Concrete-filled fiber reinforced polymer tube-footing interaction in bending

Transcription

1 Fourth International Conference on FRP Composites in Civil Engineering (CICE2008) 22-24July 2008, Zurich, Switzerland Concrete-filled fiber reinforced polymer tube-footing interaction in bending Y. C. Lai, M. Nelson & A. Z. Fam Queen s University, Kingston, Ontario, Canada ABSTRACT: Concrete-filled fiber reinforced polymer (FRP) tubes (CFFTs) are increasingly becoming popular in various structural applications. A large body of research has been established in the area of member properties of CFFTs, but little work has been done on connection details between CFFTs and other structural members. Such information would be useful for a number of practical CFFT applications, particularly monopoles and bridge piers. This paper examined the case of a moment connection composed of a CFFT embedded into a concrete footing. The objective is to establish the optimal depth of embedment, which allows the CFFT to develop its full flexural resistance without the need to use reinforcing bars. Using laboratory testing of CFFT-footing specimens, this critical embedment depth was determined by increasing the embedment depth until failure occurred by FRP tube rupture instead of debonding. The critical embedment length in this case was found to be 0.73 of the diameter. Additionally, push-off tests were conducted on CFFT stubs embedded in concrete blocks to quantify the bond stress-slip response of the FRP tube against external concrete. It was shown that the slip performance is initially governed by adhesion to concrete, after which it is dominated by friction and a mechanical interlock mechanism at the interface. 1 INTRODUCTION Recent interest in using structural FRP forms for construction has sparked a large body of research into their properties and applications. Bridge piers, monopoles and piles can all benefit from the superior durability and corrosion resistance of FRP structural forms as an alternative to conventional reinforcement. Circular FRP tubes have excellent potential in the concrete filled FRP tubes (CFFTs) applications. The system consists of a prefabricated thin FRP tube, which is then filled partially or entirely with concrete and used as a vertical member to assume axial and lateral loads. The tube could be the sole reinforcement by means of fibers oriented in different directions, or could include some mild steel reinforcement, prestressing, or conventional FRP reinforcement. The FRP tube also serves as a protective jacket for the concrete core, which allows for durability in harsh environments, particularly where traditional reinforced concrete would be highly susceptible to deterioration. Confinement pressure provided by the jackets also benefits the axial strength of CFFTs. Much research has been conducted into the structural properties of CFFTs of different configurations, as well as field applications. Early work in the area of CFFTs was conducted using carbon-frp shells, examining both steel anchorage dowel bar and embedment techniques of the tube into footing (Seible et al, 1996). This work, however, did not investigate the effect of embedment length. Investigation of axial performance of CFFTs has been conducted to asses the contribution of FRP tubes to concrete column strength (Mirmiran and Shahawy, 1997 and Fam and Rizkalla, 2002). Effects of different types of reinforcement and reinforcement ratios on flexural performance have been investigated (Cole and Fam, 2006). Prestressing of CFFTs has been investigated to assess its improvement to flexural stiffness (Fam and Mandal, 2004). In

2 novative solutions for casting CFFTs have also been developed and researched, including spincasting CFFT poles (Yazan, 2007). Despite the wealth of information available on structural properties of CFFT members, little research has been effectuated to examine the performance of connections between CFFTs and other structural members. Research has mainly focused on using steel dowels to connect CFFTs to concrete footings (Zhu et al., 2004). This program aims at studying the concept of a CFFT-footing connection by direct embedment, without connecting bars, using scale laboratory testing. In this case, the CFFT is embedded into a concrete footing by a known depth and moment is resisted by this mechanism alone. Specimens of various embedment lengths have been tested to failure to establish the optimum length, whereby a CFFT tube does not require any additional connection reinforcement to develop its full flexural resistance with no failure in the connection itself. Ancillary push-off tests on CFFT stubs embedded into concrete blocks were also carried out to study the bond-slip behavior of the system. 2 EXPERIMENTAL PROGRAM 2.1 Description of specimens Two distinct types of specimens were prepared for the testing program, one set for the scale testing of the CFFT-footing joint tested in bending and the other being the CFFT-footing push off specimens. All specimens used identical glass-frp (GFRP) tubes having an outside diameter of 220mm and a wall thickness of 4.15mm. The tubes were fabricated using the filament winding method, and are composed of 10 layers with the ratio of hoop to longitudinal fibers be- 500mm 500mm CFFT L (variable) 1100mm Stiff reaction beam 220mm Hydraulic jack Load Cell (a) Bending test Load CFFT Concrete block (500x500mm) 200/400mm (b) Push-off test Figure 1. Schematic of test setups - 2 -

3 ing 1.9. This configuration of tube, although may not be the optimum, is readily available commercially and as such represents an economical alternative to custom fabricated filament wound tubes. All specimens were also cast using the same concrete having aggregate size of 14mm and mean ultimate strength of 42MPa at time of testing. The primary flexure phase of the program consisted of five CFFT-footing specimens, each being an unreinforced CFFT embedded into a concrete footing. The embedment depth was varied in the footings. The footings had dimensions of 500x500x500mm and were reinforced transversely to mitigate cracking. The CFFTs protruded from the block at least 1150mm, for a moment arm of 1100mm for all specimens. Embedment depth was chosen as a function of the pipe diameter (D) as 0.3D, 0.5D, 0.7D, 1D and 1.5D, in order to ensure that the critical embedment depth would be within the range of experimentation. No additional reinforcement (dowels) was added to the connection. The test setup consisted of a horizontal stiff reaction beam with a spacer at one end, on which the footing block rested (Fig.1 (a)). The footing block was then clamped to the reaction beam through the spacer using grade 9 threaded rods and heavy walled HSS sections to ensure a rigid connection. A 30ton 2-way hydraulic ram with a 25000lb load cell was placed at the end of the CFFT (1100mm from the face of the footing) to apply a transverse force on the end of the specimen. This rig achieved an effective cantilever load on the end of the CFFT, allowing the application of moment to the CFFT-footing connection. Figure 2(a) shows a general view of the test setup. Data collected includes end deflection and load from an LPs and the load cell, respectively. Three 5mm 120 ohm electrical resistance strain gauges were used on the FRP tube, 5mm from the face of the footing to establish the compression and tension strains in the tube. Two LPs were also used to measure longitudinal slip between the CFFT and the footing at the extreme tension and compression fibers of the tube at the fixed end. Figure 2. General view of (a) CFFT-footing bending test setup and (b) push-off test setup The secondary phase included six push-off specimens. The specimens consisted of a short stub of CFFT cast into the center of a concrete block (Fig.1 (b)). Three specimens were cast into blocks 200mm deep and another three were cast into blocks 400mm deep to examine two different contact surface areas. In both arrangements, the blocks measured 500mm long by 500mm wide. The CFFT was cast flush with the base of the blocks and protruded at least 25mm from the top. The blocks were reinforced transversely using 4mm wire in a 40mm pitch spiral and the CFFT stubs were entirely unreinforced. The specimens were then tested using a 900kN Riehle load machine using a ramp program at a stroke control of 1mm/min. An internal load cell provided load data and four 100mm linear potentiometers measured the displacement of the top and the bottom of the CFFT stub, relative to the concrete block. The specimens were supported uniformly by two long parallel spacers placed close to the CFFT, to minimize any bending in the block. Data collected included stroke and load from the machine and displacement from the four LPs. Figure 2 shows a general view of the test setup. 2.2 Test results and analysis Results for the secondary push-off tests were transferred from an axial load to a shear stress by dividing the load (N) by the gross contact area (mm 2 ). These stress values could then be plotted against the known displacements to form stress-slip curves (Fig.3). Initial observation showed - 3 -

4 that the shear resistance of the concrete-frp interface was dominated by separate mechanisms. Initially, very little slip occurred as adhesion between the FRP tube and the concrete provided the resistance and the load increased rapidly. The average peak shear resistance of all the specimens (bond strength) was 0.75MPa with a standard deviation of 0.09MPa (Table 1). Table 1. Interfacial bond strength for push-off tests Specimen Bond strength (MPa) P P P P P P Average Standard Deviation Figure 3. Interfacial stress-slip response of push through specimens Upon reaching a peak adhesion stress at the interface, the load dropped instantaneously as the mechanism of resistance changed from adhesion to friction and mechanical interlock. After this point, the load began a cyclic pattern of peaks and troughs in the slip resistance as displacement of the CFFT progressed. A survey of the FRP tube surface was effectuated to explain the cyclic nature of the slip resistance in the push-off tests. A Scan-CONTROL Micro-optic scanning laser was used to create a digital profile of a section of pipe, establishing coordinates on approximately a 1mm by 1mm grid across the surface. A longitudinal cross section of the pipe in the x-z plane (with x being the longitudinal axis) showed a clear geometric wave pattern along the length of the pipe with an amplitude of 0.2mm and a wavelength of 29mm. The wave in the surface exists due to the manufacturing process. The wavelength of the geometric pattern matches that of the slip pattern in the push-off specimens, indicating a direct correlation. It is therefore postulated that an interlock mechanism between the FRP tube and the concrete block produces the cyclic nature of the resistance, as friction increases and decreases as peaks and troughs travel past each other during slippage. The primary flexure tests performed as expected, with the two most deeply embedded CFFTs failing in flexure by rupture of the FRP tube outside the footing and the other three failing by excessive slip at the CFFT-footing interface. Figure 4(a) shows the responses for the specimens with 0.3D, 0.5D and 0.7D embedments, all of which failed by excessive slip, associated with cracking of the concrete footing. As shown in Figure 5(a), the load increased until it reached a peak value in the 0.3D and 0.5D specimens and then dropped. This was followed by a continuing slippage. It is also noted from Figure 4(a) that the peak load increased with em

5 bedment length. Specimen 1D and 1.5D showed similar behavior (Fig. 4(b)) as both failed by rupture of the tube in tension with some slip also taking place (Fig. 5(b)). This slip has contributed to the quasi-plastic behavior noted in Figure 4(b). The 0.7D specimen behaved most similarly to that of the 1D and 1.5D, initially mobilizing increasing slip and eventually deforming quasi-plastically. However, despite the applied load reaching 28kN, the specimen slipped excessively without failing in flexure. This is likely because the level of slip occurring did not allow failure strains to develop in the pipe. The 0.5D specimen initially showed similar stiffness to the 0.7D, 1.0D and 1.5D specimens, up to an applied load of 15kN. At this point, slip was fully mobilized and the resistance of the system began dropping gradually with increasing deflection. The 0.3D specimen was not able to mobilize enough bond resistance due to its shallow depth, and therefore, showed lower stiffness than the other specimens, until adhesion failure occurred at 8kN. The tensile strains at failure were , , , and for the 0.3D, 0.5D, 0.7D, 1.0D and 1.5D specimens, respectively, where tension failure occurred at 1D and 1.5D only. 0.5D 0.3D 0.7D 1.0D 1.5D Figure 4. Load-deflection response of CFFT-footing tests 0.7D 0.5D 0.3D 1.5D 1.0D Figure 5. Load-slip response of CFFT-footing tests 0.73D (Critical embedment) Figure 6. Peak load vs. embedment length showing critical embedment length - 5 -

6 The maximum load resistance of tests 0.3D, 0.5D, 0.7D, 1D and 1.5D were plotted versus the embedment length in order to establish a trend in the data (Fig. 6). The behavior shows an increasing trend which stabilizes at a constant load, at a certain point. A second order polynomial was used to fit the ascending par of the curve. The upper limit load of 29.8kN is plotted on the same graph as a horizontal line that fits the 1D and 1.5D points. The point at which the polynomial intersects the upper limit was taken as the point of critical embedment and was extrapolated as 0.73D. Notably, a specimen with embedment 0.7D was tested in the program and exhibited nearly the maximum moment resistance without failure of the FRP tube, thereby supporting this finding. 3 CONCLUSIONS The following conclusions are drawn: 1. The maximum bond strength between the outer surface of the CFFT tested in this program and concrete was 0.75MPa. This value is based on adhesion. When adhesion fails, excessive slip takes place and lower bond strength is maintained, which is based on mechanical interlock that is dependant on surface texture. 2. The critical embedment length for the CFFT tested in this study was 0.73 times the diameter. Specimens with longer embedment length failed by fracture of the tube in tension, while shorter embedment lengths resulted in excessive slip without a flexural failure. 3. Construction is simplified by avoiding reinforcing bars connecting the CFFT to the footing and using the proposed direct embedment method to develop the moment capacity. REFERENCES Cole, B., Fam, A Flexural Load Testing of Concrete-Filled FRP Tubes with Longitudinal Steel and FRP Rebar. Journal of Composites for Construction, 10(2): Fam, A. and Mandal, S Prestressed Concrete-Filled Fibre-Reinforced Polymer Circular Tubes Tested in Flexure. PCI Journal, 51(4): Fam, A. Z., and Rizkalla, S. H Behavior of axially loaded concrete-filled circular fiber-reinforced polymer tubes. ACI Structural Journal, 98(3): Mirmiran, A. and Shahawy, M Behavior of Concrete Columns Confined by Fiber Composites. Journal of Structural Engineering, ASCE, 123(5), Seible, F., Davol, A., Burgueno, R., Nuismer, R. J., and Abdallah, M. G Structural behavior of concrete filled carbon fiber composite tubular columns. Technology Transfer in a Global Community; Seattle, Washington; USA; 4-7 Nov pp Yazan, Q Flexural behavior of spun-cast concrete-filled fiber reinforced polymer tubes for pole applications. M.Sc. Thesis, Queen s University, Kingston, Canada. Zhu, Z., Mirmiran, A., and Shahawy, M Stay-in-Place FRP Forms for Precast Modular Bridge Pier System. Journal of Composites for Construction, ASCE, 8(6),