Compressive Strength, Concrete

Technology: Past, Present, and Future, SP-144, P. K. Mehta, ed., American Concrete Institute, Farmington Hills, MI, 1994, pp. 507-518. 3. Roux, N.; Andrade, C.; and Sanjuan, M. A., Experimental Study of Durability of Reactive Powder Concretes, Journal of Materials in Civil Engineering, V. 8, No. 1, 1996, pp. 1-6. 4. Bonneau, O.; Pouline, C.; Dugat, J.; Richard, P.; and Atcin, P. C., Reactive Powder Concretes: From Theory to Practice, Concrete International, V. 18, No. 4, Apr. 1996, pp. 47-49. 5. Nanni, A., and Bradford, N. M., FRP Jacketed Concrete under Uniaxial Compression, Construction and Building Materials, V. 9, No. 2, 1995, pp. 115-124. 6. Mirmiran, A., and Shahawy, M., Behavior of Concrete Confined by Fiber Composites, Journal of Structural Engineering, ASCE, V. 123, No. 5, 1997, pp. 583-590. 7. Fam, A. Z., and Rizkalla, S. H., Behavior of Axially Loaded Concrete-Filled Circular Fiber-Reinforced Polymer Tubes, ACI Structural Journal, V. 98, No. 3, May-June 2001, pp. 280-289. 8. Matthys, S.; Toutanji, H.; Audenaert, K.; and Taerwe, L., Axial Load Behavior of Large-Scale Columns Confined with Fiber-Reinforced Polymer Composites, ACI Structural Journal, V. 102, No. 2, Mar.-Apr. 2005, pp. 258-267. 9. Harajli, M. H.; Hantouche, E.; and Soudki, K., Stress-Strain Model for Fiber-Reinforced Polymer Jacketed Concrete Columns, ACI Structural Journal, V. 103, No. 5, Sept.-Oct. 2006, pp. 672-682. 10. Dallaire, E.; Atcin., P. C.; and Lachemi, M., High-Performance Powder, Civil Engineering, V. 68, No. 1, 1998, pp. 49-51. 11. Berthet, J. F.; Ferrier, E.; and Hamelin, P., Compressive Behavior of Concrete Externally Confined by Composite Jackets Part B: Modeling, Construction and Building Materials, V. 20, 2006, pp. 338-347. 12. Chen, W. F., and Yuan, R. L., Tensile Strength of Concrete: DoublePunch Test, Journal of the Structural Division, ASCE, V. 106, No. ST. 8, 1980, pp. 1673-1693. 13. ASTM D3039, Standard Test Method for Tensile Properties of Polymer Matrix Composite Materials, ASTM International, West Conshohocken, PA, 2005, 13 pp. 14. ISO 10406-2:2008, Fibre-Reinforced Polymer (FRP) Reinforcement of ConcreteTest Methods, Part 2: FRP Sheets, ISO International Standards Organization, Geneva, Switzerland, 2008. 15. Fam, A.; Flisak, B.; and Rizkalla, S., Experimental and Analytical Modeling of Concrete-Filled Fiber-Reinforced Polymer Tubes Subjected to Combined Bending and Axial Loads, ACI Structural Journal, V. 100, No. 4, July-Aug. 2003, pp. 499-509. 16. ACI Committee 440, Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures (ACI 440.2R-08), American Concrete Institute, Farmington Hills, MI, 2008, 76 pp. 17. Lam, L., and Teng, J. G., Strength Models for Fiber-Reinforced Plastic-Confined Concrete, Journal of Structural Engineering, ASCE, V. 128, No. 5, 2002, pp. 612-623. 18. Saafi, M.; Toutanji, H. A.; and Li, Z., Behavior of Concrete Columns Confined with Fiber-Reinforced Polymer Tubes, ACI Materials Journal, V. 96, No. 4, July-Aug. 1999, pp. 500-509. 19. Samaan, M.; Mirmiran, A.; and Shahawy, M., Model of Concrete Confined by Fiber Composites, Journal of Structural Engineering, ASCE, V. 124, No. 9, 1998, pp. 1025-1031. 20. Toutanji, H. A., Stress-Strain Characteristics of Concrete Columns Externally Confined with Advanced Fiber Composite Sheets, ACI Materials Journal, V. 96, No. 3, May-June 1999, pp. 397-405. 21. Mander, B. J.; Priestley, M. J. N.; and Park, R., Theoretical Stress Strain Model for Confined Concrete, Journal of Structural Engineering, ASCE, V. 114, No. 8, 1988, pp. 1804-1825. 22. Youssef, M. N.; Feng, M. Q.; and Mosallam, A. S., Stress-Strain Model for Concrete Confined by FRP Composites, Composites Part B: Engineering, V. 38, No. 5-6, 2007, pp. 614-628.

ACI Structural Journal/May-June 2010