- Email: [email protected]
Early-age restrained shrinkage cracking of GFRP-RC bridge deck slabs: Effect of environmental conditions
Accepted Manuscript Early-Age Restrained Shrinkage Cracking Of Gfrp-Rc Bridge Deck Slabs: Effect Of Environmental Conditions Amir Ghatefar, Ph.D. Candidate, Ehab El-Salakawy, Professor and Canada Research Chair in Durability and Modernization of Civil Structures, M.T. Bassuoni, Associate Professor PII:
S0958-9465(15)30012-3
DOI:
10.1016/j.cemconcomp.2015.07.010
Reference:
CECO 2534
To appear in:
Cement and Concrete Composites
Received Date: 27 December 2014 Revised Date:
22 April 2015
Accepted Date: 28 July 2015
Please cite this article as: A. Ghatefar, E. El-Salakawy, M.T. Bassuoni, Early-Age Restrained Shrinkage Cracking Of Gfrp-Rc Bridge Deck Slabs: Effect Of Environmental Conditions, Cement and Concrete Composites (2015), doi: 10.1016/j.cemconcomp.2015.07.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
EARLY-AGE RESTRAINED SHRINKAGE CRACKING OF GFRP-RC BRIDGE DECK SLABS: EFFECT OF ENVIRONMENTAL CONDITIONS
RI PT
Amir Ghatefar1, Ehab El-Salakawy2* and M. T. Bassuoni3
Abstract: Most of codes and guidelines for glass fiber reinforced polymers (GFRP) - Reinforced Concrete (RC) are based on modifying corresponding formulas, originally developed for steel
SC
bars, taking into account the differences in properties and behavior between FRP and steel. The main objective of this research is to investigate the effect of cyclic environments on early-age
M AN U
cracking of GFRP-RC bridge deck slabs experimentally. Two full-scale (measuring 2500-mm long × 765-mm wide × 180-mm thick) cast-in-place slabs reinforced with similar amounts of reinforcement ratio of 0.7% with GFRP and steel bars, respectively, were tested in adiabatic laboratory conditions as control specimens. In comparison, two other GFRP-RC deck slabs were
TE D
tested under freezing-thawing and wetting-drying conditions. The test results are presented in terms of materials degradation, cracking pattern, crack width, and spacing, and strains in reinforcement and concrete. Test results indicate that the minimum reinforcement ratio (0.7%)
EP
recommended by the Canadian Highway Bridge Design Code 2006 (CHBDC 2006) for bridge deck slabs reinforced with GFRP bars satisfied the serviceability requirements after being
AC C
subjected to the simulated cyclic exposures. Key words: GFRP, Deck slabs, Early-age cracking, Wetting-drying cycles, Freezing-thawing cycles, Serviceability 1
Ph.D. Candidate, Department of Civil Engineering, University of Manitoba, Winnipeg, Manitoba, Canada, E-mail: [email protected] 2* Professor and Canada Research Chair in Durability and Modernization of Civil Structures, Department of Civil Engineering, University of Manitoba, Winnipeg, Manitoba, Canada, R3T 5V6, Tel: (204) 4748319 Fax: (204) 474-7513, E-mail: [email protected] (corresponding author)* 3 Associate Professor, Department of Civil Engineering, University of Manitoba, Winnipeg, Manitoba, Canada, Email: [email protected]
1
ACCEPTED MANUSCRIPT
INTRODUCTION If the volumetric change of concrete due to shrinkage and thermal stresses is restrained, tensile stresses will develop in concrete. If the induced tensile stresses are higher than the tensile
RI PT
capacity of the concrete, the concrete will crack (TRB 2006). The restraint can be internal, from reinforcement and aggregate, or external, from the sub-base or continuity of elements. If volumetric instabilities are not uniform throughout a member, as produced by thermal or
SC
humidity gradients, the member itself can serve as a restraint (Frosch et al. 2003). The magnitude of induced tensile stresses depends on both the amount of volumetric changes and the degree of
M AN U
restraint, i.e. how much movement is restricted (Hadidi and Saadeghvaziri 2005). Bridge deck slabs are typically much longer in one direction than the other, thus volumetric changes due to shrinkage and temperature changes are more pronounced in the longitudinal direction, which induces tensile stresses resulting in transverse cracks.
TE D
Recently, the non-corrodible fiber reinforced polymer (FRP) bars have been used as reinforcement for bridge deck slabs, barrier walls, and parking garages to mitigate the corrosion problem of conventional steel. Among different types of FRP materials, the lower cost of glass
EP
FRP (GFRP) bars makes them attractive for the construction industry. Nevertheless, since GFRP bars have a lower modulus of elasticity than steel, concrete elements reinforced with GFRP
AC C
exhibit larger deformation leading to wider cracks. Unlike steel, GFRP materials do not corrode by nature; however, GFRP-RC are still susceptible to other forms of deterioration due to harsh environments involving de-icing chemicals, sulfate salts and alkalis, which can readily penetrate concrete through cracks (ISIS Canada 2006). This serviceability issue can be aggravated by harsh environmental conditions such as freezing-thawing and wetting-drying cycles. A maximum crack width of 0.5 mm is recommended by CHBDC (CSA S6 2006) for FRP-RC
2
ACCEPTED MANUSCRIPT
components subjected to aggressive environments and 0.7 mm for other members. In the field, transportation elements such as bridge deck slabs and barrier walls are subjected to temperature and humidity changes due to daily or seasonal conditions. Typically, due to the low tensile
RI PT
strength of concrete during the first 24 hours after casting, thermal changes from hydration reactions and moisture loss from the surface of concrete, due to hot, dry and windy conditions, increase the tendency to cracking (Byard et al. 2010). After hardening, ambient temperature and
SC
humidity fluctuations can induce volume instability of concrete. Cracks are the easiest locations for moisture and aggressive chemicals to accelerate the deterioration of reinforcement as well as
M AN U
to reduce the service life of concrete structures. While non-pre-stressed reinforcement cannot completely eliminate cracking, it can control crack spacing and width in bridge deck slabs, thus achieving a key serviceability requirement for RC structures under different environmental conditions.
TE D
PREVIOUS WORK
From an experimental study on concrete panels reinforced with one layer of GFRP bars, Koenigsfeld and Myers (2003) concluded that the equation listed in ACI-440.1R-03 (ACI
EP
Committee 440 2003) [same as in the most recent ACI Committee 440 2006] for minimum FRP reinforcement ratio was overly conservative; however, they did not recommend any alternate
AC C
equation to calculate the minimum FRP reinforcement ratio. Koenigsfeld and Myers (2003) found three times larger crack widths for GFRP-RC panels (1830 mm × 591 mm × 127 mm) than that of their counterpart specimens with similar amounts of steel reinforcement when subjected to restraint and daily cyclic temperature between 21°C to 41°C to simulate pouring concrete on a hot summer day. They also concluded that twice as much GFRP reinforcement as steel is required to achieve similar crack control limits when subjected to flexural loading. Due to lower
3
ACCEPTED MANUSCRIPT
stiffness of GFRP bars, lower internal tensile stresses in concrete will develop due to internal restraint against concrete shrinkage or temperature variations, which leads to larger crack spacing followed by wider crack widths (Chen and Choi 2002).
RI PT
The authors of the current paper previously concluded that the minimum GFRP reinforcement ratio of 0.7% recommended by CHBDC for bridge deck slabs can reasonably control early-age crack width and reinforcement strain under normal laboratory conditions (22 ± 2 °C and 50-
SC
70 % RH) (Ghatefar et al. 2014, Marti-Vergas 2015, Ghatefar et al. 2015). However, bridge deck slabs are usually exposed to harsh environmental conditions such as freezing-thawing cycles,
M AN U
temperature fluctuations and wetting-drying cycles within temperature and humidity ranges from –40°C to +35°C and 30 % to 100 %, respectively (Laoubi et al. 2006). Therefore, at early ages, these structural elements may be subjected to different combinations of shrinkage and swelling. In RC structures, damage due to severe environmental conditions can take various forms such as
TE D
reinforcement de-bonding, scaling, and micro cracking (Bishnoi 2004 and Alves et al. 2011). Other forms of damage include large-scale spalling and crumbling of concrete and material fatigue, resulting in loss of strength and stiffness.
EP
During thermal variations, self-equilibrating stresses develop in FRP due to the difference in coefficients of thermal expansion of constituent elements (fibers and resin), resulting in matrix
AC C
micro-cracking. Further deterioration can occur due to expansion of absorbed moisture in FRP and concrete under freezing-thaw cycling. In addition, freezing-thaw cycles can lead to degradation of the fiber-matrix bond and further damage the fibers through local notching due to ice formation on their surfaces (El-badry et al. 2000). GFRP can experience an expansion of 4-6 times greater than that of concrete in the transverse direction due to temperature variations. This thermal incompatibility can cause bursting stresses to build up in the concrete surrounding the
4
ACCEPTED MANUSCRIPT
reinforcement or deboning of bars from concrete under temperature changes (Gentry et al. 1999). Moreover, under freezing-thawing cycles, ice formation at the interface between FRP and concrete leads to further loss of the FRP bond to concrete and increases existing crack width
RI PT
under sustained loads (Alves et al. 2011). Similarly, wetting-drying cycles are considered critical in the durability-based design of concrete structures since volume changes due to repetitive shrinkage/swelling may lead to material fatigue and de-bonding of reinforcement (Zhang et al.
SC
2012, Ayano et al. 2002). While numerous studies have investigated the structural behavior of FRP-RC elements (Ghatefar et al. 2014, Koenigsfeld and Myers 2003), scarce data on the
M AN U
restrained shrinkage cracking in FRP-RC elements have been reported under different environmental conditions. Hence, further research is needed in this area. This contribution presents an experimental study evaluating the effect of longitudinal (secondary) GFRP reinforcement ratio on early-age cracking of bridge deck slabs subjected to freezing-thawing and
TE D
wetting-drying cycles.
EXPERIMENTAL PROGRAM
This study included four full-size, cast-in-place deck slab prototypes; three reinforced with
EP
GFRP and one reinforced with steel. The slabs were designed according to Section 16 of the CHBDC (Clause 16.8.8.1) with a total reinforcement ratio of 0.7% to investigate the effect of
AC C
different environmental conditions on early-age cracking for a period of 112 days after casting. To determine the internal conditions of the cementations matrix, the dynamic modulus of elasticity (DME) was determined for all cores from the ultrasonic pulse velocity (UPV) measurements according to ASTM C597. Also, to evaluate the interconnectivity of the pore structure in the concrete slabs subjected to different environmental conditions, the rapid chloride penetrability test (RCPT) was performed for the cores according to ASTM C1202 (Standard Test
5
ACCEPTED MANUSCRIPT
Method for Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration) in comparison to non-exposed specimens at the same age (112 days). After the RCPT, the specimens were axially split and sprayed with a silver nitrate solution, which forms a white
RI PT
precipitate of silver chloride in approximately 15 minutes, to measure the physical penetration depth of chloride ions. The average depth of the white precipitate was determined at five different locations along the diameter of each half specimen. This depth is considered to be an
SC
indication of the ease of ingress of chloride ions, and thus the connectivity/deterioration of the microstructure (Bassuoni et al. 2006).
M AN U
Material Properties
Normal-strength, ready-mixed concrete incorporating 13% silica fume by mass of binder (Table 1) with a target 28-day compressive strength of 40 MPa was purposefully used to provoke high tendency for shrinkage as an extreme scenario that might be encountered in practice. The slump
TE D
and fresh air content of this concrete were in the ranges of 100-120 mm and 6±1%, respectively. The US-Federal Highway Administration Guidelines [Silica fume user manual (Holland 2005)] reports on the use of 4 to 15% silica fume in concrete for various infrastructure applications. For
EP
GFRP and steel reinforcement, sand-coated GFRP bars (Pultrall Inc. 2012) and normal steel grade 40, respectively, were used to reinforce the slab prototypes in both layers (top and bottom).
AC C
The GFRP bars are made of continuous E-glass fibers with modified vinyl-ester resin. The longitudinal and transverse coefficients of thermal expansion for the GFRP bars used are 6.2 and 23.8 [×10-6/°C], respectively, while the longitudinal and transverse coefficients of thermal expansion for the steel bars used are 11.7 [×10-6/°C]. Table 2 summarizes the mechanical properties of the steel and GFRP bars used in this research, according to standard laboratory tests (CSA S806-12 and ASTM A400 – 69 -12, respectively).
6
ACCEPTED MANUSCRIPT
Test Setup and Prototypes According to Clause 14.13.1.2 of CHBDC (CSA 2006), the minimum allowable thickness of bridge deck slabs is 175 mm. Therefore, in this study, a thickness of 180 mm for all test slabs
RI PT
was selected. A total of four full-size, cast-in-place concrete bridge deck slabs (2500-mm long by 765 mm wide as shown in Fig. 1) were constructed and tested in the laboratory. The effective width-to-length ratio was selected less than 1/3 to ensure that the amount of shrinkage in the
SC
longitudinal direction is much more than that in the transverse direction; there was three times as much concrete that tended to shrink in the longitudinal direction than in the transverse direction
M AN U
(Frosch et al. 2003). Consequently, a transverse crack was expected to develop in the transvers direction in order to relieve the larger tensile stress in the longitudinal direction. The reinforcement configuration of the test specimens was selected based on the empirical design method recommended by Section 16 - Clause 16.8.8.1 of the CHBDC (CSA 2006).
TE D
According to this section, the minimum FRP reinforcement ratio in the longitudinal bottom and top assemblies is 0.35% with top and bottom covers equal to 35±10 mm (CHBDC, Clause16.4.4). All test prototypes had similar bar size, top and bottom clear covers (25 and 30
EP
mm, respectively) and a constant spacing of 255 mm for the longitudinal reinforcement. Slabs G1, G1-FT, and G1-WD (G: GFRP-RC, FT: under freezing-thawing, WD: under wetting-drying
AC C
conditions) were reinforced with the minimum GFRP reinforcement ratio (0.7 %) while specimen S1 (S: steel-RC) contained similar amounts of steel reinforcement (0.7 %). Prototypes G1 and S1 were tested under normal laboratory conditions (22 ± 2 °C and 50-70 % RH), while slabs G1-FT and G1-WD were subjected to freezing-thawing and wetting-drying cycles respectively at 7 days (Table 3). Dimensions and details of the slabs’ cross-section are shown in Fig. 2.
7
ACCEPTED MANUSCRIPT
Instrumentation To measure strains in the GFRP bars in the vicinity of the first crack, three 6-mm electrical resistance strain gauges [FLA-6-11 series (accuracy: ±1%)] were attached to each bar at the top
RI PT
and bottom layers; one centered at the mid-span, and the other two at 50 mm on each side as shown in Fig. 3. In addition, to measure concrete strain two types of strain gauges were used; embedment strain gauges [EGP series (accuracy: ±2%)] in concrete and 10-mm linear pattern
SC
strain gauges [20CBW series (accuracy: ±1%)] on the surface of the slabs. For each slab, one strain gauge was embedded at the cracking (mid-length) location. The other strain gauge was
M AN U
attached to the surface of concrete at an arbitrary distance of 270 mm (1.5 times slab thickness) away from the mid-length (cracking location) to avoid gauge damage upon the occurrence of first cracking. The internal strain gauge was used to capture the development of tensile strains within concrete up to failure by first cracking, while the surface strain gauge measured the
TE D
deformation of concrete in the vicinity of the cracking location during the entire period of exposure. The width of the cracks developed was recorded throughout the test using two PIgauges. Also, the internal relative humidity was monitored by humidity sensors [Rapid RH®
EP
4.0EX series (accuracy:±3%)] embedded at the level of the reinforcement layers and mid-depth at an arbitrary distance of 625 mm away from the mid-length (cracking location) to avoid further
AC C
stress concentration at cracking place. All instrumentation (except humidity sensors) was connected to a DAQ (data acquisition system) controlled by a computer. Test Procedure
Prior to casting, the 2500 mm long slabs were effectively anchored at its ends by 1473 × 1000 × 1200 mm concrete blocks, which were clamped (pre-stressed) to the laboratory strong floor using 38-mm diameter dywidag bars (Fig. 4). Afterwards, the inside surface of the formwork
8
ACCEPTED MANUSCRIPT
was cleaned and thinly coated with a releasing agent (oil) to prevent adhesion of the concrete. Also, a total of 63 cylinders (150 × 300 mm) were cast using the same concrete as the prototypes. For the first 24 hours after casting, a plastic tent was built around the test prototypes
RI PT
and cylinders while electrical heaters were used to maintain the internal concrete temperature at 35 °C without moist curing. Subsequently, the tent and formwork was removed, then PI-gauges were attached to the concrete surface, and initial strain measurements were recorded. During the
SC
first 24 hours, the ambient conditions around the slab (under tent) were 40°C with 30-40 % RH. The average internal temperature and RH measured at the top reinforcement level were 35 °C
M AN U
and 90 %, respectively. After the tent was removed, the specimens were left in laboratory conditions (22±2 °C and 50-70 % RH) for 7 days. During that period, the internal temperature at the top reinforcement level was 22±2 °C while the RH decreased from 90 to 70 %. After 7 days, freezing-thawing and wetting-drying cycles were applied for specimens G1-FT and G1-WD,
TE D
respectively. Using the concrete cylinders, the average compressive (ASTM C39M-03) and splitting tensile (ASTM C496M-04) strengths and the modulus of elasticity (ASTM C469-02) were obtained after 1, 3, 7, 14, 21 and 28 days (Table 4) to determine the development of
EP
concrete properties (within a standard deviation of 5%). At the mid-length of each slab, the cross-section was locally reduced (notched) to 565 × 150 mm
AC C
to control the first crack occurrence at this location (Fig. 5). The bottom surface of the slab was supported by three stay-in-place smooth greasy plates (300 mm × 300 mm, spaced at 1250 mm as shown in Fig. 6) to reduce the effect of slab’s self-weight on the reinforcement strains. Also, fresh water was poured into the surface reservoir (approximately 5-mm thick) created by peripheral foam dykes at the outer edge of the slabs G1-FT and G1-WD for freezing-thawing
9
ACCEPTED MANUSCRIPT
and wetting cycles. Concrete surface strain gauges were damaged for the slab G1-FT in the first freezing-thawing cycle. Environmental Conditioning Schemes
RI PT
Freezing-thawing cycles
Different freezing-thawing conditioning schemes had been used by researchers to study the behavior of RC elements externally or internally reinforced with FRP composites (Laoubi et al.
SC
2006 and Alves et al. 2011). In the current research, the temperature profile of Standard Test Method for Resistance of Concrete to Rapid freezing-thawing (ASTM C-666 M-03 2008) was
M AN U
adopted. In this standard, the freezing-thawing cycles consist of alternately lowering the temperature from +4 to -18ºC for freezing and raising it from -18 to +4 ºC for thawing. Thawing time should not be less than 25% of the total freezing-thawing time. In order to reach the standard conditions in the bottom reinforcement level of the slab, the applied freezing-thawing
TE D
cycles consisted of alternately decreasing the environmental chamber temperature from +22 to 25ºC for freezing and raising it from -25 to +35ºC for thawing at a rate of 1.55 cycles/day to achieve the ASTM temperature and duration requirements at the level of GFRP reinforcement.
EP
Figure 7 shows the reading of thermocouples embedded in specimen G1-FT at the level of bottom reinforcement compared to the air temperature inside the chamber. Specimen G1-FT was
AC C
subjected to 163 freezing-thawing cycles over 105 days. Figure 8 (a) indicates that ponded water (3-5 mm) on the top of slab G1-FT increased the average internal humidity to approximately 99 % (i.e. beyond the critical saturation level of 90%). Wetting-drying cycles
It should be noted that there are no standard test methods for the wetting-drying exposure of concrete. The cyclic regime and the total number of cycles (five cycles) in this study were
10
ACCEPTED MANUSCRIPT
selected similar to that adopted by Zhang et al. (2012) to achieve significant humidity changes at the reinforcement level of the bridge deck slabs. Each wetting-drying cycle started with 14 days of drying at 35±2°C and 30% RH followed by 7 days of wetting at 22±2°C and 100% RH.
RI PT
Figure 8 (b) shows the reading of humidity sensors embedded in the test specimen at the different levels of the cross section compared to the humidity in the chamber. RESULTS AND DISCUSSION
SC
In this section, the experimental results are primarily presented in terms of concrete cracking pattern, width, and spacing, and strains in the reinforcement and concrete, with links to the
M AN U
materials tests. Finally, the main experimental results from this study are compared to predictions from a theoretical model for RC restrained members that are not subjected to significant bending (Gilbert 1992). General Behavior
TE D
The width of cracks in the restrained slabs varied according to the environmental exposure and type of reinforcement. While the magnitude of crack width depends on several factors such as degree of restraint, quality of bond between concrete and reinforcement, size and distribution of
EP
bars, concrete quality and ambient conditions, the variables tested in this program were the environmental conditions and properties of reinforcement. For each slab, the crack width was
AC C
considered as the average of the measured value at two locations across the slab width at midlength. Generally, the first crack in all specimens was observed within the first three days after casting in the transverse direction before exposure. Figure 9 shows the cracking pattern for all the slabs at the notched (mid-length) location. The cracks, which usually extended into the full depth of slabs, typically occurred at mid-length. The internal concrete strain at cracking location was measured by the embedment concrete strain
11
ACCEPTED MANUSCRIPT
gauge up to cracking. Since concrete cracking damaged the internal strain gauge, the surface strain gauges, located at a distance of 270 mm from the cracking location, were used to measure concrete strain after cracking. Figure 10 shows the internal strain of the concrete up to cracking,
RI PT
for specimens G1, S1, G1-FT and G1-WD. The measured internal tensile strains of concrete were 300, 270, 280 and 310 µε, respectively, when the early-age cracks became visible. Once the crack formed at mid-length (crack location), the stress in the concrete dropped to zero. For all
SC
specimens after cracking, the concrete surface strain changed to a compressive strain [negative values in Figs. 11 (c), 12 (e) and 13 (c)]. When cracked, the stiffness of the slab in the vicinity of
the crack shortened elastically.
M AN U
a crack reduced, depending on the cross-section axial stiffness, and the concrete on either side of
The strain in the reinforcing bars was presented as the average strain readings of all instrumented bars (top and bottom) in the vicinity of the crack at mid-length. Figures 11, 12 and 13 indicate
TE D
that, once the crack formed at mid-length, the reinforcement carried the full restraining force for the slabs G1, S1, G1-FT and G1-WD, consequently, the average reinforcement strains suddenly increased to 1450, 680, 1270, 1370 µε, respectively. In the meantime, the average reinforcement
EP
strains, before cracking, were 290, 285, 275 and 300 µε, respectively. Normal laboratory exposure
AC C
The first crack occurred for slabs G1 and S1 within 48 ad 62 hours, respectively. Figure 11 (a) shows the change in crack width over the test period under normal laboratory conditions. The final crack width in slab G1 was 0.33 mm, while in slab S1 it reached 0.18 mm after 112 days of casting. This was expected due to the lower section stiffness of the GFRP-RC slab compared to that of the steel-RC slab. Reduction of the reinforcement modulus of elasticity increased the crack width as well as average crack spacing. In slab G1, the average crack spacing was 2500
12
ACCEPTED MANUSCRIPT
mm whereas in slab S1 it reduced to 833 mm (the second and third cracks occurred in S1 after 19 days on both sides of the first crack, Fig. 9). An increase in the reinforcement modulus of elasticity leads to less stiffness reduction at first cracking (mid-length), therefore the restraining
RI PT
force after cracking remains relatively high resulting in further cracking. Generally with a higher restraining force, due to further drying shrinkage or any ambient temperature variation, the concrete in regions away from the first crack tends to experience further cracking.
SC
Figure 11 (b) shows that the average strain in the steel bars (S1) in the vicinity of the first crack was 410 µε after 112 days under laboratory conditions. Comparatively, the strains in the GFRP
M AN U
bars of specimen G1 reached 1520 µε, because of the lower section stiffness in the GFRP-RC slab. This complies with the crack width trend as increasing the reinforcement modulus of elasticity from 62 to 200 GPa, the surface strain of concrete reduced by approximately 28%. Correspondingly, the concrete surface compressive strains for slabs S1 and G1 were 210 and 290
TE D
µε, respectively (Fig. 11 (c)) Freezing-thawing exposure
The behavior of restrained concrete elements under freezing-thawing conditions is affected by
EP
multiple variables (e.g. internal water expansion, and material contraction due to low temperature). It is well documented that if concrete elements are critically saturated (internal RH
AC C
> 90%), the water volume expansion phenomenon in larger capillary pores induces considerable volume changes of concrete during freezing. At the onset of ice crystallization, the frictional resistance to ice growth creates internal pressure in the pores leading to concrete expansion (Scherer et al. 2002). In addition, ice formation in the void space imbibes water from the smaller (gel) pores, creating negative (suction) pressure in the matrix and thus contraction (Towers and
13
ACCEPTED MANUSCRIPT
Helmuth 2008). Hence, the total volume change of concrete is a combination of expansion and contraction from hydraulic and osmotic pressures. The transverse full-depth crack occurred for slab G1-FT within 47 hours (before exposure) at the
RI PT
mid-length (notched section). Figure 12 (a) shows the change in crack width in this slab over 163 freezing-thawing cycles (112 days). Crack width reached to its maximum and minimum point at +4°C (thawing) and -18°C (freezing), respectively, in each cycle. At the last cycle (163),
SC
the crack width varied between 0.29 and 0.42 mm corresponding to the freezing-thawing stages, respectively (Fig. 12 (b)). The lower crack width recorded during freezing periods can be
M AN U
attributed to the volumetric expansion of the critically saturated slab, which led to partial closure of the crack opening. Upon relieving the expansion pressure during thawing periods, the crack width increased up to 0.42 mm (in the last cycle), which is 40% and 27% higher than the crack width measured before the freezing-thawing exposure and in the normal exposure (slab G1),
TE D
respectively. Also, it was observed that slab G1-FT suffered from moderate surface scaling (less than 0.5 kg/m2 at the end of exposure (according to BNQ 2002 test standard)), which is a typical damage manifestation of concrete exposed to freezing-thawing cycles. This trend might be
EP
ascribed to over finishing the surface of slab G1-FT, which led to reducing the volume of air entrainment in the surface as shown by the Scanning Electron Microscopy (SEM) analysis.
AC C
For slab G1-FT, Fig. 12 (c) shows the average fluctuation of the strains in the reinforcement at cracking after 163 freezing-thawing cycles. In the last cycle, the strains of the reinforcement at the cracking location were 1020 and 1690 µε during the freezing-thawing stages, respectively (Fig. 12 (d)). Complying with the crack width results, this trend is attributed to the repetitive volumetric change associated with frost action as discussed earlier.
14
ACCEPTED MANUSCRIPT
Wetting and drying exposure Wetting-drying conditions may significantly affect RC elements due to the variation of moisture distribution with depth and accelerated shrinkage during drying periods. For slab G1-WD, the
RI PT
relative humidity of the concrete surface markedly changed during wetting-drying cycles relative to the inner core, which led to further deformations (Fig. 8 (b)). Fig. 13 (a) shows the change in crack width for slab G1-WD which was subjected to 5 wetting-drying cycles over 112 days. In
SC
the last cycle (22°C with 100% RH and 35°C with 30% RH) the crack width varied between 0.23 and 0.46 mm. The lower crack width recorded during the wetting periods can be attributed to
M AN U
swelling of the slab due to the increase in relative humidity, which led to partial closure of the crack opening. Subsequently, excessive drying of the slab increased the crack width up to 0.46 mm (in the last cycle), which is 0.92% and 0.39% higher than that crack width measured before the wetting and drying exposure and in the normal exposure (slab G1 after exposure for 112
TE D
days), respectively.
Corresponding to the crack width trend, the strain in the GFRP bars increased from 1400 (wetting portion) to 2250 µε (drying portion) in the last cycle (Fig. 13 (b)), due to the additional
EP
drying shrinkage deformation under hot-arid conditions, which was restrained by the GFRP reinforcement. It should be noted that this value is significantly higher than the maximum strains
AC C
recorded in the normal (1520 µε) and freezing-thawing (1690 µε) exposures. This behavior was confirmatory to the measurements of concrete surface strain that increased up to 480 µε due to the additional shrinkage during the drying portion (Fig. 13 (c)). Materials Tests:
In additional to the early-age cracking of restrained concrete slabs exposed to harsh conditions, they were vulnerable to material degradation especially at the mid-length (in the vicinity of the
15
ACCEPTED MANUSCRIPT
crack) due to temperature and humidity variations. To capture this trend, three materials tests: Dynamic Modulus of Elasticity (DME), Rapid Chloride Permeability Test (RCPT), and Scanning Electron Microscopy (SEM) were conducted on cores extracted from the three GFRP-
RI PT
RC slabs. A total of twelve 100-mm diameter cores were extracted; four cores from each slab (close and away from left and right sides of the crack). All cores were extracted from slabs at the end of the test period (112 days). For the DME test, full length cores were used, while for the
SC
other two tests, the top 50-mm thick slices were cut from the cores extracted from different locations in the slabs.
M AN U
UPV test (Ultrasonic Pulse Velocity test)
Table 5 shows the dynamic modulus of elasticity results for the unexposed concrete (cores from specimen G1 under laboratory conditions) and after being subjected to freezing-thawing and wetting-drying conditions (cores from G1-FT and G1-WD). While the test results were in the
TE D
narrow range of 45-50 GPa, they showed a general reduction (maximum of 10%) of DME for the concrete exposed to cyclic conditions, which indicates the existence of fissures and micro-cracks in the cementitious matrix.
EP
RCPT test (Rapid Chloride Permeability test) After operating the RCPT for 6 hours according to ASTM C1202 (Standard Test Method for
AC C
Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration), the penetration depth was measured in the concrete specimens. The whitish color of the penetration depth was clearly visible as depicted in Fig. 14, and the results are listed in Table 5. In contrast to the freezing-thawing exposure, Table 5 shows that the specimens extracted from the slab subjected to wetting-drying cycles yielded relatively higher penetration depths in the vicinity of the crack location, which signifies that the pore structure was highly interconnected in
16
ACCEPTED MANUSCRIPT
these specimens. This can be attributed to a higher intensity of micro-cracks due to the matrix fatigue resulting from high strain fluctuations of repetitive swelling and shrinkage (Fig. 13 (c)) in the wetting-drying exposure. These results are consistent with the higher concrete and
RI PT
reinforcement strain values for the spacemen G1-WD exposure to drying conditions (Fig. 13 (b), and (b)). Microstructural analysis
SC
To supplement the results of UPV and RCPT, the alteration of microstructure of concrete was also assessed by backscattered scanning electron microscopy (BSEM) on thin sections from the
M AN U
cores extracted from G1-FT and G1-WD in the vicinity and away from the main crack. The polished sections were prepared from fracture surfaces that were dried at 40°C for 24 h, impregnated with low-viscosity epoxy resin under pressure, cut, polished and carbon coated. The SEM micrographs show that the specimen subjected to wetting-drying conditions
TE D
particularly in the vicinity of the main crack (Fig. 15 (a)) had higher intensity of micro-cracks and internal damage than that of the concrete exposed to freezing-thawing cycles (Fig. 15 (b)). This trend is consistent with the RCPT test, the higher recorded concrete and reinforcement
EP
strain values in the vicinity of the crack, and the crack width for the specimen under wettingdrying conditions (Fig. 13 (b) and (c)).
AC C
Theoretical vs. experimental results Limited studies provided formulas to predict cracking characteristics of RC deck slabs and stress distribution in bars at cracking locations due to restrained shrinkage. In Gilbert’s 1992 theoretical analysis, it was explained that shrinkage causes an axial force built-up (Eq. 1) in restrained members, which leads to direct tension cracks. Equations 2 and 3 were proposed to calculate the stress in the bars, in the vicinity of the crack, and the crack width, respectively. In this model, the
17
ACCEPTED MANUSCRIPT
restraint is provided to the longitudinal movement caused by shrinkage and temperature changes. Table 6 illustrates input values for parameters used in Equations 1 to 3.
∗ =
. ∗ . ∆
'(
+∗ ,∗
∗ .
Eq. 2
.
) = −[
−
./ − 01 2 + # ∗ $ 0]
!
+ # ∗ $ % ∗ &
Eq. 1
RI PT
∞ =
Eq. 3
SC
Considering the fully restrained member in direct tension as the concrete shrinks, the restraining force gradually increases until the first crack occurs, which is usually within two days from the
M AN U
commencement of drying. Immediately after the first cracking, the restraining force reduces and the concrete stress away from the crack is less than the tensile strength of the concrete. The concrete on either side of the crack shortens elastically and the crack opens to a width, w. At the crack location, the reinforcement carries the entire force (the concrete strain drops to zero. In the
TE D
region immediately adjacent to the crack the concrete and the reinforcement stresses vary considerably, and a region of partial bond breakdown exists. At a distance so from the crack, which was earlier proposed by Favre et al. (1983) for a member containing deformed bars or
EP
welded wire mesh (Eq. 1), the concrete and the reinforcement stresses are no longer influenced
AC C
directly by the presence of the crack. It was suggested that the value of so to be multiplied by 1.33 to achieve better predictions for RC member with steel bars (Gilbert 1992). Ghatefar et al. (2014) concluded that the coefficient of 0.8 instead of 1.33 for so led to better predictions for FRP-RC members.
In order to calculate the width of crack and reinforcement stress in the vicinity of the main crack (under freezing-thawing conditions) by Gilbert’s equations, the total strain of the concrete was calculated based on the combination of concrete contraction due to temperature reduction to 18
ACCEPTED MANUSCRIPT
18°C (using the measured concrete thermal coefficient of 5.5×10−6/℃), and the expansion (up to 9%) of water in large capillary pores (greater than 100 nm) upon freezing at critically saturated conditions. Also, to evaluate the crack width and reinforcement stress in the vicinity of the main
calculate concrete strain under wetting-drying conditions.
RI PT
crack in specimen G1-WD by Gilbert’s equations, ACI 209.2R-08 guidance was used to
For the specimen subjected to freezing-thawing conditions, the measured crack widths, using
SC
Gilbert’s model (Gilbert 1992), were 0.28 and 0.31 mm for freezing and thawing, respectively. The values of the crack width agree with the results of the analytical model with an error of
M AN U
approximately 20%. Also, the reinforcement strains in the vicinity of the crack were comparable to those obtained from the theoretical model (1205 and 1845 µε during freezing and thawing, respectively) within 13% error.
For the specimen subjected to wetting-drying conditions, the measured widths of cracks, using
TE D
Gilbert’s 1992 theory, were 0.33 and 0.43 mm respectively. The values for crack width within first wetting-drying cycle agree with the results of the analytical model with an error of approximately 13%.
While For the specimen exposed to wetting-drying conditions, the
EP
measured strains in the GFRP bars, using Gilbert’s 1992 model, were 1889 and 568 µε, respectively. The values for GFRP reinforcement strain within the first wetting-drying cycle
AC C
agree with the results of the analytical model with an error of approximately 17%. CONCLUSIONS
Based on the experimental variables and environmental conditions implemented in the current research, the following conclusions can be drawn:
19
ACCEPTED MANUSCRIPT
1. Due to the relatively lower modulus of elasticity of GFRP bars, the crack width and average reinforcement strain in slab G1 were larger than those of the counterpart slab reinforced with similar steel reinforcement ratio (S1).
RI PT
2. Under normal laboratory conditions, increasing the reinforcement modulus of elasticity from 62 GPa (GFRP) to 200 GPa (steel) reduced the average crack spacing. An increase in the reinforcement modulus of elasticity leads to less stiffness reduction at first cracking
SC
(mid-length), therefore the restraining force after cracking remains high resulting in further cracking.
M AN U
3. Under freezing-thawing conditions, the crack width reached its maximum and minimum points at +4°C (thawing) and -18°C (freezing), respectively, in each cycle. This behavior is attributed to the volumetric expansion of the critically saturated slab during freezing, which led to partial closure of the crack opening. Upon relieving the expansion pressure
TE D
during thawing periods, the crack width increased up to 0.42 mm (in the last cycle), which is 40 and 27% higher than the crack width measured before the freezing-thawing exposure and in the normal exposure (slab G1), respectively.
EP
4. In the specimen under wetting-drying conditions, the lower crack width and reinforcement strain recorded during the wetting periods can be attributed to swelling of
AC C
the slab due to the increase in relative humidity, which led to partial closure of the crack opening. Subsequently, excessive drying of the slab increased the crack width and reinforcement strain in the drying period.
5. The UPV test results show a general reduction of DME for the concrete exposed to cyclic conditions, which indicates the existence of fissures and micro-cracks in the cementitious matrix. While RCPT and microstructural analysis indicate that the specimens extracted
20
ACCEPTED MANUSCRIPT
from the slab subjected to wetting-drying cycles yielded relatively higher penetration depths and more internal micro cracks in the vicinity of the crack location, which signifies that the pore structure was highly interconnected in these specimens. This can be
RI PT
attributed to a higher intensity of micro-cracks due to the matrix fatigue resulting from high strain fluctuations of repetitive swelling and shrinkage in the wetting-drying exposure. These results are consistent with the higher concrete and reinforcement strain
SC
values for specimen G1-WD exposed to wetting-drying conditions.
6. Test results indicate that the minimum reinforcement ratio (0.7%) recommended by
M AN U
CHBDC for bridge deck slabs reinforced with GFRP bars is conservative under laboratory condition. However, this ratio satisfied the serviceability requirements of the CHBDC (crack width of 0.5 mm and 5800 µε, which represents 25% of GFRP ultimate strain) after being subjected to the simulated exposures of freezing-thawing and wetting-
TE D
drying cycles. The maximum measured crack width and GFRP strains did not exceed 0.46 mm and 2250 µ, respectively.
7. The measured width of shrinkage cracks and stresses in GFRP agreed with most results
EP
from Gilbert analytical model (Gilbert 1992) within 17% error. However, more refinement to Gilbert’s model is still needed to be fully applicable to FRP-RC slabs under
AC C
different environmental conditions (especially for structures reinforced with different rebar diameter and spacing), which is recommended for future research.
Finally, similar to the case of normal laboratory conditions, it can be inferred that a corresponding steel-RC slab, subjected to wetting-drying and freezing-thawing cycles, would achieve the same trend but with lower crack width and strain in reinforcement due to the higher modulus of elasticity of steel.
21
ACCEPTED MANUSCRIPT
ACKNOWLEDGMENTS The financial support provided by the Natural Science and Engineering Research Council of Canada (NSERC) through Canada Research Chairs (CRC) program and the Network of Centers
RI PT
of Excellence on Intelligent Sensing for Innovative Structures (ISIS Canada) are gratefully appreciated. Also, the efforts of A. Ghazy and the technical staff of the McQuade Heavy Structures Laboratory and IKO Construction Materials Testing Facility at the University of
AC C
EP
TE D
M AN U
SC
Manitoba are greatly acknowledged.
22
ACCEPTED MANUSCRIPT
NOTATION Areinf.: total area of bars in a section (mm2)
Ec (day): concrete modulus of elasticity at different age (MPa) E*e: effective concrete modulus of elasticity (MPa)
fc (day): concrete compressive strength at different age (MPa) ft (day): concrete tensile strength at different age (MPa)
l: slab length (mm) m: number of cracks N (∞): final tensile force (kN)
M AN U
ft (reinf): tensile strength of reinforcement bars (MPa)
SC
Ereinf.: reinforcement modulus of elasticity (GPa)
RI PT
db: bar diameter (mm)
TE D
5∗ : reinforcement modulus of elasticity divided by concrete effective modulus of elasticity S: average crack spacing (mm)
/1 : reinforcement stress transfer length at cracking location (mm)
EP
t: slab thickness (mm)
w: final cracking width (mm)
AC C
∆u: support displacement (mm)
ε*sh: ultimate shrinkage strain (using ACI 209.2R-08) ρreinf.: reinforcement ratio
σav: estimate of the average concrete stress in the period after first cracking (MPa) σ*c1: final concrete stress away from the crack (MPa) ϕ*: creep coefficient (according to ACI 209.2R-08)
23
ACCEPTED MANUSCRIPT
REFERENCES ACI Committee 440. 2006. “Guide for the Design and Construction of Structural Concrete Reinforced with FRP Bars”. ACI 440.IR-06, American Concrete Institute (ACI), Detroit, MI.
RI PT
Alves, J., El-Ragaby, A. and El-Salakawy, E. 2011. “Durability of GFRP Bars Bond to Concrete under Different Loading and Environmental Conditions”. Journal of Composite for Construction, ASCE, Vol. 15, No. 3, pp. 249-269.
SC
ASTM C666 M-03. 2008. “Standard Test Method for Resistance of Concrete to Rapid Freezing and Thawing”. ASTM International, US.
M AN U
ASTM C597. 2009. “Standard Test Method for Pulse Velocity through Concrete,” ASTM International, US.
ASTM C1202. 2012. “Standard Test Method for Electrical Indication of Concrete's Ability to Resist Chloride Ion Penetration,” ASTM International, US.
TE D
ASTM A400. 2012. “Standard Practice for Steel Bars, Selection Guide, and Mechanical Properties,” ASTM International, US.
Ayano, T. and Wittmann, F.H. 2002. “Drying, moisture distribution, and Shrinkage of cement-
EP
based materials”. Journal of Materials and Structures, Vol. 35, No. 247, pp. 134-140. Bassuoni M. T., Nehdi M. L. and Greenough, T. R. 2006. “Enhancing the Reliability of
AC C
Evaluating Chloride Ingress in Concrete Using the ASTM C 1202 Rapid Chloride Penetrability Test,” Journal of ASTM International, Vol. 3, No. 3, p. 13. BNQ NQ 2621-900. 2002. “Determination of the Scaling Resistance of Concrete Surfaces Exposed to Freezing-and-Thawing Cycles in the Presence of De-icing Chemicals,” Bureau de normalisation du Québec, Annex A, pp. 19-22. Move up, alphabetical order!!!
24
ACCEPTED MANUSCRIPT
Byard, B. E., Schindler, A. K., Barnes, R. W. and Rao, A. 2010. “Cracking Tendency of Bridge Deck Concrete”. Journal of the Transportation Research Board, Washington, D.C., No. 2164, pp.122-131.
RI PT
CSA. 2006. “Canadian Highway Bridge Design Code - CHBDC”. CAN/CSA-S6-06, Canadian Standards Associations, Rexdale, Ontario.
CSA. 2012. “Design and Construction of Building Structures with Fibre-Reinforced Polymers”.
SC
CSA/S806-12, Canadian Standards Association. Toronto, Ontario, Canada.
Favre, R. 1983. “Fissuration et deformations.” Manual du Comité Euro-Internationale du Beton
M AN U
(CEB). Ecole Polytechnique Federale de Lausanne, Swizerland, pp. 249. Frosch, R. J., Blackman, D. T. and Radabaugh, R. D. 2003. “Investigation of Bridge Deck Cracking in Various Bridge Superstructure Systems,” Joint Transportation Research Program, E-port No: FHWA/IN/JTRP-2002/25, Indiana Department of Transportation and
TE D
Purdue University, West Lafayette, Indiana.
Gilbert, R. I. 1992. “Shrinkage Cracking in Fully Restrained Concrete Members”, ACI Structural Journal, Vol. 89, No. 2, pp.141-149.
EP
Ghatefar, A., El-Salakawy, E. and Bassuoni, M.T. 2014. “Effect of Reinforcement Ratio on Transverse Early-Age Cracking of GFRP-RC Bridge Deck Slabs”. Journal of Composite for
AC C
Construction, ASCE, Vol. 18, No. 6, 04014018, 9 p. Ghatefar, A., El-Salakawy, E. and Bassuoni, M.T. 2015. “Closure to the discussion on the paper: Effect of Reinforcement Ratio on Transverse Early-Age Cracking of GFRP-RC Bridge Deck Slabs”. Journal of Composite for Construction, ASCE, Vol. 19, No. 1, 070140001, 1 p. Hadidi, R. and Saadeghvaziri, M. A. 2005. “Transverse Cracking of Concrete Bridge Decks: State-of-the-Art”. Journal of Bridge Engineering, ASCE, Vol. 10, No. 5, pp. 503-510.
25
ACCEPTED MANUSCRIPT
Holland, T.C. 2005. “Silica Fume User’s Manual.” US Federal Highway Administration Technical report, Washington DC. U.S. ISIS Canada. 2006. “Durability of Fiber Reinforced Polymers in Civil Infrastructure”. Durability
RI PT
Monograph, Winnipeg, Manitoba, Canada: ISIS Canada Research Network.
Krasus, P. D. and Rogalla, E. A. 1996. “Transverse Cracking in Newly Constructed Bridge Decks”. NCHRP Report 380, Transportation Research Board, National Research Council,
SC
Washington, D.C.
Laoubi, K., El-Salakawy, E. and Benmokrane, B. 2006. “Creep and durability of sand-coated
M AN U
glass FRP bars in concrete elements under freeze/thaw cycling and sustained loads”. Journal of Concrete and Composites, Vol. 10, No. 28, pp. 869-878.
Marti-Vargas, J.R. 2015. “Discussion of the paper: Effect of Reinforcement Ratio on Transverse Early-Age Cracking of GFRP-RC Bridge Deck Slabs by Amir Ghatefar, Ehab El-Salakawy
p.
TE D
and M. T. Bassuoni” Journal of Composite for Construction, ASCE, Vol. 19, 070140001, 1
Pultrall Inc. 2012. “V-RODTM-Technical data sheet.” ADS Composites Group Inc., Thetford
EP
Mines, Quebec, Canada. http://www.pultrall.com (visited on October 15, 2012). Scherer, W., Chen, J. and Valenza, J. 2002. “Method of Protecting Concrete from Freeze
AC C
Damage.” U.S. Patent No. 6,485,560. TRB. 2006. “Control of Cracking in Concrete-State of the Art”. Transportation Research Circular E-C107, Transportation Research Board of the National Academic, Basic research and emerging technologies related concrete committee, NY.
26
ACCEPTED MANUSCRIPT
Towers, T.C. and Helmuth, R.A. 2008. “Theory of Volume Changes in Hardened PortlandCement Paste during Freezing"., American Concrete Institution (ACI). Special Publication, SP-249, pp. 141-160.
RI PT
U.S. Army Corps of Engineers. 1981. “Test Method for Coefficient of Linear Thermal Expansion of Concrete” CRD C 39–81, issued 1 June 1981. Available at: http://www.wes.army.mil/SL/MTC/handbook/crd_c39.pdf
SC
Zhang, J., Gao, Y., Han, Y. and Sun, W. 2012. “Shrinkage and Interior Humidity of Concrete under Dry–Wet Cycles”. International Journal of Drying Technology, Vol. 30, No. 6, pp.
AC C
EP
TE D
M AN U
583-596.
27
ACCEPTED MANUSCRIPT
Table 1: Proportions of concrete per cubic meter
365 kg
Coarse aggregate
1020 kg
(max. aggregate size, 20 mm) 650 kg
High-range water reducer
2500 ml
Air entraining agent
250 ml
Silica Fume
54.6 kg
M AN U
Fine aggregate
Water
170 kg
EP
TE D
GU = General use.
AC C
*
SC
Cement Type GU*
RI PT
Quantity/m3
Ingredient
1
ACCEPTED MANUSCRIPT
Table 2: Mechanical properties of GFRP and steel bars Bar area in each layer (top and bottom)
Modulus of
strength
tensile
(mm2)
elasticity
(MPa)
strain (%)
Bar size (mm)
(GPa) Nominal
CSA S806-12 Annex A
15.9
197.8
282.7
62
16
200
------
200
1450
SC
GFRP, No. 16
AC C
EP
TE D
M AN U
Steel, 15M
Ultimate
RI PT
Bar Dia.
Tensile
2
590
2.33
ɛy = 0.2
ACCEPTED MANUSCRIPT
Table 3: Test Matrix Areinf*
ρ**
(mm2)
(%)
S1
400
0.35
1.2
Normal laboratory conditions (22°C and 50 to 60% RH)
G1
382
0. 35
1.0
Normal laboratory conditions (22°C and 50 to 60% RH)
G1-FT
382
0. 35
1.0
Freezing-thawing cycles (-18°C to +4°C)
G1-WD
382
0. 35
1.0
RI PT
Environmental Conditions
M AN U
Wetting-drying cycles ( 22°C to 35°C and 100 to 30 % RH)
The area and ratio of longitudinal bars per layer (top and bottom)
TE D
Reinforcement spacing: 255 mm
EP
**
AC C
*
ρ/ρmin
SC
Prototypes
3
ACCEPTED MANUSCRIPT
Table 4: Compressive, tensile and E-modulus test results for concrete under different environmental conditions Ambient N
N
N
N
W/D
F/T
N
W/D
F/T
N
W/D
F/T
Age (days)
Tensile Strength (MPa) Modulus of Elasticity (GPa)
3
7
14
14
14
21
21
21
28
28
28
7
14
34
35
37
33
36
38
32
38
41
35
1.3
3.4
3.5
3.9
3.3
3.6
3.1
3.7
3.9
3.3
18
21.1
21.2
21.2
21.8
22.2
21
0. 6 -
SC
Compressive Strength (MPa)
1
3.7
M AN U
Property
RI PT
conditions
21.8
21.3
21.1
21.8
N: Normal laboratory conditions (22°C and 50-60% RH),
W/D: Wetting and drying cycling (22°C to 35 °C and 100 to 30 % RH respectively), and
AC C
EP
TE D
F/T: Freezing and thawing cycling (-18 °C to +4 °C and 100% RH).
4
ACCEPTED MANUSCRIPT
Table 5: DME and RCPT results
Wetting and Drying
(mm)
C-R
50
7
C-L A-R A-L
49 52 51
C-R C-L A-R A-L C-R C-L A-R A-L
48 49 46 48 48 48 47 45
RI PT
Freezing and Thawing
(GPa)
Cores
6 5 6
SC
Normal (un-exposed)
Average Penetration Depth
M AN U
Exposure
Dynamic Modulus of Elasticity
AC C
EP
TE D
C: Center, A: Away from center, L: Left, R: Right
5
8 8 7 8 9 12 8 8
ACCEPTED MANUSCRIPT
Table 6: Input data for parameters used in equations 1 to 3 db
(mm2)
(mm)
S1
800
16
-1.71e-4
3.7
38
21800
200000
fy:= 400
G1
764
15.9
-1.71e-4
3.7
38
21800
68297
1450
G1-WD
764
15.9
3.9
41
22200
68297
1450
3.1
35
ε*sh
ft (28)
fc (28)
E c (28)
EGFRP
ft (reinf)
(MPa)
(MPa)
(MPa)
(MPa)
(MPa)
-2.23e-4(dry)
-1.07e-4(thaw) G1-FT
764
15.9
21200
M AN U
-0.83e-4(freeze)
SC
-1.09e-4(wet)
RI PT
Areinf. Slab
68297
AC C
EP
TE D
l = 2500 mm, t = 180 mm, ϕ* = 0.6, ft (7) = 3.4 MPa, fc(3) = 25 MPa, Ec(3) = 20200 MPa,
6
1450
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 1: General view of the test setup and specimen (all dimensions are in mm).
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Fig. 2: Deck slab dimensions – (a) side view, (b) top view, and (c) cross-section A-A (all dimensions are in mm).
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 3: Typical instrumentation of a deck slab (all dimensions are in mm).
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 4: Slab ends effectively held in position and restrained against translation.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 5: Mid-length details.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 6: Smooth supports at the bottom surface of the slabs to reduce self-weight effect
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 7: A part of the freeze/thaw profile for specimen G1-FT.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 8: Relative humidity readings for slab prototypes (a) G1-FT and (b) G1-WD.
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Fig. 9: Crack pattern in the specimens at the notched location [S1: Steel-RC under laboratory conditions, G1: GFRP-RC under laboratory conditions, G1-FT: GFRP-RC under freeze/thaw conditions, G1-WD: GFRP-RC under wet/dry conditions] (all dimensions are in mm).
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 10: Internal strain of concrete up to cracking.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 11: Control slabs (G1, S1) under normal laboratory conditions: (a) crack width development, (b) development of the bar strains, and (c) surface strain of the concrete in the vicinity of the main crack.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 12: Slab G1-FT under freeze/thaw conditions: (a) crack width development, (b) crack width development in the slab during first and last cycles, (c) reinforcement strain development, (d) development of strains in the reinforcement during first and last cycles, and (e) surface strain of the concrete in the vicinity of the main crack.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 13: Slabs G1-WD under wetting-drying conditions: (a) crack width development, (b) development of the bar strains, and (c) surface strain of the concrete in the vicinity of the first crack.
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 14: Chloride penetration depth in cores extracted from: (a) slab G1-FT close to the crack area, (b) slab G1-FT away from the crack area, (c) slab G1-WD close to the crack area, and (d) slab G1-WD away from the crack area.
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 15: Typical SEM micrographs from: (a) specimen G1-WD (slab under wetting-drying conditions), and (b) specimen G1-FT (slab under freezing-thawing conditions) in the vicinity of the main crack.
S0958-9465(15)30012-3
DOI:
10.1016/j.cemconcomp.2015.07.010
Reference:
CECO 2534
To appear in:
Cement and Concrete Composites
Received Date: 27 December 2014 Revised Date:
22 April 2015
Accepted Date: 28 July 2015
Please cite this article as: A. Ghatefar, E. El-Salakawy, M.T. Bassuoni, Early-Age Restrained Shrinkage Cracking Of Gfrp-Rc Bridge Deck Slabs: Effect Of Environmental Conditions, Cement and Concrete Composites (2015), doi: 10.1016/j.cemconcomp.2015.07.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
EARLY-AGE RESTRAINED SHRINKAGE CRACKING OF GFRP-RC BRIDGE DECK SLABS: EFFECT OF ENVIRONMENTAL CONDITIONS
RI PT
Amir Ghatefar1, Ehab El-Salakawy2* and M. T. Bassuoni3
Abstract: Most of codes and guidelines for glass fiber reinforced polymers (GFRP) - Reinforced Concrete (RC) are based on modifying corresponding formulas, originally developed for steel
SC
bars, taking into account the differences in properties and behavior between FRP and steel. The main objective of this research is to investigate the effect of cyclic environments on early-age
M AN U
cracking of GFRP-RC bridge deck slabs experimentally. Two full-scale (measuring 2500-mm long × 765-mm wide × 180-mm thick) cast-in-place slabs reinforced with similar amounts of reinforcement ratio of 0.7% with GFRP and steel bars, respectively, were tested in adiabatic laboratory conditions as control specimens. In comparison, two other GFRP-RC deck slabs were
TE D
tested under freezing-thawing and wetting-drying conditions. The test results are presented in terms of materials degradation, cracking pattern, crack width, and spacing, and strains in reinforcement and concrete. Test results indicate that the minimum reinforcement ratio (0.7%)
EP
recommended by the Canadian Highway Bridge Design Code 2006 (CHBDC 2006) for bridge deck slabs reinforced with GFRP bars satisfied the serviceability requirements after being
AC C
subjected to the simulated cyclic exposures. Key words: GFRP, Deck slabs, Early-age cracking, Wetting-drying cycles, Freezing-thawing cycles, Serviceability 1
Ph.D. Candidate, Department of Civil Engineering, University of Manitoba, Winnipeg, Manitoba, Canada, E-mail: [email protected] 2* Professor and Canada Research Chair in Durability and Modernization of Civil Structures, Department of Civil Engineering, University of Manitoba, Winnipeg, Manitoba, Canada, R3T 5V6, Tel: (204) 4748319 Fax: (204) 474-7513, E-mail: [email protected] (corresponding author)* 3 Associate Professor, Department of Civil Engineering, University of Manitoba, Winnipeg, Manitoba, Canada, Email: [email protected]
1
ACCEPTED MANUSCRIPT
INTRODUCTION If the volumetric change of concrete due to shrinkage and thermal stresses is restrained, tensile stresses will develop in concrete. If the induced tensile stresses are higher than the tensile
RI PT
capacity of the concrete, the concrete will crack (TRB 2006). The restraint can be internal, from reinforcement and aggregate, or external, from the sub-base or continuity of elements. If volumetric instabilities are not uniform throughout a member, as produced by thermal or
SC
humidity gradients, the member itself can serve as a restraint (Frosch et al. 2003). The magnitude of induced tensile stresses depends on both the amount of volumetric changes and the degree of
M AN U
restraint, i.e. how much movement is restricted (Hadidi and Saadeghvaziri 2005). Bridge deck slabs are typically much longer in one direction than the other, thus volumetric changes due to shrinkage and temperature changes are more pronounced in the longitudinal direction, which induces tensile stresses resulting in transverse cracks.
TE D
Recently, the non-corrodible fiber reinforced polymer (FRP) bars have been used as reinforcement for bridge deck slabs, barrier walls, and parking garages to mitigate the corrosion problem of conventional steel. Among different types of FRP materials, the lower cost of glass
EP
FRP (GFRP) bars makes them attractive for the construction industry. Nevertheless, since GFRP bars have a lower modulus of elasticity than steel, concrete elements reinforced with GFRP
AC C
exhibit larger deformation leading to wider cracks. Unlike steel, GFRP materials do not corrode by nature; however, GFRP-RC are still susceptible to other forms of deterioration due to harsh environments involving de-icing chemicals, sulfate salts and alkalis, which can readily penetrate concrete through cracks (ISIS Canada 2006). This serviceability issue can be aggravated by harsh environmental conditions such as freezing-thawing and wetting-drying cycles. A maximum crack width of 0.5 mm is recommended by CHBDC (CSA S6 2006) for FRP-RC
2
ACCEPTED MANUSCRIPT
components subjected to aggressive environments and 0.7 mm for other members. In the field, transportation elements such as bridge deck slabs and barrier walls are subjected to temperature and humidity changes due to daily or seasonal conditions. Typically, due to the low tensile
RI PT
strength of concrete during the first 24 hours after casting, thermal changes from hydration reactions and moisture loss from the surface of concrete, due to hot, dry and windy conditions, increase the tendency to cracking (Byard et al. 2010). After hardening, ambient temperature and
SC
humidity fluctuations can induce volume instability of concrete. Cracks are the easiest locations for moisture and aggressive chemicals to accelerate the deterioration of reinforcement as well as
M AN U
to reduce the service life of concrete structures. While non-pre-stressed reinforcement cannot completely eliminate cracking, it can control crack spacing and width in bridge deck slabs, thus achieving a key serviceability requirement for RC structures under different environmental conditions.
TE D
PREVIOUS WORK
From an experimental study on concrete panels reinforced with one layer of GFRP bars, Koenigsfeld and Myers (2003) concluded that the equation listed in ACI-440.1R-03 (ACI
EP
Committee 440 2003) [same as in the most recent ACI Committee 440 2006] for minimum FRP reinforcement ratio was overly conservative; however, they did not recommend any alternate
AC C
equation to calculate the minimum FRP reinforcement ratio. Koenigsfeld and Myers (2003) found three times larger crack widths for GFRP-RC panels (1830 mm × 591 mm × 127 mm) than that of their counterpart specimens with similar amounts of steel reinforcement when subjected to restraint and daily cyclic temperature between 21°C to 41°C to simulate pouring concrete on a hot summer day. They also concluded that twice as much GFRP reinforcement as steel is required to achieve similar crack control limits when subjected to flexural loading. Due to lower
3
ACCEPTED MANUSCRIPT
stiffness of GFRP bars, lower internal tensile stresses in concrete will develop due to internal restraint against concrete shrinkage or temperature variations, which leads to larger crack spacing followed by wider crack widths (Chen and Choi 2002).
RI PT
The authors of the current paper previously concluded that the minimum GFRP reinforcement ratio of 0.7% recommended by CHBDC for bridge deck slabs can reasonably control early-age crack width and reinforcement strain under normal laboratory conditions (22 ± 2 °C and 50-
SC
70 % RH) (Ghatefar et al. 2014, Marti-Vergas 2015, Ghatefar et al. 2015). However, bridge deck slabs are usually exposed to harsh environmental conditions such as freezing-thawing cycles,
M AN U
temperature fluctuations and wetting-drying cycles within temperature and humidity ranges from –40°C to +35°C and 30 % to 100 %, respectively (Laoubi et al. 2006). Therefore, at early ages, these structural elements may be subjected to different combinations of shrinkage and swelling. In RC structures, damage due to severe environmental conditions can take various forms such as
TE D
reinforcement de-bonding, scaling, and micro cracking (Bishnoi 2004 and Alves et al. 2011). Other forms of damage include large-scale spalling and crumbling of concrete and material fatigue, resulting in loss of strength and stiffness.
EP
During thermal variations, self-equilibrating stresses develop in FRP due to the difference in coefficients of thermal expansion of constituent elements (fibers and resin), resulting in matrix
AC C
micro-cracking. Further deterioration can occur due to expansion of absorbed moisture in FRP and concrete under freezing-thaw cycling. In addition, freezing-thaw cycles can lead to degradation of the fiber-matrix bond and further damage the fibers through local notching due to ice formation on their surfaces (El-badry et al. 2000). GFRP can experience an expansion of 4-6 times greater than that of concrete in the transverse direction due to temperature variations. This thermal incompatibility can cause bursting stresses to build up in the concrete surrounding the
4
ACCEPTED MANUSCRIPT
reinforcement or deboning of bars from concrete under temperature changes (Gentry et al. 1999). Moreover, under freezing-thawing cycles, ice formation at the interface between FRP and concrete leads to further loss of the FRP bond to concrete and increases existing crack width
RI PT
under sustained loads (Alves et al. 2011). Similarly, wetting-drying cycles are considered critical in the durability-based design of concrete structures since volume changes due to repetitive shrinkage/swelling may lead to material fatigue and de-bonding of reinforcement (Zhang et al.
SC
2012, Ayano et al. 2002). While numerous studies have investigated the structural behavior of FRP-RC elements (Ghatefar et al. 2014, Koenigsfeld and Myers 2003), scarce data on the
M AN U
restrained shrinkage cracking in FRP-RC elements have been reported under different environmental conditions. Hence, further research is needed in this area. This contribution presents an experimental study evaluating the effect of longitudinal (secondary) GFRP reinforcement ratio on early-age cracking of bridge deck slabs subjected to freezing-thawing and
TE D
wetting-drying cycles.
EXPERIMENTAL PROGRAM
This study included four full-size, cast-in-place deck slab prototypes; three reinforced with
EP
GFRP and one reinforced with steel. The slabs were designed according to Section 16 of the CHBDC (Clause 16.8.8.1) with a total reinforcement ratio of 0.7% to investigate the effect of
AC C
different environmental conditions on early-age cracking for a period of 112 days after casting. To determine the internal conditions of the cementations matrix, the dynamic modulus of elasticity (DME) was determined for all cores from the ultrasonic pulse velocity (UPV) measurements according to ASTM C597. Also, to evaluate the interconnectivity of the pore structure in the concrete slabs subjected to different environmental conditions, the rapid chloride penetrability test (RCPT) was performed for the cores according to ASTM C1202 (Standard Test
5
ACCEPTED MANUSCRIPT
Method for Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration) in comparison to non-exposed specimens at the same age (112 days). After the RCPT, the specimens were axially split and sprayed with a silver nitrate solution, which forms a white
RI PT
precipitate of silver chloride in approximately 15 minutes, to measure the physical penetration depth of chloride ions. The average depth of the white precipitate was determined at five different locations along the diameter of each half specimen. This depth is considered to be an
SC
indication of the ease of ingress of chloride ions, and thus the connectivity/deterioration of the microstructure (Bassuoni et al. 2006).
M AN U
Material Properties
Normal-strength, ready-mixed concrete incorporating 13% silica fume by mass of binder (Table 1) with a target 28-day compressive strength of 40 MPa was purposefully used to provoke high tendency for shrinkage as an extreme scenario that might be encountered in practice. The slump
TE D
and fresh air content of this concrete were in the ranges of 100-120 mm and 6±1%, respectively. The US-Federal Highway Administration Guidelines [Silica fume user manual (Holland 2005)] reports on the use of 4 to 15% silica fume in concrete for various infrastructure applications. For
EP
GFRP and steel reinforcement, sand-coated GFRP bars (Pultrall Inc. 2012) and normal steel grade 40, respectively, were used to reinforce the slab prototypes in both layers (top and bottom).
AC C
The GFRP bars are made of continuous E-glass fibers with modified vinyl-ester resin. The longitudinal and transverse coefficients of thermal expansion for the GFRP bars used are 6.2 and 23.8 [×10-6/°C], respectively, while the longitudinal and transverse coefficients of thermal expansion for the steel bars used are 11.7 [×10-6/°C]. Table 2 summarizes the mechanical properties of the steel and GFRP bars used in this research, according to standard laboratory tests (CSA S806-12 and ASTM A400 – 69 -12, respectively).
6
ACCEPTED MANUSCRIPT
Test Setup and Prototypes According to Clause 14.13.1.2 of CHBDC (CSA 2006), the minimum allowable thickness of bridge deck slabs is 175 mm. Therefore, in this study, a thickness of 180 mm for all test slabs
RI PT
was selected. A total of four full-size, cast-in-place concrete bridge deck slabs (2500-mm long by 765 mm wide as shown in Fig. 1) were constructed and tested in the laboratory. The effective width-to-length ratio was selected less than 1/3 to ensure that the amount of shrinkage in the
SC
longitudinal direction is much more than that in the transverse direction; there was three times as much concrete that tended to shrink in the longitudinal direction than in the transverse direction
M AN U
(Frosch et al. 2003). Consequently, a transverse crack was expected to develop in the transvers direction in order to relieve the larger tensile stress in the longitudinal direction. The reinforcement configuration of the test specimens was selected based on the empirical design method recommended by Section 16 - Clause 16.8.8.1 of the CHBDC (CSA 2006).
TE D
According to this section, the minimum FRP reinforcement ratio in the longitudinal bottom and top assemblies is 0.35% with top and bottom covers equal to 35±10 mm (CHBDC, Clause16.4.4). All test prototypes had similar bar size, top and bottom clear covers (25 and 30
EP
mm, respectively) and a constant spacing of 255 mm for the longitudinal reinforcement. Slabs G1, G1-FT, and G1-WD (G: GFRP-RC, FT: under freezing-thawing, WD: under wetting-drying
AC C
conditions) were reinforced with the minimum GFRP reinforcement ratio (0.7 %) while specimen S1 (S: steel-RC) contained similar amounts of steel reinforcement (0.7 %). Prototypes G1 and S1 were tested under normal laboratory conditions (22 ± 2 °C and 50-70 % RH), while slabs G1-FT and G1-WD were subjected to freezing-thawing and wetting-drying cycles respectively at 7 days (Table 3). Dimensions and details of the slabs’ cross-section are shown in Fig. 2.
7
ACCEPTED MANUSCRIPT
Instrumentation To measure strains in the GFRP bars in the vicinity of the first crack, three 6-mm electrical resistance strain gauges [FLA-6-11 series (accuracy: ±1%)] were attached to each bar at the top
RI PT
and bottom layers; one centered at the mid-span, and the other two at 50 mm on each side as shown in Fig. 3. In addition, to measure concrete strain two types of strain gauges were used; embedment strain gauges [EGP series (accuracy: ±2%)] in concrete and 10-mm linear pattern
SC
strain gauges [20CBW series (accuracy: ±1%)] on the surface of the slabs. For each slab, one strain gauge was embedded at the cracking (mid-length) location. The other strain gauge was
M AN U
attached to the surface of concrete at an arbitrary distance of 270 mm (1.5 times slab thickness) away from the mid-length (cracking location) to avoid gauge damage upon the occurrence of first cracking. The internal strain gauge was used to capture the development of tensile strains within concrete up to failure by first cracking, while the surface strain gauge measured the
TE D
deformation of concrete in the vicinity of the cracking location during the entire period of exposure. The width of the cracks developed was recorded throughout the test using two PIgauges. Also, the internal relative humidity was monitored by humidity sensors [Rapid RH®
EP
4.0EX series (accuracy:±3%)] embedded at the level of the reinforcement layers and mid-depth at an arbitrary distance of 625 mm away from the mid-length (cracking location) to avoid further
AC C
stress concentration at cracking place. All instrumentation (except humidity sensors) was connected to a DAQ (data acquisition system) controlled by a computer. Test Procedure
Prior to casting, the 2500 mm long slabs were effectively anchored at its ends by 1473 × 1000 × 1200 mm concrete blocks, which were clamped (pre-stressed) to the laboratory strong floor using 38-mm diameter dywidag bars (Fig. 4). Afterwards, the inside surface of the formwork
8
ACCEPTED MANUSCRIPT
was cleaned and thinly coated with a releasing agent (oil) to prevent adhesion of the concrete. Also, a total of 63 cylinders (150 × 300 mm) were cast using the same concrete as the prototypes. For the first 24 hours after casting, a plastic tent was built around the test prototypes
RI PT
and cylinders while electrical heaters were used to maintain the internal concrete temperature at 35 °C without moist curing. Subsequently, the tent and formwork was removed, then PI-gauges were attached to the concrete surface, and initial strain measurements were recorded. During the
SC
first 24 hours, the ambient conditions around the slab (under tent) were 40°C with 30-40 % RH. The average internal temperature and RH measured at the top reinforcement level were 35 °C
M AN U
and 90 %, respectively. After the tent was removed, the specimens were left in laboratory conditions (22±2 °C and 50-70 % RH) for 7 days. During that period, the internal temperature at the top reinforcement level was 22±2 °C while the RH decreased from 90 to 70 %. After 7 days, freezing-thawing and wetting-drying cycles were applied for specimens G1-FT and G1-WD,
TE D
respectively. Using the concrete cylinders, the average compressive (ASTM C39M-03) and splitting tensile (ASTM C496M-04) strengths and the modulus of elasticity (ASTM C469-02) were obtained after 1, 3, 7, 14, 21 and 28 days (Table 4) to determine the development of
EP
concrete properties (within a standard deviation of 5%). At the mid-length of each slab, the cross-section was locally reduced (notched) to 565 × 150 mm
AC C
to control the first crack occurrence at this location (Fig. 5). The bottom surface of the slab was supported by three stay-in-place smooth greasy plates (300 mm × 300 mm, spaced at 1250 mm as shown in Fig. 6) to reduce the effect of slab’s self-weight on the reinforcement strains. Also, fresh water was poured into the surface reservoir (approximately 5-mm thick) created by peripheral foam dykes at the outer edge of the slabs G1-FT and G1-WD for freezing-thawing
9
ACCEPTED MANUSCRIPT
and wetting cycles. Concrete surface strain gauges were damaged for the slab G1-FT in the first freezing-thawing cycle. Environmental Conditioning Schemes
RI PT
Freezing-thawing cycles
Different freezing-thawing conditioning schemes had been used by researchers to study the behavior of RC elements externally or internally reinforced with FRP composites (Laoubi et al.
SC
2006 and Alves et al. 2011). In the current research, the temperature profile of Standard Test Method for Resistance of Concrete to Rapid freezing-thawing (ASTM C-666 M-03 2008) was
M AN U
adopted. In this standard, the freezing-thawing cycles consist of alternately lowering the temperature from +4 to -18ºC for freezing and raising it from -18 to +4 ºC for thawing. Thawing time should not be less than 25% of the total freezing-thawing time. In order to reach the standard conditions in the bottom reinforcement level of the slab, the applied freezing-thawing
TE D
cycles consisted of alternately decreasing the environmental chamber temperature from +22 to 25ºC for freezing and raising it from -25 to +35ºC for thawing at a rate of 1.55 cycles/day to achieve the ASTM temperature and duration requirements at the level of GFRP reinforcement.
EP
Figure 7 shows the reading of thermocouples embedded in specimen G1-FT at the level of bottom reinforcement compared to the air temperature inside the chamber. Specimen G1-FT was
AC C
subjected to 163 freezing-thawing cycles over 105 days. Figure 8 (a) indicates that ponded water (3-5 mm) on the top of slab G1-FT increased the average internal humidity to approximately 99 % (i.e. beyond the critical saturation level of 90%). Wetting-drying cycles
It should be noted that there are no standard test methods for the wetting-drying exposure of concrete. The cyclic regime and the total number of cycles (five cycles) in this study were
10
ACCEPTED MANUSCRIPT
selected similar to that adopted by Zhang et al. (2012) to achieve significant humidity changes at the reinforcement level of the bridge deck slabs. Each wetting-drying cycle started with 14 days of drying at 35±2°C and 30% RH followed by 7 days of wetting at 22±2°C and 100% RH.
RI PT
Figure 8 (b) shows the reading of humidity sensors embedded in the test specimen at the different levels of the cross section compared to the humidity in the chamber. RESULTS AND DISCUSSION
SC
In this section, the experimental results are primarily presented in terms of concrete cracking pattern, width, and spacing, and strains in the reinforcement and concrete, with links to the
M AN U
materials tests. Finally, the main experimental results from this study are compared to predictions from a theoretical model for RC restrained members that are not subjected to significant bending (Gilbert 1992). General Behavior
TE D
The width of cracks in the restrained slabs varied according to the environmental exposure and type of reinforcement. While the magnitude of crack width depends on several factors such as degree of restraint, quality of bond between concrete and reinforcement, size and distribution of
EP
bars, concrete quality and ambient conditions, the variables tested in this program were the environmental conditions and properties of reinforcement. For each slab, the crack width was
AC C
considered as the average of the measured value at two locations across the slab width at midlength. Generally, the first crack in all specimens was observed within the first three days after casting in the transverse direction before exposure. Figure 9 shows the cracking pattern for all the slabs at the notched (mid-length) location. The cracks, which usually extended into the full depth of slabs, typically occurred at mid-length. The internal concrete strain at cracking location was measured by the embedment concrete strain
11
ACCEPTED MANUSCRIPT
gauge up to cracking. Since concrete cracking damaged the internal strain gauge, the surface strain gauges, located at a distance of 270 mm from the cracking location, were used to measure concrete strain after cracking. Figure 10 shows the internal strain of the concrete up to cracking,
RI PT
for specimens G1, S1, G1-FT and G1-WD. The measured internal tensile strains of concrete were 300, 270, 280 and 310 µε, respectively, when the early-age cracks became visible. Once the crack formed at mid-length (crack location), the stress in the concrete dropped to zero. For all
SC
specimens after cracking, the concrete surface strain changed to a compressive strain [negative values in Figs. 11 (c), 12 (e) and 13 (c)]. When cracked, the stiffness of the slab in the vicinity of
the crack shortened elastically.
M AN U
a crack reduced, depending on the cross-section axial stiffness, and the concrete on either side of
The strain in the reinforcing bars was presented as the average strain readings of all instrumented bars (top and bottom) in the vicinity of the crack at mid-length. Figures 11, 12 and 13 indicate
TE D
that, once the crack formed at mid-length, the reinforcement carried the full restraining force for the slabs G1, S1, G1-FT and G1-WD, consequently, the average reinforcement strains suddenly increased to 1450, 680, 1270, 1370 µε, respectively. In the meantime, the average reinforcement
EP
strains, before cracking, were 290, 285, 275 and 300 µε, respectively. Normal laboratory exposure
AC C
The first crack occurred for slabs G1 and S1 within 48 ad 62 hours, respectively. Figure 11 (a) shows the change in crack width over the test period under normal laboratory conditions. The final crack width in slab G1 was 0.33 mm, while in slab S1 it reached 0.18 mm after 112 days of casting. This was expected due to the lower section stiffness of the GFRP-RC slab compared to that of the steel-RC slab. Reduction of the reinforcement modulus of elasticity increased the crack width as well as average crack spacing. In slab G1, the average crack spacing was 2500
12
ACCEPTED MANUSCRIPT
mm whereas in slab S1 it reduced to 833 mm (the second and third cracks occurred in S1 after 19 days on both sides of the first crack, Fig. 9). An increase in the reinforcement modulus of elasticity leads to less stiffness reduction at first cracking (mid-length), therefore the restraining
RI PT
force after cracking remains relatively high resulting in further cracking. Generally with a higher restraining force, due to further drying shrinkage or any ambient temperature variation, the concrete in regions away from the first crack tends to experience further cracking.
SC
Figure 11 (b) shows that the average strain in the steel bars (S1) in the vicinity of the first crack was 410 µε after 112 days under laboratory conditions. Comparatively, the strains in the GFRP
M AN U
bars of specimen G1 reached 1520 µε, because of the lower section stiffness in the GFRP-RC slab. This complies with the crack width trend as increasing the reinforcement modulus of elasticity from 62 to 200 GPa, the surface strain of concrete reduced by approximately 28%. Correspondingly, the concrete surface compressive strains for slabs S1 and G1 were 210 and 290
TE D
µε, respectively (Fig. 11 (c)) Freezing-thawing exposure
The behavior of restrained concrete elements under freezing-thawing conditions is affected by
EP
multiple variables (e.g. internal water expansion, and material contraction due to low temperature). It is well documented that if concrete elements are critically saturated (internal RH
AC C
> 90%), the water volume expansion phenomenon in larger capillary pores induces considerable volume changes of concrete during freezing. At the onset of ice crystallization, the frictional resistance to ice growth creates internal pressure in the pores leading to concrete expansion (Scherer et al. 2002). In addition, ice formation in the void space imbibes water from the smaller (gel) pores, creating negative (suction) pressure in the matrix and thus contraction (Towers and
13
ACCEPTED MANUSCRIPT
Helmuth 2008). Hence, the total volume change of concrete is a combination of expansion and contraction from hydraulic and osmotic pressures. The transverse full-depth crack occurred for slab G1-FT within 47 hours (before exposure) at the
RI PT
mid-length (notched section). Figure 12 (a) shows the change in crack width in this slab over 163 freezing-thawing cycles (112 days). Crack width reached to its maximum and minimum point at +4°C (thawing) and -18°C (freezing), respectively, in each cycle. At the last cycle (163),
SC
the crack width varied between 0.29 and 0.42 mm corresponding to the freezing-thawing stages, respectively (Fig. 12 (b)). The lower crack width recorded during freezing periods can be
M AN U
attributed to the volumetric expansion of the critically saturated slab, which led to partial closure of the crack opening. Upon relieving the expansion pressure during thawing periods, the crack width increased up to 0.42 mm (in the last cycle), which is 40% and 27% higher than the crack width measured before the freezing-thawing exposure and in the normal exposure (slab G1),
TE D
respectively. Also, it was observed that slab G1-FT suffered from moderate surface scaling (less than 0.5 kg/m2 at the end of exposure (according to BNQ 2002 test standard)), which is a typical damage manifestation of concrete exposed to freezing-thawing cycles. This trend might be
EP
ascribed to over finishing the surface of slab G1-FT, which led to reducing the volume of air entrainment in the surface as shown by the Scanning Electron Microscopy (SEM) analysis.
AC C
For slab G1-FT, Fig. 12 (c) shows the average fluctuation of the strains in the reinforcement at cracking after 163 freezing-thawing cycles. In the last cycle, the strains of the reinforcement at the cracking location were 1020 and 1690 µε during the freezing-thawing stages, respectively (Fig. 12 (d)). Complying with the crack width results, this trend is attributed to the repetitive volumetric change associated with frost action as discussed earlier.
14
ACCEPTED MANUSCRIPT
Wetting and drying exposure Wetting-drying conditions may significantly affect RC elements due to the variation of moisture distribution with depth and accelerated shrinkage during drying periods. For slab G1-WD, the
RI PT
relative humidity of the concrete surface markedly changed during wetting-drying cycles relative to the inner core, which led to further deformations (Fig. 8 (b)). Fig. 13 (a) shows the change in crack width for slab G1-WD which was subjected to 5 wetting-drying cycles over 112 days. In
SC
the last cycle (22°C with 100% RH and 35°C with 30% RH) the crack width varied between 0.23 and 0.46 mm. The lower crack width recorded during the wetting periods can be attributed to
M AN U
swelling of the slab due to the increase in relative humidity, which led to partial closure of the crack opening. Subsequently, excessive drying of the slab increased the crack width up to 0.46 mm (in the last cycle), which is 0.92% and 0.39% higher than that crack width measured before the wetting and drying exposure and in the normal exposure (slab G1 after exposure for 112
TE D
days), respectively.
Corresponding to the crack width trend, the strain in the GFRP bars increased from 1400 (wetting portion) to 2250 µε (drying portion) in the last cycle (Fig. 13 (b)), due to the additional
EP
drying shrinkage deformation under hot-arid conditions, which was restrained by the GFRP reinforcement. It should be noted that this value is significantly higher than the maximum strains
AC C
recorded in the normal (1520 µε) and freezing-thawing (1690 µε) exposures. This behavior was confirmatory to the measurements of concrete surface strain that increased up to 480 µε due to the additional shrinkage during the drying portion (Fig. 13 (c)). Materials Tests:
In additional to the early-age cracking of restrained concrete slabs exposed to harsh conditions, they were vulnerable to material degradation especially at the mid-length (in the vicinity of the
15
ACCEPTED MANUSCRIPT
crack) due to temperature and humidity variations. To capture this trend, three materials tests: Dynamic Modulus of Elasticity (DME), Rapid Chloride Permeability Test (RCPT), and Scanning Electron Microscopy (SEM) were conducted on cores extracted from the three GFRP-
RI PT
RC slabs. A total of twelve 100-mm diameter cores were extracted; four cores from each slab (close and away from left and right sides of the crack). All cores were extracted from slabs at the end of the test period (112 days). For the DME test, full length cores were used, while for the
SC
other two tests, the top 50-mm thick slices were cut from the cores extracted from different locations in the slabs.
M AN U
UPV test (Ultrasonic Pulse Velocity test)
Table 5 shows the dynamic modulus of elasticity results for the unexposed concrete (cores from specimen G1 under laboratory conditions) and after being subjected to freezing-thawing and wetting-drying conditions (cores from G1-FT and G1-WD). While the test results were in the
TE D
narrow range of 45-50 GPa, they showed a general reduction (maximum of 10%) of DME for the concrete exposed to cyclic conditions, which indicates the existence of fissures and micro-cracks in the cementitious matrix.
EP
RCPT test (Rapid Chloride Permeability test) After operating the RCPT for 6 hours according to ASTM C1202 (Standard Test Method for
AC C
Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration), the penetration depth was measured in the concrete specimens. The whitish color of the penetration depth was clearly visible as depicted in Fig. 14, and the results are listed in Table 5. In contrast to the freezing-thawing exposure, Table 5 shows that the specimens extracted from the slab subjected to wetting-drying cycles yielded relatively higher penetration depths in the vicinity of the crack location, which signifies that the pore structure was highly interconnected in
16
ACCEPTED MANUSCRIPT
these specimens. This can be attributed to a higher intensity of micro-cracks due to the matrix fatigue resulting from high strain fluctuations of repetitive swelling and shrinkage (Fig. 13 (c)) in the wetting-drying exposure. These results are consistent with the higher concrete and
RI PT
reinforcement strain values for the spacemen G1-WD exposure to drying conditions (Fig. 13 (b), and (b)). Microstructural analysis
SC
To supplement the results of UPV and RCPT, the alteration of microstructure of concrete was also assessed by backscattered scanning electron microscopy (BSEM) on thin sections from the
M AN U
cores extracted from G1-FT and G1-WD in the vicinity and away from the main crack. The polished sections were prepared from fracture surfaces that were dried at 40°C for 24 h, impregnated with low-viscosity epoxy resin under pressure, cut, polished and carbon coated. The SEM micrographs show that the specimen subjected to wetting-drying conditions
TE D
particularly in the vicinity of the main crack (Fig. 15 (a)) had higher intensity of micro-cracks and internal damage than that of the concrete exposed to freezing-thawing cycles (Fig. 15 (b)). This trend is consistent with the RCPT test, the higher recorded concrete and reinforcement
EP
strain values in the vicinity of the crack, and the crack width for the specimen under wettingdrying conditions (Fig. 13 (b) and (c)).
AC C
Theoretical vs. experimental results Limited studies provided formulas to predict cracking characteristics of RC deck slabs and stress distribution in bars at cracking locations due to restrained shrinkage. In Gilbert’s 1992 theoretical analysis, it was explained that shrinkage causes an axial force built-up (Eq. 1) in restrained members, which leads to direct tension cracks. Equations 2 and 3 were proposed to calculate the stress in the bars, in the vicinity of the crack, and the crack width, respectively. In this model, the
17
ACCEPTED MANUSCRIPT
restraint is provided to the longitudinal movement caused by shrinkage and temperature changes. Table 6 illustrates input values for parameters used in Equations 1 to 3.
∗ =
. ∗ . ∆
'(
+∗ ,∗
∗ .
Eq. 2
.
) = −[
−
./ − 01 2 + # ∗ $ 0]
!
+ # ∗ $ % ∗ &
Eq. 1
RI PT
∞ =
Eq. 3
SC
Considering the fully restrained member in direct tension as the concrete shrinks, the restraining force gradually increases until the first crack occurs, which is usually within two days from the
M AN U
commencement of drying. Immediately after the first cracking, the restraining force reduces and the concrete stress away from the crack is less than the tensile strength of the concrete. The concrete on either side of the crack shortens elastically and the crack opens to a width, w. At the crack location, the reinforcement carries the entire force (the concrete strain drops to zero. In the
TE D
region immediately adjacent to the crack the concrete and the reinforcement stresses vary considerably, and a region of partial bond breakdown exists. At a distance so from the crack, which was earlier proposed by Favre et al. (1983) for a member containing deformed bars or
EP
welded wire mesh (Eq. 1), the concrete and the reinforcement stresses are no longer influenced
AC C
directly by the presence of the crack. It was suggested that the value of so to be multiplied by 1.33 to achieve better predictions for RC member with steel bars (Gilbert 1992). Ghatefar et al. (2014) concluded that the coefficient of 0.8 instead of 1.33 for so led to better predictions for FRP-RC members.
In order to calculate the width of crack and reinforcement stress in the vicinity of the main crack (under freezing-thawing conditions) by Gilbert’s equations, the total strain of the concrete was calculated based on the combination of concrete contraction due to temperature reduction to 18
ACCEPTED MANUSCRIPT
18°C (using the measured concrete thermal coefficient of 5.5×10−6/℃), and the expansion (up to 9%) of water in large capillary pores (greater than 100 nm) upon freezing at critically saturated conditions. Also, to evaluate the crack width and reinforcement stress in the vicinity of the main
calculate concrete strain under wetting-drying conditions.
RI PT
crack in specimen G1-WD by Gilbert’s equations, ACI 209.2R-08 guidance was used to
For the specimen subjected to freezing-thawing conditions, the measured crack widths, using
SC
Gilbert’s model (Gilbert 1992), were 0.28 and 0.31 mm for freezing and thawing, respectively. The values of the crack width agree with the results of the analytical model with an error of
M AN U
approximately 20%. Also, the reinforcement strains in the vicinity of the crack were comparable to those obtained from the theoretical model (1205 and 1845 µε during freezing and thawing, respectively) within 13% error.
For the specimen subjected to wetting-drying conditions, the measured widths of cracks, using
TE D
Gilbert’s 1992 theory, were 0.33 and 0.43 mm respectively. The values for crack width within first wetting-drying cycle agree with the results of the analytical model with an error of approximately 13%.
While For the specimen exposed to wetting-drying conditions, the
EP
measured strains in the GFRP bars, using Gilbert’s 1992 model, were 1889 and 568 µε, respectively. The values for GFRP reinforcement strain within the first wetting-drying cycle
AC C
agree with the results of the analytical model with an error of approximately 17%. CONCLUSIONS
Based on the experimental variables and environmental conditions implemented in the current research, the following conclusions can be drawn:
19
ACCEPTED MANUSCRIPT
1. Due to the relatively lower modulus of elasticity of GFRP bars, the crack width and average reinforcement strain in slab G1 were larger than those of the counterpart slab reinforced with similar steel reinforcement ratio (S1).
RI PT
2. Under normal laboratory conditions, increasing the reinforcement modulus of elasticity from 62 GPa (GFRP) to 200 GPa (steel) reduced the average crack spacing. An increase in the reinforcement modulus of elasticity leads to less stiffness reduction at first cracking
SC
(mid-length), therefore the restraining force after cracking remains high resulting in further cracking.
M AN U
3. Under freezing-thawing conditions, the crack width reached its maximum and minimum points at +4°C (thawing) and -18°C (freezing), respectively, in each cycle. This behavior is attributed to the volumetric expansion of the critically saturated slab during freezing, which led to partial closure of the crack opening. Upon relieving the expansion pressure
TE D
during thawing periods, the crack width increased up to 0.42 mm (in the last cycle), which is 40 and 27% higher than the crack width measured before the freezing-thawing exposure and in the normal exposure (slab G1), respectively.
EP
4. In the specimen under wetting-drying conditions, the lower crack width and reinforcement strain recorded during the wetting periods can be attributed to swelling of
AC C
the slab due to the increase in relative humidity, which led to partial closure of the crack opening. Subsequently, excessive drying of the slab increased the crack width and reinforcement strain in the drying period.
5. The UPV test results show a general reduction of DME for the concrete exposed to cyclic conditions, which indicates the existence of fissures and micro-cracks in the cementitious matrix. While RCPT and microstructural analysis indicate that the specimens extracted
20
ACCEPTED MANUSCRIPT
from the slab subjected to wetting-drying cycles yielded relatively higher penetration depths and more internal micro cracks in the vicinity of the crack location, which signifies that the pore structure was highly interconnected in these specimens. This can be
RI PT
attributed to a higher intensity of micro-cracks due to the matrix fatigue resulting from high strain fluctuations of repetitive swelling and shrinkage in the wetting-drying exposure. These results are consistent with the higher concrete and reinforcement strain
SC
values for specimen G1-WD exposed to wetting-drying conditions.
6. Test results indicate that the minimum reinforcement ratio (0.7%) recommended by
M AN U
CHBDC for bridge deck slabs reinforced with GFRP bars is conservative under laboratory condition. However, this ratio satisfied the serviceability requirements of the CHBDC (crack width of 0.5 mm and 5800 µε, which represents 25% of GFRP ultimate strain) after being subjected to the simulated exposures of freezing-thawing and wetting-
TE D
drying cycles. The maximum measured crack width and GFRP strains did not exceed 0.46 mm and 2250 µ, respectively.
7. The measured width of shrinkage cracks and stresses in GFRP agreed with most results
EP
from Gilbert analytical model (Gilbert 1992) within 17% error. However, more refinement to Gilbert’s model is still needed to be fully applicable to FRP-RC slabs under
AC C
different environmental conditions (especially for structures reinforced with different rebar diameter and spacing), which is recommended for future research.
Finally, similar to the case of normal laboratory conditions, it can be inferred that a corresponding steel-RC slab, subjected to wetting-drying and freezing-thawing cycles, would achieve the same trend but with lower crack width and strain in reinforcement due to the higher modulus of elasticity of steel.
21
ACCEPTED MANUSCRIPT
ACKNOWLEDGMENTS The financial support provided by the Natural Science and Engineering Research Council of Canada (NSERC) through Canada Research Chairs (CRC) program and the Network of Centers
RI PT
of Excellence on Intelligent Sensing for Innovative Structures (ISIS Canada) are gratefully appreciated. Also, the efforts of A. Ghazy and the technical staff of the McQuade Heavy Structures Laboratory and IKO Construction Materials Testing Facility at the University of
AC C
EP
TE D
M AN U
SC
Manitoba are greatly acknowledged.
22
ACCEPTED MANUSCRIPT
NOTATION Areinf.: total area of bars in a section (mm2)
Ec (day): concrete modulus of elasticity at different age (MPa) E*e: effective concrete modulus of elasticity (MPa)
fc (day): concrete compressive strength at different age (MPa) ft (day): concrete tensile strength at different age (MPa)
l: slab length (mm) m: number of cracks N (∞): final tensile force (kN)
M AN U
ft (reinf): tensile strength of reinforcement bars (MPa)
SC
Ereinf.: reinforcement modulus of elasticity (GPa)
RI PT
db: bar diameter (mm)
TE D
5∗ : reinforcement modulus of elasticity divided by concrete effective modulus of elasticity S: average crack spacing (mm)
/1 : reinforcement stress transfer length at cracking location (mm)
EP
t: slab thickness (mm)
w: final cracking width (mm)
AC C
∆u: support displacement (mm)
ε*sh: ultimate shrinkage strain (using ACI 209.2R-08) ρreinf.: reinforcement ratio
σav: estimate of the average concrete stress in the period after first cracking (MPa) σ*c1: final concrete stress away from the crack (MPa) ϕ*: creep coefficient (according to ACI 209.2R-08)
23
ACCEPTED MANUSCRIPT
REFERENCES ACI Committee 440. 2006. “Guide for the Design and Construction of Structural Concrete Reinforced with FRP Bars”. ACI 440.IR-06, American Concrete Institute (ACI), Detroit, MI.
RI PT
Alves, J., El-Ragaby, A. and El-Salakawy, E. 2011. “Durability of GFRP Bars Bond to Concrete under Different Loading and Environmental Conditions”. Journal of Composite for Construction, ASCE, Vol. 15, No. 3, pp. 249-269.
SC
ASTM C666 M-03. 2008. “Standard Test Method for Resistance of Concrete to Rapid Freezing and Thawing”. ASTM International, US.
M AN U
ASTM C597. 2009. “Standard Test Method for Pulse Velocity through Concrete,” ASTM International, US.
ASTM C1202. 2012. “Standard Test Method for Electrical Indication of Concrete's Ability to Resist Chloride Ion Penetration,” ASTM International, US.
TE D
ASTM A400. 2012. “Standard Practice for Steel Bars, Selection Guide, and Mechanical Properties,” ASTM International, US.
Ayano, T. and Wittmann, F.H. 2002. “Drying, moisture distribution, and Shrinkage of cement-
EP
based materials”. Journal of Materials and Structures, Vol. 35, No. 247, pp. 134-140. Bassuoni M. T., Nehdi M. L. and Greenough, T. R. 2006. “Enhancing the Reliability of
AC C
Evaluating Chloride Ingress in Concrete Using the ASTM C 1202 Rapid Chloride Penetrability Test,” Journal of ASTM International, Vol. 3, No. 3, p. 13. BNQ NQ 2621-900. 2002. “Determination of the Scaling Resistance of Concrete Surfaces Exposed to Freezing-and-Thawing Cycles in the Presence of De-icing Chemicals,” Bureau de normalisation du Québec, Annex A, pp. 19-22. Move up, alphabetical order!!!
24
ACCEPTED MANUSCRIPT
Byard, B. E., Schindler, A. K., Barnes, R. W. and Rao, A. 2010. “Cracking Tendency of Bridge Deck Concrete”. Journal of the Transportation Research Board, Washington, D.C., No. 2164, pp.122-131.
RI PT
CSA. 2006. “Canadian Highway Bridge Design Code - CHBDC”. CAN/CSA-S6-06, Canadian Standards Associations, Rexdale, Ontario.
CSA. 2012. “Design and Construction of Building Structures with Fibre-Reinforced Polymers”.
SC
CSA/S806-12, Canadian Standards Association. Toronto, Ontario, Canada.
Favre, R. 1983. “Fissuration et deformations.” Manual du Comité Euro-Internationale du Beton
M AN U
(CEB). Ecole Polytechnique Federale de Lausanne, Swizerland, pp. 249. Frosch, R. J., Blackman, D. T. and Radabaugh, R. D. 2003. “Investigation of Bridge Deck Cracking in Various Bridge Superstructure Systems,” Joint Transportation Research Program, E-port No: FHWA/IN/JTRP-2002/25, Indiana Department of Transportation and
TE D
Purdue University, West Lafayette, Indiana.
Gilbert, R. I. 1992. “Shrinkage Cracking in Fully Restrained Concrete Members”, ACI Structural Journal, Vol. 89, No. 2, pp.141-149.
EP
Ghatefar, A., El-Salakawy, E. and Bassuoni, M.T. 2014. “Effect of Reinforcement Ratio on Transverse Early-Age Cracking of GFRP-RC Bridge Deck Slabs”. Journal of Composite for
AC C
Construction, ASCE, Vol. 18, No. 6, 04014018, 9 p. Ghatefar, A., El-Salakawy, E. and Bassuoni, M.T. 2015. “Closure to the discussion on the paper: Effect of Reinforcement Ratio on Transverse Early-Age Cracking of GFRP-RC Bridge Deck Slabs”. Journal of Composite for Construction, ASCE, Vol. 19, No. 1, 070140001, 1 p. Hadidi, R. and Saadeghvaziri, M. A. 2005. “Transverse Cracking of Concrete Bridge Decks: State-of-the-Art”. Journal of Bridge Engineering, ASCE, Vol. 10, No. 5, pp. 503-510.
25
ACCEPTED MANUSCRIPT
Holland, T.C. 2005. “Silica Fume User’s Manual.” US Federal Highway Administration Technical report, Washington DC. U.S. ISIS Canada. 2006. “Durability of Fiber Reinforced Polymers in Civil Infrastructure”. Durability
RI PT
Monograph, Winnipeg, Manitoba, Canada: ISIS Canada Research Network.
Krasus, P. D. and Rogalla, E. A. 1996. “Transverse Cracking in Newly Constructed Bridge Decks”. NCHRP Report 380, Transportation Research Board, National Research Council,
SC
Washington, D.C.
Laoubi, K., El-Salakawy, E. and Benmokrane, B. 2006. “Creep and durability of sand-coated
M AN U
glass FRP bars in concrete elements under freeze/thaw cycling and sustained loads”. Journal of Concrete and Composites, Vol. 10, No. 28, pp. 869-878.
Marti-Vargas, J.R. 2015. “Discussion of the paper: Effect of Reinforcement Ratio on Transverse Early-Age Cracking of GFRP-RC Bridge Deck Slabs by Amir Ghatefar, Ehab El-Salakawy
p.
TE D
and M. T. Bassuoni” Journal of Composite for Construction, ASCE, Vol. 19, 070140001, 1
Pultrall Inc. 2012. “V-RODTM-Technical data sheet.” ADS Composites Group Inc., Thetford
EP
Mines, Quebec, Canada. http://www.pultrall.com (visited on October 15, 2012). Scherer, W., Chen, J. and Valenza, J. 2002. “Method of Protecting Concrete from Freeze
AC C
Damage.” U.S. Patent No. 6,485,560. TRB. 2006. “Control of Cracking in Concrete-State of the Art”. Transportation Research Circular E-C107, Transportation Research Board of the National Academic, Basic research and emerging technologies related concrete committee, NY.
26
ACCEPTED MANUSCRIPT
Towers, T.C. and Helmuth, R.A. 2008. “Theory of Volume Changes in Hardened PortlandCement Paste during Freezing"., American Concrete Institution (ACI). Special Publication, SP-249, pp. 141-160.
RI PT
U.S. Army Corps of Engineers. 1981. “Test Method for Coefficient of Linear Thermal Expansion of Concrete” CRD C 39–81, issued 1 June 1981. Available at: http://www.wes.army.mil/SL/MTC/handbook/crd_c39.pdf
SC
Zhang, J., Gao, Y., Han, Y. and Sun, W. 2012. “Shrinkage and Interior Humidity of Concrete under Dry–Wet Cycles”. International Journal of Drying Technology, Vol. 30, No. 6, pp.
AC C
EP
TE D
M AN U
583-596.
27
ACCEPTED MANUSCRIPT
Table 1: Proportions of concrete per cubic meter
365 kg
Coarse aggregate
1020 kg
(max. aggregate size, 20 mm) 650 kg
High-range water reducer
2500 ml
Air entraining agent
250 ml
Silica Fume
54.6 kg
M AN U
Fine aggregate
Water
170 kg
EP
TE D
GU = General use.
AC C
*
SC
Cement Type GU*
RI PT
Quantity/m3
Ingredient
1
ACCEPTED MANUSCRIPT
Table 2: Mechanical properties of GFRP and steel bars Bar area in each layer (top and bottom)
Modulus of
strength
tensile
(mm2)
elasticity
(MPa)
strain (%)
Bar size (mm)
(GPa) Nominal
CSA S806-12 Annex A
15.9
197.8
282.7
62
16
200
------
200
1450
SC
GFRP, No. 16
AC C
EP
TE D
M AN U
Steel, 15M
Ultimate
RI PT
Bar Dia.
Tensile
2
590
2.33
ɛy = 0.2
ACCEPTED MANUSCRIPT
Table 3: Test Matrix Areinf*
ρ**
(mm2)
(%)
S1
400
0.35
1.2
Normal laboratory conditions (22°C and 50 to 60% RH)
G1
382
0. 35
1.0
Normal laboratory conditions (22°C and 50 to 60% RH)
G1-FT
382
0. 35
1.0
Freezing-thawing cycles (-18°C to +4°C)
G1-WD
382
0. 35
1.0
RI PT
Environmental Conditions
M AN U
Wetting-drying cycles ( 22°C to 35°C and 100 to 30 % RH)
The area and ratio of longitudinal bars per layer (top and bottom)
TE D
Reinforcement spacing: 255 mm
EP
**
AC C
*
ρ/ρmin
SC
Prototypes
3
ACCEPTED MANUSCRIPT
Table 4: Compressive, tensile and E-modulus test results for concrete under different environmental conditions Ambient N
N
N
N
W/D
F/T
N
W/D
F/T
N
W/D
F/T
Age (days)
Tensile Strength (MPa) Modulus of Elasticity (GPa)
3
7
14
14
14
21
21
21
28
28
28
7
14
34
35
37
33
36
38
32
38
41
35
1.3
3.4
3.5
3.9
3.3
3.6
3.1
3.7
3.9
3.3
18
21.1
21.2
21.2
21.8
22.2
21
0. 6 -
SC
Compressive Strength (MPa)
1
3.7
M AN U
Property
RI PT
conditions
21.8
21.3
21.1
21.8
N: Normal laboratory conditions (22°C and 50-60% RH),
W/D: Wetting and drying cycling (22°C to 35 °C and 100 to 30 % RH respectively), and
AC C
EP
TE D
F/T: Freezing and thawing cycling (-18 °C to +4 °C and 100% RH).
4
ACCEPTED MANUSCRIPT
Table 5: DME and RCPT results
Wetting and Drying
(mm)
C-R
50
7
C-L A-R A-L
49 52 51
C-R C-L A-R A-L C-R C-L A-R A-L
48 49 46 48 48 48 47 45
RI PT
Freezing and Thawing
(GPa)
Cores
6 5 6
SC
Normal (un-exposed)
Average Penetration Depth
M AN U
Exposure
Dynamic Modulus of Elasticity
AC C
EP
TE D
C: Center, A: Away from center, L: Left, R: Right
5
8 8 7 8 9 12 8 8
ACCEPTED MANUSCRIPT
Table 6: Input data for parameters used in equations 1 to 3 db
(mm2)
(mm)
S1
800
16
-1.71e-4
3.7
38
21800
200000
fy:= 400
G1
764
15.9
-1.71e-4
3.7
38
21800
68297
1450
G1-WD
764
15.9
3.9
41
22200
68297
1450
3.1
35
ε*sh
ft (28)
fc (28)
E c (28)
EGFRP
ft (reinf)
(MPa)
(MPa)
(MPa)
(MPa)
(MPa)
-2.23e-4(dry)
-1.07e-4(thaw) G1-FT
764
15.9
21200
M AN U
-0.83e-4(freeze)
SC
-1.09e-4(wet)
RI PT
Areinf. Slab
68297
AC C
EP
TE D
l = 2500 mm, t = 180 mm, ϕ* = 0.6, ft (7) = 3.4 MPa, fc(3) = 25 MPa, Ec(3) = 20200 MPa,
6
1450
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 1: General view of the test setup and specimen (all dimensions are in mm).
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Fig. 2: Deck slab dimensions – (a) side view, (b) top view, and (c) cross-section A-A (all dimensions are in mm).
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 3: Typical instrumentation of a deck slab (all dimensions are in mm).
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 4: Slab ends effectively held in position and restrained against translation.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 5: Mid-length details.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 6: Smooth supports at the bottom surface of the slabs to reduce self-weight effect
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 7: A part of the freeze/thaw profile for specimen G1-FT.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 8: Relative humidity readings for slab prototypes (a) G1-FT and (b) G1-WD.
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Fig. 9: Crack pattern in the specimens at the notched location [S1: Steel-RC under laboratory conditions, G1: GFRP-RC under laboratory conditions, G1-FT: GFRP-RC under freeze/thaw conditions, G1-WD: GFRP-RC under wet/dry conditions] (all dimensions are in mm).
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 10: Internal strain of concrete up to cracking.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 11: Control slabs (G1, S1) under normal laboratory conditions: (a) crack width development, (b) development of the bar strains, and (c) surface strain of the concrete in the vicinity of the main crack.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 12: Slab G1-FT under freeze/thaw conditions: (a) crack width development, (b) crack width development in the slab during first and last cycles, (c) reinforcement strain development, (d) development of strains in the reinforcement during first and last cycles, and (e) surface strain of the concrete in the vicinity of the main crack.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 13: Slabs G1-WD under wetting-drying conditions: (a) crack width development, (b) development of the bar strains, and (c) surface strain of the concrete in the vicinity of the first crack.
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 14: Chloride penetration depth in cores extracted from: (a) slab G1-FT close to the crack area, (b) slab G1-FT away from the crack area, (c) slab G1-WD close to the crack area, and (d) slab G1-WD away from the crack area.
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 15: Typical SEM micrographs from: (a) specimen G1-WD (slab under wetting-drying conditions), and (b) specimen G1-FT (slab under freezing-thawing conditions) in the vicinity of the main crack.