STRENGTHENING A PRESTRESSED CONCRETE SLAB BY EPOXY-BONDED FRP COMPOSITES AND SFRC OVERLAYER

Transcription

1 STRENGTHENING A PRESTRESSED CONCRETE SLAB BY EPOXY-BONDED FRP COMPOSITES AND OVERLAYER Joaquim A. O. Barros (1) and José M. Sena Cruz (2) Deartment of Civil Eng., School of Eng., University of Minho, Camus de Azurém, Guimarães, Portugal. (1) barros@eng.uminho.t, (2) jsena@eng.uminho.t Abstract Laminates of carbon fibre reinforced olymer (), layer of steel fibre reinforced concrete () and the simultaneous use of laminates and overlayer were the strengthening strategies analysed for increasing the load bearing caacity of a concrete slab. A numerical model develoed for simulating these strengthening techniques is described and the results obtained are analysed. Keywords: strengthening, carbon fibre reinforced olymer, steel fibre reinforced concrete 1 - Introduction In a building with insufficient suort conditions a textile machine laced on a concrete slab will be relaced by a new one much heavy. To enhance the deficient suort conditions of the building, additional reinforced concrete frames suorted on micro-iles were built nearest as ossible of the slab suorts, in order to disturb, as minimum as ossible, the functionality of the warehouse underneath the slab (see scheme of Figure 1 and hotos in Figure 2). 5. m ~ 7.5 m ~ 7.5 m Slab on the grade So il Slab to strength Existing mortar brick masonry suorting the slab Concrete frames for adding extra suort to the slab Soil weekly comacted Micro-iles 1- m Rive r Bed rock Figure 1 Strategy for adding extra suorts to the slab. Figure 2 Frames suorting the slab.

2 For increasing the slab load bearing caacity the three reinforcing techniques, schematically reresented in Figure were analysed: alying eoxy-bonded laminates of carbon fibre reinforced olymer () [1]; using a to over-layer of steel fibre reinforced concrete () [2]; combining the two last ones. The numerical model develoed for designing the strengthening techniques alied on the slab will be described in the resent work, and the results obtained will be analysed. Casted concrete Casted concrete Casted concrete Ceramic brick Pre-casting re-stressed beams (a) (b) (c) Figure Techniques for increasing the slab load bearing caacity: (a) using ; (b) alying a to overlayer of ; (c) using the two last strategies. 2 Numerical model 2.1 Introduction The resent model can be alied on a cross section of any shae and submitted to an axial force and two bending curvatures. The cross section can be comosed of different concrete, steel and comosite materials. To be useful for strengthening design, the model can simulate materials with a given initial strain and stress. The finite element mesh discretizing the cross section is deendent on the geometry and on the materials comosing this cross section (see Figure 4). The moment-curvature relationshis are obtained taking into account the material constitutive laws and the equilibrium and the cinematic equations. Aroriated constitutive laws simulate the monotonic and the cyclic behaviour of the aforementioned materials. Steel law θ z Unconfined concrete law θ y Confined concrete law ε x Figure 4 Cross section model reresentation.

3 2.2 - Materials Plain concrete (PC): The monotonic comression behaviour is simulated by the exressions roosed by the CEB-FIP Model Code []. The confinement rovided by conventional stirrus or hoos is taken into account using the model develoed by Scott et al [4]. The formulation roosed by Thomson and Park [5] was used for governing the unload-reload branches. U to cracking, the PC tensile behaviour is simulated by the Young modulus and the axial tensile strength, whereas the ost-cracking behaviour is modelled by the fracture arameters [6]. The tensile stiffness of the unload-reload branches of the cracked PC is considered constant and equal to the tensile stiffness of the uncracked PC. Steel fibre reinforced concrete (): Previous work [6] has shown that the comlete stress-strain exressions roosed for lain concrete cannot adequately fit the ost-eak resonse of the fibre concrete. Based on exerimental research, a new exression was roosed [6] that was imlemented in the resent model. The energy absortion caacity of the cracked concrete is the roerty most benefited by adding fibres to a concrete mix. This roerty can be evaluated from three oints bending tests, determining the fracture energy of the material [6,7]. A strain softening law was imlemented in the numerical model for simulating the tensile ost-cracking behaviour of. This law was resented elsewhere [8]. Reinforced cracked concrete: A tension-stiffening model was develoed for simulating the ostcracking behaviour of concrete (or ) reinforced with conventional rebars. This model takes into account the main roerties of the conventional reinforcement and the fracture arameters of the lain concrete (or ). This model was described in another work [8]. Conventional reinforcement: Steel rebars are modelled using the laws roosed by Menegotto and Pinto []. Laminates of carbon fibre reinforced olymer: A linear stress-strain relationshi was imlemented for modelling the linear-elastic brittle behaviour of the laminates. - Model araisal To assess the model erformance, the design resistant moment of the slab cross section reresented in Figure 5 will be determined and comared with the data furnished by the slab s sulier, included in Table 1. 4 mm Ceramic brick Casted concrete Ceramic brick 2 mm G Figure 5 A modulo of the slab. Pre-casting re-stressed beams Figure 6 Wires of the beams. The slab is comosed by re-casting re-stressed beams of C25/ concrete and ceramic bricks. According to the slab sulier, the values on Table 1 were obtained considering a casted concrete of class C2/25. Table 2 includes the characteristics of the reinforcement on the re-casting re-stressed concrete beams (see also Figure 6).

4 Thickness (mm) Table 1 Slab characteristics according to the sulier. Limit states Reference of the Ultimate Serviciability Slab reference Total Above the brick re-stressed beam Dead load (kn/m 2 ) M Rd (knm/m) V Rd (kn/m) M fctk (knm/m) EI (knm 2 /m) 2S4-22B S Table 2 Characteristics of the steel wires of the re-casting re-stressed concrete beams (see Fig. 6). Wire number Diameter (mm) Pre-stress at long time (MPa) 1, 2 e Figure 7 reresents the cross-section aroach adoted for design uroses. The corresonding four-nodded finite element mesh used is shown in Figure 8. S 45 4 Casted concrete C16/ G S' Pre-casting re-stressed C25/ concrete beams Figure 7 Geometry assumed for the numerical simulation. Figure 8 Four-nodded finite element mesh. The re-stress on wires indicated on Table 2 introduces on the concrete re-casting beams the stress field reresented in Figure. To evaluate these stresses it was assumed a design comression strength of.85 f cd =14.2MPa (C25/) for the concrete, and the finite element mesh reresented in Figure 1. In the bottom surface the comression stress is about 1MPa, less than the concrete design comression strength. To evaluate the comression strength of the casted concrete alied in the slab, six cylinder-drilled cores of 1mm diameter and 1mm length were tested in a comressive load frame. Using the recommendations roosed by ASTM [1] the average comression strength and the standard deviation were f cms =27.4MPa and s =.1MPa, resectively. To evaluate the characteristic and design values it was used the exression roosed by RILEM [11] f ck s t1 s t1 = f s cms n (1) fcms n where n is the number of secimens and t 1 is the value for the student distribution at 1% fractile, which for n=6 is Using these data it was obtained a f ck =2.MPa and a f cd =1.5MPa, which corresonds to a C2/25 concrete class. However, taking into account that shrinkage cracks were observed in some drilled cores (see Figure 11) it was considered the C16/2 concrete class for design uroses. For the maximum comression strain allowed in concrete it was considered a value of.5, in agreement with the CEB-FIP Model Code. Assuming the concrete of the re-casting re-stressed beams has the initial stress/strain field reresented in Figure and E =1GPa, f ud =1715/1.15=141MPa and f.1d =14/1.15=126MPa for the design constitutive law of the steel wires, the moment-curvature relationshi, reresented in

5 Figure 12, was obtained, from which it can be concluded that the model estimates quite well the design resistant moment, M Rd, given in the technical data of the slab s sulier (see Table 1). In all the numerical simulations it was neglected the concrete tensile contribution. The M Rd value was conditioned by the maximum strain limit in the steel wires (see figure inset of Figure 12). Strain () Element Figure Comression stress on concrete of the re-casting re-stressed beams Stress (MPa) Figure 1 Finite element mesh for the re-casting re-stressed beam (.225; 11.8) Moment (knm/m) 5 (.15; 86.41) ( f.1k /γ s) σ (f uk/γ s) Corresonding to M Rd 25 σ ε =1 ε ε Figure 11 Some drilled cores had shrinkage cracks Figure 12 Moment-curvature relationshi for the slab. 4 Numerical model alied for strengthening the slab 4.1 Introduction Any strengthening strategy will be alied when the slab is submitted to its dead load and to the revetments. For these loads the design moment (M Sd ) is 16kNm/m. Figure 1a illustrates the strains and the stresses in the slab cross-section for this loading regime. Figure 1b gives some suort for understanding the stress ath observed in the re-casting re-stressed concrete beams. For this urosed the layers 1 and 2 were considered. Under the re-stress loading the stress in these layers is σ ' C1 and σ ' C 2, resectively (see Figure 1b). When the slab dead weight and the slab revetments are actuating, these two concrete layers unload, following the branches reresented in Figure 1b. Since the

6 strain increment in layer 2 is higher than in layer 1, and the concrete tensile strength is neglected, the residual comression strength in layer 1, σ C1, is higher than the stress in layer 2, σ C 2. S 12 σ' c2 Layer 1 εc1 σc1 Stress (MPa) 6 σ' c1 Layer 1 Layer 2 Layer 2 S' εc Strain (1/1) σc Stress (MPa) ε' c2 σc2 εc1 ε' c1.2 εc Strain (1/1) (a) (b) Figure 1 Strain and stress field for the slab dead load and revetments. σc1 4.2 Using laminates of carbon fibre reinforced olymer () Comaratively to reinforcing traditional materials, the advanced comosite materials have higher tensile strength, higher resistance to the environmental aggressiveness, are much more lightweight and more easy to aly. In some alications, the higher costs of the comosite materials can be comensated by a significant decrease in the time sent on the strengthening rocedures. Another imortant asect is the marginal changing in the geometrical configuration of the element to be reaired when comosite materials are used. In the bottom surface of each re-casting re-stressed concrete beams a laminate of 5mm width and 1.2mm thick was glued using an eoxi comound. According to the technical data furnished by the sulier [12] the Young modulus and the tensile strength of the laminates roosed are E L =15GPa and f ul =2MPa, resectively. To avoid the occurrence of the eeling henomenon [1] in these comosites, the design stress was limited to the half of its ultimate tensile strength. For the roerties of the other materials it was used the values reviously indicated. Figure 14 reresents the moment-curvature relationshi for the slab reinforced with laminates. The design resistant moment, M Rd, is 111kNm/m, which is higher than the design moment of the loading actions (M Sd =16kNm/m). The M Rd value was conditioned by the concrete comression strain limit (.5 ). For this M Rd the stress in laminates is 616MPa (see Figure 15), which is less than the maximum allowed tensile stress (f Ld =f Lu /2) and the tensile stress in the highest strained steel wire is 18MPa, less than the design ultimate stress (see Figure 16). 4. Using a to overlayer of The increase on the slab load bearing caacity rovided by the revious strengthening strategy was limited by the allowed maximum concrete comression deformability. To retrieve the high tensile strength of the laminates without occurring the crushing of the casted concrete, a mm thick C2/25 to overlayer of Kg/m of Dramix ZP5 steel fibres [14] was alied. This layer was bonded to the casted concrete using metallic connectors fixed by an eoxy comound. These connectors were designed for resisting to the shear stresses between these two concrete layers. From Figure 14 it can be observed that the resistant moment is 112kNm/m, higher than the M sd, and the resonse is much more ductile (see also Figure 17). This last observation can also be taken from the relationshi between curvature and the comression stress at the to surface, illustrated in Figure 18. The M Rd value was conditioned by the strain limit in the steel wires. A total strain of 1. ( ε o + ε =. +1 ) was attained for M Rd =112kNm/m, which corresonds to a stress of σ =1416MPa.

7 2 (.85; ) 12 (.85; 75.) 15 Moment (knm/m) 1 5 (.525; ) SF RC Stress in (MPa) 6 (.25; 615.7) (.25; 11.85) Figure 14 - Moment-curvature relationshis for the reinforcing strategies. Stress in steel wire number 2 (MPa) (.85; 157.7) (.25; 17.76) (.525; ) Figure 16 Relationshi between the stress in the highest strained wire and the curvature Figure 15 Relationshi between the stress in the laminates and the curvature. Stress in the to surface of casted concrete (MPa) 12 6 (.85; 8.74) (.25; 8.71) (.525; 8.65) + Figure 17 Relationshi between the curvature and the comression stress at the casted concrete to surface. 4.4 Using and Alying simultaneously laminates and a overlayer the ultimate moment is increased significantly (see Figure 14). The maximum moment was conditioned by the maximum stress allowed in laminates, but the comression strain limit at overlayer was almost attained (.2 ), and the strain at to surface of the casted concrete is also significant (2. ), indicating that this strategy leads to a well rofit of the all intervening materials. 5 Conclusions Three strengthening strategies for increasing the load bearing caacity of a concrete slab were analysed using a numerical model develoed for this urose. The first strategy was based on gluing laminates on the re-casting re-stressed C25/ concrete beams. The design resistant moment (M Rd ) is 5% higher than the design loading moment (M Sd ), and its value is conditioned by the comression strain at the to surface of the casted concrete. In the second strategy a mm thick to overlayer of was

8 alied. In this case the M Rd is also 5% higher than the M Sd but it is conditioned by the strain limit in the wires, being the resonse much more ductile. The third strengthening strategy is a conjugation of the last two strategies, resulting a M Rd 5% higher than the M Sd, limited by the stress at laminates. For this M Rd, the comression strain at to surface of the casted concrete and at to surface of the overlayer is near their limit values and the strain at the highest strained wire is a little bit above its limit, which indicates to be an effective reinforcing strategy. However, for the loads acting on the slab and taking into account economical reasons, the strategy based on overlayer was roosed to solve the roblem. Figure 1 shows the asect of the slab after to be strengthened. 12 Stress in the to surface of (MPa) 6.85; 11.1) (.525; 11.2) Figure 18 Relationshi between the curvature and the comression stress at to surface. Figure 1 View of the concrete slab after has been strengthened using a over-layer. References [1] ACI 44R-6, State-of-the-art reort on fiber reinforced lastic reinforcement for concrete structures, reorted by ACI Committee 44, American Concrete Institute, Farmington Hills, Michigan, 68. [2] ACI 544.1R-6, State-of-the-art reort on fiber reinforced concrete, ACI Manual of Concrete Practice Part 5 ACI International, 17, 66. [] CEB-FIP Model Code 1. Comite Euro-International du Beton, Bulletin d Information nº 21/214, Ed. Thomas Telford, (1). [4] Scott, B.D.; Park, R.; Priestley, M. J. N. (182), Stress-strain behavior of concrete confined by overlaing hoos at low and high strain rates, Journal of the American Concrete Institute, January-February. [5] Thomson, K.J.; Park, R. (18), Moment-curvature behaviour of cyclically loaded structural concrete members, Journal Structural Division, Proc. Inst. Civ. Engrs. Part 2, June, [6] Barros, J.A.O., Figueiras, J.A. Flexural behavior of steel fiber reinforced concrete: testing and modelling, Journal of Materials in Civil Engineering, ASCE, Vol. 11, Nº 4, 1-, 1. [7] Barros, J.A.O., Sena Cruz, J. Fracture energy of steel fibre reinforced concrete, Journal of Mechanics of Comosite Materials and Structures, Vol. 8, No , January-March 21. [8] Barros, J.A.O., Figueiras, J.A. Nonlinear analysis of steel fibre reinforced concrete slabs on grade, Comuters & Structures, Vol.7, No.1,. 7-16, January 21. [] Menegotto, M.; Pinto, P. (17), Method of analysis for cyclically loaded R.C. lane frames including changes in geometry and non-elastic behavior of elements under combined normal force and bending, Symosium on Resistance and Ultimate Deformability os Structures Acted on by Well Defined Reeated Loads, IABS Reorts Vol. 1, Lisbon. [1] ASTM Committee C-, Subcommittee C.61, Standard test method for obtaining and testing drilled cores and sawed beams of concrete, ASTM,.24-27, 1. [11] RILEM TC 162 TDF, Materials and Structures, Vol., 75-81, March 2. [12] Technical data sheet of MBRACE Laminates, LM5 1.2, 2. [1] Rostásy, F.S., Assessment of the suitability of CRP lates from the S&P CRP system for use as adhesive-bonded reinforcement to strengthen concrete constructional elements and bases of assessment for their general aroval by the construction suervisory authorities, exert oinion nº 8/22, S&P Reinforcement, TU Braunschweig, 46., 18. [14] Technical data sheet of Dramix ZP5 steel fibres, N.V. Bekaert S.A., 18.