Concrete Column Shape Modification with FRP and Expansive Cement Concrete

Transcription

1 CICE The 5th International Conference on FRP Composites in Civil Engineering September 27-29, 2010 Beiing, China Concrete Column Shape Modification with FRP and Expansive Cement Concrete Zihan Yan (yanzh214@yahoo.com) & Jeffrey B. Duffin Department of Engineering Services, California Department of Transportation, Sacramento, USA Chris P. Pantelides Department of Civil and Environmental Engineering, University of Utah, Salt Lake City, USA Jeffrey B. Duffin Department of Engineering Services, California Department of Transportation, Sacramento, USA ABSTRACT: Fibre Reinforced Polymer (FRP) composites are an effective material for confining circular concrete columns. FRP confinement for square and rectangular columns is less effective due to stress concentrations at the sharp corners and loss of the membrane effect. Shape modification using post-tensioning of FRP shells through expansive cement concrete is described. In the field, shape modification can be achieved by utilizing pre-fabricated FRP shells as the permanent forms. An analytical model is briefly introduced to predict results of the experiments regarding the enhanced stress-strain behaviour of FRP-confined concrete. The confinement model and shape modification technique with FRP shells and expansive cement concrete are used in simulations of seismic rehabilitation of square columns for existing reinforced concrete bridges and are compared to in-situ tests of square columns with bonded FRP ackets. 1 INTRODUCTION Externally bonded FRP composite ackets can provide effective confinement for circular concrete columns (Karbhari et al. 1997). However, FRP confinement is much less effective in increasing the axial compressive strength of square and rectangular columns compared to circular columns (Rochette and Labossière 2000, Pessiki et al. 2001) due to stress concentrations at the corners and ineffective confinement at the flat sides. The presence of steel ties limits rounding of the corner radius in existing square or rectangular columns. Lower confinement effectiveness for square/rectangular columns results in softening behavior and premature FRP composite rupture; therefore, the inherent high tensile strength of FRP composite materials cannot be fully utilized. One approach for improving the effectiveness of FRP ackets for rectangular columns is to use prefabricated (non-bonded) FRP composite shells with expansive cement concrete. A prefabricated elliptical/oval/circular FRP shell may be used as stay-inplace formwork for casting additional expansive cement concrete around the square or rectangular cross-section to achieve shape modification. Expansive cement normally consists of a Portland cement and a calcium-sulfoaluminate anhydrite component. The mechanism of expansive cement concrete can be used with FRP composite shells for confinement: when expansive cement concrete is applied to prefabricated FRP shells, expansion of the grout is restrained by the FRP shell, thus creating a posttensioning effect which confines the original concrete core as well as the expansive grout. It is apparent that this post-tensioning effect would increase the confinement behavior of FRP shells and change the confinement action from passive to active. Experiments were conducted to investigate the shape modification effect on large concrete specimens, and analytical model for developing the axial stress-strain relationship of shape-modified columns is briefly described. The confinement model and shape modification technique with FRP shells and expansive cement concrete are used in simulations of seismic rehabilitation of square columns for existing reinforced concrete bridges and are compared to actual in-situ tests of square columns with bonded FRP ackets. 2 EXPERIMENTAL PROGRAMS AND RESULTS 2.1 Experimental program The experiments involved FRP-acketed specimens bonded for the full specimen height, as well as shape-modified specimens confined with FRP. Two FRP composite systems, a Carbon Fiber Reinforced Polymer (CFRP) system and a Glass Fiber Reinforced Polymer (GFRP) system were used. Three

2 groups of specimens: square specimens (S-type), rectangular specimens with an aspect ratio of 2:1 (R2-type) and 3:1(R3-type), respectively. All specimens were 914 mm high; no steel reinforcement was used inside the specimens. Each group included an unconfined (baseline) specimen, two specimens with the original square or rectangular cross-section confined by bonded CFRP or GFRP ackets and two shape-modified specimens using prefabricated CFRP or GFRP shells with expansive cement concrete. For S-type the shape-modified cross-section was circular, and for R2-type and R3-type it was elliptical, as shown in Fig. 1. S-TYPE R2-TYPE R3-TYPE Figure 1. Cross-sections of shape-modified specimens: S-Type (square to circle); R2-Type (2:1 rectangle to ellipse); R3-Type (3:1 rectangle to ellipse). Table 1 lists the details of all specimens. The specimens are identified using a three-code scheme. The first part is the shape of the column (Square or Rectangular), and the aspect ratio of the rectangular cross-section (2:1 or 3:1). The second part indicates the type of FRP composite (CFRP or GFRP) and the number of FRP layers (2 or 6, respectively). The third part denotes the type of material used to achieve shape modification; expansive cement concrete is denoted as (E) and (0) denotes no shape modification. Table 1. Dimensions of column specimens. Specimen Original crosssection Modified cross-section (B xd )* (mm) (mm) S-C S-G R2-C R2-G R3-C R3-G S-C2-E S-G6-E R2-C2-E R2-G6-E R3-C2-E R3-G6-E * Denote maor axis length (B ) and minor axis length (D ) Regular and expansive cement concrete were used in this experimental program. Regular concrete was used to cast the original square or rectangular specimens; expansive cement concrete was used to perform shape modification. The mix design of expansive cement concrete was investigated by Yan et al. (2006). The 28-day compressive strength of the original concrete specimens was 15 MPa. For the shape-modified specimens, unconfined concrete strength f co was calculated by obtaining a mean value as: A A o f fco fco1 fco2 (1) A A where f co1 = unconfined concrete strength of square or rectangular specimens; f co2 = unconfined strength of expansive cement concrete obtained from cylinder compression tests, which was 10 MPa; Ao = area of square or rectangular cross-section; Af = area of expansive cement concrete; A = total area of shape-modified cross-section. Using this method f co is calculated to be 13 MPa for specimens with expansive cement concrete. Two FRP composite materials were used to confine the concrete columns. One was SikaWrap Hex 103C which is a high strength, unidirectional carbon fiber fabric with epoxy resin. The other was Aquawrap G-06, which is a unidirectional preimpregnated glass fiber fabric with urethane resin. The material properties were determined from tensile coupon tests per ASTM Standard D3039 (2001). The properties of the two FRP composite systems were: (1) for the CFRP composite, which was cured with epoxy resin, tensile strength was 1220 MPa, tensile modulus was 87 GPa, and ply thickness was 1.0 mm; and (2) for the GFRP composite, which was cured with urethane resin, tensile strength was 228 MPa, tensile modulus was 17 GPa, and ply thickness was 1.6 mm; for both FRP composite systems, the ultimate tensile strain was 1.4%. All concrete specimens were subected to uniaxial compression until failure. The load was applied using displacement control at a constant rate of 1.3 mm/min. 2.2 Experimental results For FRP-confined square/rectangular columns without shape modification, failure started with concrete crushing followed by fracture of the FRP composite acket at a corner. Failure was brittle due to the corner and flat side effects, which eliminate membrane action of the FRP acket and result in ineffective confinement except at the diagonals and the four corners. Failure of shape-modified specimens with non-bonded FRP shells and expansive cement concrete was fracture of the FRP shell and cracking of the expansive cement concrete. Fracture of the FRP shell extended over the specimen height, demon-

3 strating extensive participation of the FRP shell in confinement. At the end of the test, vertical and diagonal cracks were observed in the expansive cement concrete, but the original concrete column cross-section was protected. These specimens achieved a higher compressive strain compared to FRP-bonded specimens. Specimens with a smaller aspect ratio reached a higher axial strength. GFRPconfined specimens failed less explosively and at a higher hoop strain than CFRP-confined specimens. Axial strain was measured using the average of two linear variable displacement transducers; axial stress was computed by dividing the axial compression load by the total cross-sectional area. CFRPconfined square specimen S-C2-0 showed limited hardening behavior and CFRP-confined rectangular specimen R3-C2-0 demonstrated softening behavior; a drop of axial stress was observed after the initial axial strength was reached, and the degree of softening increased with aspect ratio. For shape-modified specimens, the stress-strain curves show ascending branches without softening behavior. Improvement of compressive strength and ultimate strain capacity is significant for shape-modified square columns S- C2-E since their modified shape was circular; improvement was less significant for rectangular columns, especially the sections with the higher aspect ratio as the section becomes a flatter ellipse. A summary of the experimental results in terms of ultimate compressive strength f cc and corresponding compressive strain cc is given in Table 2. Stress-strain curves are not presented herein due to the page limitations. More details of the experimental results are described elsewhere (Yan et al. 2006). Table 2. Main experimental results for column specimens. Column Designation f cc (MPa) f cc /f co cc (mm/mm) S-C S-G R2-C R2-G R3-C R3-G S-C2-E S-G6-E R2-C2-E R2-G6-E R3-C2-E R3-G6-E ANALYTICAL MODEL 3.1 Dilatancy behavior of shape-modified columns In this study, the dilatancy behavior of FRPconfined concrete is represented by the volumetric strain versus axial strain relationship. olumetric strain is defined as the FRP area strain in the two transverse orthogonal directions minus the axial strain in the concrete column c : 2 c (2) where is defined as the FRP hoop strain for circular cross-sections or average FRP hoop strain for elliptical cross-sections. Figure 6 shows the volumetric strain versus axial strain relations for shape-modified specimens with expansive cement concrete. Since the FRP shell was already posttensioned prior to axial loading through chemical post-tensioning, the amount of radial expansion was smaller compared to bonded FRP ackets. Therefore, the axial strain was larger than the hoop area strain, 2 ; this reveals that the axial strain was dominant in the volumetric strain versus axial strain curve. This dilatancy behavior is extremely important for shape-modified FRP specimens with expansive cement concrete because in this case the FRP confinement becomes active instead of passive. The authors suggest the relationship between volumetric strain and axial strain is approximately linear and is the slope of a straight line that is: c (3) where is determined by the FRP effective confinement ratio f lu / fco and is proposed as (Yan & Pantelides 2006): flu (4) fco The effective confinement ratio f lu / fco is defined as the ratio of the ultimate FRP confining pressure flu at rupture of the FRP acket to the unconfined concrete strength f, where f is: co flu 1 FRPE k fu (5) 2 where FRP = FRP volumetric ratio; E = elastic modulus of FRP composite; k = FRP acket efficiency factor; and fu = ultimate FRP tensile strain obtained from material coupon tests. In Eq. (5), k is used to account for the reduction factor of the FRP ultimate hoop strain compared to the material coupon tests and depends on the aspect ratio of the cross-section. Since the FRP composite shells are already post-tensioned prior to axial loading, k is smaller than that of the corresponding bonded FRP ackets. Based on this study, k was found to be in the range of 0.30 and 0.50 for circular nonbonded FRP shells, and a value of 0.40 is recommended; for elliptical cross-sections, k is controlled by the aspect ratio as (Yan et al. 2006): lu

4 k B D (6) 0.65 B D 1.5 Hydraulic Actuator 2.57 m m 0.30 m 7.26 m 7.26 m 2.57 m 3.2 Axial stress versus axial strain relationship The Popovics (1973) model and a modified Willam- Warnke (Willam & Warnke 1975, Yan et el. 2006) model were used for developing the analytical stress-strain relationship for shape-modified specimens with expansive cement concrete. Considering that the confining pressure provided by FRP shells varies continuously and exhibits an approximately linear behavior until failure; the analytical FRPconfined concrete model was implemented based on an incremental approach (Yan & Pantelides 2006) to account for the variable FRP confining pressure by which the axial loading was divided into a number of steps. This incremental approach also utilizes the specific dilatancy behavior of shape-modified columns with expansive cement concrete. The model predictions agree well with the experimental results (Yan & Pantelides 2006) m Concrete filled steel pile 0.91 x Typ. exterior pile cap 2.13 x 0.46 x 0.46 m RC strut 2.74 m As-Built Design Interior pile cap 2.74 x 2.74 m Figure 2. Dimensions and test layout of typical bridge bent at South Temple Bridge in Salt Lake City m 0.53 m 4.57 m (1L) 3.35 m (1L) 3.35 m (1L) 2.74 m 0.46 m (4L) 52.00? 52.00? RC grade beam retrofit RC strut 0.53 m (6L) 0.46 m (3L) 0.46 m (3L) (14L) 7.34 m (a) 4 SIMULATION OF SEISMIC RETROFIT FOR SHAPE MODIFIED SQUARE COLUMNS To investigate the advantage of shape modification, the confinement model and shape modification technique with FRP shells and expansive cement concrete are used in a simulation of seismic rehabilitation of square columns for existing reinforced concrete bridges and are compared to actual in-situ tests of square columns with bonded FRP ackets. The details of a typical bride bent of the South Temple Bridge in Salt Lake City consisting of three columns and a cap beam are shown in Fig. 2. The vertical steel reinforcement of the 914 mm square columns consisted of 16-#10 bars (32 mm diameter) with #4 confinement ties (13 mm diameter) at 305 mm on center. A seismic retrofit utilizing bonded CFRP ackets (Pantelides et al. 2007) was implemented in the field as shown in Fig. 3 and the bent was tested in-situ with a hydraulic actuator that applied a quasi-static load at the bent cap level as shown in Fig. 2. The CFRP composite rehabilitation consisted of confinement layers at the top and bottom of the columns, diagonal and horizontal CFRP sheets at the oints and CFRP straps at the oints. As shown in Fig. 3, each column top has six layers for the first, three layers for the next 0.41 m, and two layers for the next 0.41 m; each column bottom has fourteen layers for the first, three layers for the next 0.41 m, and two layers for the next 0.41 m. The analytical pushover curve for the bridge bent with columns confined with bonded CFRP ackets is 0.36 m strap 0.36 m strap Plan iew of Top of Column U-Strap Top of Bent - Elevation layers (2 layers) U-straps (5 layers) confinement layers (6 layers) U-Strap Detail End iew of Cap Beam (b) Figure 3. CFRP composite retrofit design of typical bent at South Temple Bridge: (a) ankle wrap, zero layers in the cap beam and column ackets, (b) U-strap details (xl = no. of layers). shown in Fig. 4; this pushover curve has been verified with the test results (Pantelides et al. 2007). The alternative CFRP composite retrofit of the bent is simulated here using shape modification and the same number of CFRP layers in a circular shell filled with expansive cement concrete for each top and bottom 1.83 m length of the columns. The shape modified column for the top and bottom 1.83 m of each column has a circular shape of 1.52 m diameter. For comparison purposes, the expansive cement concrete is assumed to be the same material used in the present experimental program. The CFRP composite material properties used for obtaining the push-over curve for the bridge bent with the bonded CFRP-acketed columns are used in the present analysis; the tensile strength of the CFRP composite was 720 MPa, the tensile modulus was 70 GPa, and the ultimate tensile strain was m

5 mm/mm. Due to the insufficient amount of transverse steel reinforcement in the column, the confinement effect from the transverse steel was not considered for developing the stress-strain relations. The resulting analytical curve is also shown in Fig. 4. It is clear from this simulation that the columns confined with CFRP composites show a significant increase in shear and displacement capacity compared to reinforced concrete columns. In this simulation case, it is also noted that for the case of rehabilitation with shape modification and a CFRP shell with expansive cement concrete, the shear capacity of the bridge bent would be increased by approximately 15% compared to the regular retrofit method of the bonded CFRP acket by using a wet layup. Base Shear (x10 3 kn) CFRP Shells + Expansive Cement Without CFRP Bonded CFRP Jackets Displacement (mm) Figure 4. Comparison of pushover curves for bridge bent for columns without CFRP ackets, for columns with bonded CFRP ackets, and for shape modified columns with CFRP shells and expansive cement concrete. 5 CONCLUSIONS Shape modification using expansive cement concrete and prefabricated FRP composite shells can restore the membrane effect; in addition, it can change FRP confinement from passive to active which is induced by the unrestrained expansion of the grout before the column is loaded. Shape-modified columns with expansive cement concrete achieved a higher axial strength for modified square and rectangular columns compared to original columns with the same number of FRP composite layers. The non-bonded FRP acket can be used as a stay-in-place form to save construction time and the significant expense of formwork. In addition, the difference in cost between expansive and Portland cement is small compared to the total cost of the retrofit. For lightly or moderately FRP-confined square or rectangular columns, shape modification can modify the stress-strain behavior from softening to hardening and achieve a higher axial strength and strain. The optimal column cross-sectional shape for FRP confinement is circular. For columns with a rectangular cross-section, especially those with a large aspect ratio, change to a circular cross-section requires a higher cost and significant enlargement of the foundation. Therefore, for strengthening rectangular columns by shape modification, the influence of factors such as: (a) volume increase, (b) increase in surface area, (c) foundation enlargement, and (d) material cost increase need to be considered to obtain an optimal solution. 6 AKNOWLEDGEMENT The authors would like to thank Professor Lawrence D. Reaveley of the University of Utah for his constructive suggestions. The authors acknowledge financial support provided by the Utah Department of Transportation, contribution of FRP composite materials from Sika Inc. and Air Logistics, and contribution of Type K cement and Komponent from CTS Company. The authors also acknowledge Gavin Fitzsimmons for his assistance with the bridge application example in this article. 7 REFERENCES Karbhari,.M. & Gao, Y Composite acketed concrete under uniaxial compression-verification of simple design equations. Journal of Materials in Civil Engineering ASCE 1997; 9(4): Rochette, P. & Labossière, P Axial testing of rectangular column models confined with composites. Journal of Composites for Construction ASCE 2000; 4(3): Pessiki, S., Harries, K.A., Kestner, J.T., Sause, R., Ricles J.M Axial behavior of reinforced concrete columns confined with FRP ackets. Journal of Composites for Construction ASCE 2001; 5(4): Yan, Z., Pantelides, C.P., Reaveley, L.D Fiberreinforced polymer acketed and shape-modified compression members: I-experimental behavior. ACI Structural Journal 2006; 103(6): American Society for Testing Materials (ASTM) Standard Test Method for Tensile Properties of Polymer Matrix Composite Material. ASTM Standards, 15.03, ASTM D3039, West Conshohocken, PA. Yan, Z. & Pantelides, C.P Fiber-reinforced polymer acketed and shape-modified compression members: IImodel. ACI Structural Journal 2006; 103(6): Popovics, S Numerical approach to the complete stressstrain relation for concrete. Cement Concrete 1973; 3(5): Willam, K.J. & Warnke, E.P Constitutive model for the triaxial behavior of concrete. Proceedings of International Association for Bridge and Structural Engineering 1975; 19: Pantelides, C.P., Duffin, J.B., Reaveley, L.D Seismic strengthening of reinforced-concrete multicolumn bridge piers. Earthquake Spectra EERI 2007; 23(3):