Publisher’s version / Version de l'éditeur:
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la
première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.
Questions? Contact the NRC Publications Archive team at
PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
50th Brazilian Concrete Congress "Concrete Infrastructure and Environmental Conservation" [Proceedings], pp. 1-20, 2008-09-04
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright
NRC Publications Archive Record / Notice des Archives des publications du CNRC :
https://nrc-publications.canada.ca/eng/view/object/?id=c27a946d-02d4-4f1d-b9cf-ee846181f9c2 https://publications-cnrc.canada.ca/fra/voir/objet/?id=c27a946d-02d4-4f1d-b9cf-ee846181f9c2
NRC Publications Archive
Archives des publications du CNRC
This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Internal curing of high performance concrete bridge decks and its effects on performance, service life, and life-cycle cost
http://irc.nrc-cnrc.gc.ca
I nt e r na l c uring of high pe rfor m a nc e c onc re t e
bridge de ck s a nd it s e ffe c t s on pe rfor m a nc e ,
se r vic e life , a nd life -cycle c ost
N R C C - 5 0 5 3 8
C u s s o n , D . ; D a i g l e , L . ; L o u n i s , Z .
S e p t e m b e r 2 0 0 8
A version of this document is published in / Une version de ce document se trouve dans:
50th Brazilian Concrete Congress Concrete Infrastructure and Environmental Conservation, Salvador, Brazil, September 4-9, 2008, pp. 1-20.
The material in this document is covered by the provisions of the Copyright Act, by Canadian laws, policies, regulations and international agreements. Such provisions serve to identify the information source and, in specific instances, to prohibit reproduction of materials without written permission. For more information visit http://laws.justice.gc.ca/en/showtdm/cs/C-42
Les renseignements dans ce document sont protégés par la Loi sur le droit d'auteur, par les lois, les politiques et les règlements du Canada et des accords internationaux. Ces dispositions permettent d'identifier la source de l'information et, dans certains cas, d'interdire la copie de documents sans permission écrite. Pour obtenir de plus amples renseignements : http://lois.justice.gc.ca/fr/showtdm/cs/C-42
Internal Curing of High Performance Concrete Bridge Decks and Its Effects on Performance, Service Life and Life-Cycle Cost
Daniel Cusson; Lyne Daigle; Zoubir Lounis
National Research Coucil Canada, Institute for Research in Construction 1200 Montreal Road, Building M-20, Ottawa, Ontario, Canada, K1A 0R6
Corresponding author: daniel.cusson@nrc-cnrc.gc.ca, tel.: 1-613-998-7361, fax.: 1-613-952-8102
Abstract
High performance concrete (HPC) bridge decks are prone to premature cracking if movement is restrained. Internal curing (IC) can reduce shrinkage cracking in structures, thus improving their performance and service life. However, there is very limited information in the literature on the possible extension of service life due to internal curing. This paper addresses this question by using predictive models to estimate the service life and life-cycle costs of a typical bridge deck made with concrete using internal curing as opposed to conventional curing. Four options are compared: (i) normal concrete deck; (ii) HPC deck; (iii) HPC deck with internal curing; and (iv) very high performance concrete deck with internal curing. It was found that the use of internal curing can extend the service life of HPC bridge decks by almost ten years, which is mainly due to due a slower penetration of chlorides as a result of reduced shrinkage cracking. The use of internal curing in HPC decks resulted in a 20% reduction in life-cycle costs, even though the in-place unit cost of internally-cured HPC can be up to 4% higher than that of conventionally-cured HPC. The higher initial investment for the construction of the HPC-IC deck can be offset in only 5 years when compared to that required for normal concrete deck. This is due to a longer service life and less frequent maintenance activities.
Keywords: internal curing; high performance concrete; bridge deck; service life
1 Introduction
Proper curing of concrete structures is important to ensure that they meet their intended performance and design life requirements. In low permeability concrete, conventional external curing may not be effective in preventing self-desiccation at the centre of thick concrete elements. Internal curing (IC) is a technique that can be used to provide additional moisture inside the concrete for a more effective cement hydration and reduced self-desiccation (RILEM TC-196, 2007). For that purpose, saturated porous lightweight aggregate (LWA) can be mixed into concrete in order to supply an internal source of water, which can replace the mix water consumed by chemical shrinkage during cement hydration (Weber & Reinhardt 1997). As a result, autogenous shrinkage of high-performance concrete can be eliminated (Cusson & Hoogeveen 2008).
Internal curing with LWA has been successfully used recently in large construction projects of normal-density concrete structures. For example, in January 2005, about 190 000 m3 of internally-cured concrete was used in a large paving project in Hutchins, Texas (Villarreal & Crocker 2007). Field observations reported marginal pavement cracking, and tests indicated that 7-day flexural strengths reached 90% to 100% of the required 28-day flexural strength due to an improved cement hydration. They also found that the compressive strengths of air-cured cylinders were similar to those of wet-cured cylinders at all ages, suggesting that concrete with internal curing is less sensitive to poor external curing practices or unfavourable ambient conditions.
Although the benefits of internal curing for high performance concrete (HPC) structures have been evidenced in laboratory and field investigations (such as those previously mentioned), the literature does not provide any significant quantitative information regarding the extent of service life that can be achieved by internal curing of concrete structures. The objective of this paper is to provide reasonable estimates of service life and life cycle costs for a typical concrete bridge deck made with internally-cured HPC, using available test data from the literature, along with mechanistic models and conservative engineering judgment.
2
Effect of Internal Curing on Performance of HPC
2.1 Internal curing water requirement for elimination of self-desiccation
Based the pioneering work of Powers (1948) on chemical shrinkage of cement pastes, Figure 1 was proposed by Jensen & Hansen (2001) to estimate the amount of internal curing water required in concrete to prevent self-desiccation and the resulting autogenous deformations. The two solid lines represent the minimum quantity of internal curing water required to reach the maximum degree of hydration, which can be 1.0 (i.e. 100%) for water-cement ratios (w/c) above 0.36. Below this value, cement hydration can only be partially achieved, and can be estimated as (w/c)/0.36. Figure 1 also indicates that fully saturated conditions of hydrating cement pastes can be achieved with a quantity of internal curing water of 0.064 kg per kg of cement. The latter value does in fact correspond to the chemical shrinkage typically measured on ordinary Portland cement paste.
0.00 0.02 0.04 0.06 0.08 0.0 0.1 0.2 0.3 0.4 0.5 (w/c) tot (w /c ) IC 0.36 0.42 0.064 αmax = 1 IC water needed to reach max. hydration
w/c 0.36 αmax=
Figure 1 – Minimum amount of internal curing water needed to achieve maximum hydration and to prevent self-desiccation during cement hydration (adapted from RILEM TC-196, 2007).
When using cements with different values of chemical shrinkage, the IC water requirement may vary accordingly. For example, chemical shrinkage of silica fume is typically 0.22 L/kg. Consequently, a blended cement containing 10% silica fume and 90% ordinary Portland cement may experience a chemical shrinkage of 0.08 L/kg.
Knowing the degree of saturation and absorption capacity of the LWA, one can estimate the required mass of LWA to include in concrete to provide adequate internal curing. Zhutovsky et al. (2004) and Bentz et al. (2005) suggested similar forms of this equation:
LWA LWA c LWA S CS M M φ α ⋅ ⋅ ⋅ = max with 1 36 . 0 / max = ≤ c w α (Equation 1)
where MLWA is the dry mass of LWA per unit volume of concrete (kg/m
3
); Mc is the mass of
cementitious materials in concrete (kg/m3); CS is the chemical shrinkage (kg of water per kg of cement); αmax is the maximum expected degree of hydration; SLWA is the saturation
degree of LWA; φLWA is the absorption capacity of LWA (kg of water per kg of dry LWA);
and w/c is the water-cement ratio of concrete. For optimum performance, the LWA carrying the internal curing water should possess a high water absorption capacity and high sorption and desorption rates. Its particle size distribution and stiffness should be similar to those of the substituted normal-density aggregate. If the lightweight aggregate replacement ratio is rather small, these last two requirements may be neglected.
2.2 Effect of internal curing on net shrinkage and strength at early ages
A great concern for design engineers and contractors is whether the concrete will achieve the specified compressive strength and all durability requirements in the structure during service. Recent work by Cusson & Hoogeveen (2008) demonstrated that internal curing can reduce autogenous shrinkage considerably without affecting the strength and stiffness of high performance concrete. This was achieved by reducing the amount of mix water in the concrete by an amount equal to that used in the LWA for internal curing, thus reducing the effective water-cement ratio of the concrete when using higher quantities of LWA. The reference concrete selected for the demonstration had a 0.34 water-cement ratio, a cement-sand-coarse aggregate ratio of 1:2:2, and a cement content of 450 kg/m3. Three similar concretes were made with 3 different amounts of LWA sand: 6%, 12% and 20% of the total sand content, for which normal sand was partly replaced with pre-soaked LWA sand (expanded shale) for internal curing. The size of the HPC specimens were 200 x 200 x 1000 mm. Figure 2 presents the net shrinkage strain values (defined as the difference between the initial peak expansion strain and the most significant shrinkage strain measured in 7 days) and the 7-day compressive strengths (f’c 7d) obtained for the four
concrete mixes as a function of the sand replacement ratio. It can be seen that an increase in the amount of LWA sand used for internal curing resulted in a considerable reduction in the net autogenous shrinkage strain, where a sand replacement ratio of 22% could be regarded as one that can totally eliminate the net autogenous shrinkage and the resulting tensile stress for this concrete mixture. Figure 2 also shows that an increase in the amount of pre-soaked LWA sand resulted in a small increase in compressive strength, for the quantities of LWA tested in this study (under 180 kg/m3 of dry mass, before saturation).
-400 -300 -200 -100 0 0 5 10 15 20 25
Sand replacement ratio (%)
N e t s h ri n k a g e s tra in (x 1 0 -6 ) w/c eff = 0.34 w/c IC = 0.00 f'c 7d = 50 MPa w/c eff = 0.32 w/c IC = 0.02 f'c 7d = 50 MPa w/c eff = 0.30 w/c IC = 0.04 f'c 7d = 54 MPa w/c eff = 0.28 w/c IC = 0.06 f'c 7d = 57 MPa Mix-0 Mix-L Mix-M Mix-H
2.3 Effect of internal curing on deformation and stress development
Internal curing of concrete provided by means of pre-saturated structural-grade LWA can significantly enhance the early-age behaviour of concrete structures in terms of shrinkage and creep deformations, and tensile stresses, thus resulting in lower risks of cracking and longer service life. Figure 3 presents the total strains measured in the four unrestrained large-size specimens tested by Cusson & Hoogeveen (2008). The total strain includes autogenous shrinkage and thermal strains only, as the specimens were sealed to prevent external drying shrinkage. It can be readily seen that the addition of pre-soaked LWA for internal curing allowed early-age expansion to occur, which was due to swelling and thermal expansion, with peaks observed between 8 hours and 12 hours of age. The extent of expansion increased with the quantity of pre-soaked LWA used in the concrete mix. Mix-H was the only concrete producing positive values of total strain after the first day until the end of testing.
While the mechanisms of internal curing contributing to a reduction in autogenous shrinkage are relatively well known (RILEM TC-196, 2007), the mechanisms leading to an early-age expansion are not fully understood. Current shrinkage prediction models (ACI 209, 1992; CEB 1993) do not consider the possibility of early-age expansion in concrete structures. This expansion may be related to ettringite formation and/or swelling of the gel hydration products, as it has been hypothesized in previous studies (Bentz et al. 2001, 2008). It appears that there may be a competition between the development of expansion and that of autogenous shrinkage. As shown in Figure 3, the expansion clearly dominated autogenous shrinkage in the internally-cured HPC specimens at very early ages (<12 hours), since early shrinkage had been significantly reduced by internal curing. After the expansion peak, however, autogenous shrinkage developed faster than autogenous expansion, regardless of the type of curing used.
-300 -200 -100 0 100 200 300 0 24 48 72 96 120 144 168 Time after setting (hours)
T o ta l s tr a in ( x 1 0 -6 ) Mix-0 Mix-L Mix-M Mix-H Total strain due to autogenous
shrinkage and thermal effects
Figure 3 – Measured total strains in unrestrained concrete specimens
This early-age expansion had a significant effect on the size and sign of the measured basic creep strain developing over time on restrained specimens, as shown in Figure 4. For Mix-0 concrete, all the creep strains were positive (tensile creep). On the opposite, for Mix-H concrete with a high level of internal curing, all the creep strains were negative, with negative increments of creep strain during restrained expansion, and positive increments of creep strain during the following restrained shrinkage mode. For creep modelling and structural analysis purposes, the creep coefficient in this case cannot be determined with the conventional methods, which assume either constant stress (CEB 1993; ACI 209, 1992) or restrained shrinkage only (Kovler 1994). Cusson (2008) developed a procedure
for determining the creep coefficient from the time of setting under restrained expansion followed by restrained shrinkage. It was shown that creep coefficients of HPC determined experimentally with the proposed approach (Cusson 2008) were found to be more than 50% smaller than those predicted from existing creep models (ACI 209, 1992; CEB 1993) that are mainly based on compressive creep test results obtained on normal strength concrete specimens. -300 -200 -100 0 100 200 0 24 48 72 96 120 144 168 Time after setting (hours)
B a si c cr e e p st ra in ( x 1 0 -6 ) Mix-0 Mix-L Mix-M Mix-H
Increase in tensile creep (restrained shrinkage) Increase in compressive creep (restrained expansion)
Figure 4 – Measured basic creep strains in restrained concrete specimens
The main benefit of the initial expansion obtained in Mix-H due to the use of saturated LWA can be readily observed in Figure 5. Compared to 0, the initial expansion in Mix-H considerably delayed the onset of tensile stresses in concrete to a time at which the tensile strength was substantially higher, thus providing a lower stress/strength ratio at any given time (i.e. reduced risk of cracking).
-2 -1 0 1 2 3 4 5 0 24 48 72 96 120 144 1 Time after setting (hours)
S tre s s (MP a ) Mix-0 Mix-H
(almost a third of the stress was due to auto
68 genous shrinkage) (almost none of the stress was due to autogenous shrinkage)
Thermal effects were important (but similar) in both mixes
Tensile strength
Figure 5 – Measured uniaxial stresses in restrained concrete specimens
In this experiment, the thermal effects originated from the heat of cement hydration and the subsequent cooling producing thermal stresses under restraint. Measurements of temperatures and strains indicated that thermal effects accounted for a large part of the total concrete stresses. In the case of Mix-H concrete specimen, most of the stresses were due to thermal effects since autogenous shrinkage strains were very small compared to the thermal strains.
In light of the above, if thermal effects can be controlled, and drying and autogenous shrinkage prevented by conventional and internal curing, then the risk of concrete cracking can be effectively reduced, resulting in potentially more durable concrete structures.
3
Case study – Highway Bridge Deck
To demonstrate the benefits of internal curing in high performance concrete structures in terms of extended service life and reduced life cycle costs, a typical concrete bridge deck case study was used for the analysis. The bridge element under consideration is a 200-mm thick reinforced-concrete slab (with an effective depth of 160 200-mm), which is part of a three-lane bridge as illustrated in Figure 6. The slab has a steel reinforcement ratio of 0.3% for both top and bottom mats in both transverse and longitudinal directions (CSA 2006) and a concrete cover thickness of 75 mm for the top reinforcement. The concrete deck has a width of 13 m and a length of 35 m between expansion joints, and is carried by simply-supported prestressed concrete girders. The bridge superstructure and substructure are not considered in the analysis.
Cross-section of bridge Elevation view of bridge
Cross-section of bridge slab
Width = 13 m 200 mm 75 mm 200 mm Pier Length = 35 m Girder Slab 30 mm
Cross-section of bridge Elevation view of bridge
Cross-section of bridge slab
Width = 13 m 200 mm 75 mm 200 mm Pier Length = 35 m Girder Slab 30 mm
Figure 6 - Geometry of case study bridge deck
Table 1 presents four mix design options for the concrete decks compared in this case study, including: (i) bridge deck made of normal concrete (NC) with a water-to-cementitious materials ratio (w/cm) of 0.4, and no supplementary cementing materials (SCM); (ii) deck made of high performance concrete (HPC), assuming early-age cracking due to autogenous shrinkage, with a w/cm of 0.35 and 25% SCM as partial cement replacement by mass; (iii) deck made of high performance concrete with internal curing (HPC-IC), also with a w/cm of 0.35 and 25% SCM; and (iv) deck made of very high-strength high performance concrete with internal curing (VHPC-IC), with a lower w/cm of 0.30, and 25% SCM. In this case study, the SCM consisted of 5% silica fume and 20% slag in the cement; and the required mass of LWA was calculated using Equation 1, assuming a chemical shrinkage value of 0.07 kg of water per kg of cement, and a free water content of 15% in the LWA for internal curing.
Table 1 – Mix design of selected concrete bridge deck options Deck option Initial
cracking Water (kg/m3) Cement (kg/m3) SCM (%) LWA (kg/m3)
In-place concrete unit cost ($/m3)
NC No 140 350 0 0 450
HPC Yes 160 450 25 0 600
HPC-IC No 160 450 25 200 625
VHPC-IC No 160 525 25 200 750
With respect to exposure conditions, the bridge deck is directly exposed to de-icing salts during winter periods. A surface chloride concentration of 9 kg/m3 is assumed, which is typical of the severe conditions found in Canada and Northern USA (Weyers 1998). A chloride threshold of 0.7 kg/m3 is assumed for the conventional steel reinforcement (ACI 222, 2001), above which corrosion can initiate; and a moderate corrosion rate of 0.5 μA/cm2
is assumed once corrosion has started (Rodriguez et al. 1994). It can be argued that these values may differ from actual values depending on the severity and variability of the local exposure conditions; however, the findings of this case study should be considered with respect to the normal concrete deck (base case), since all deck options are compared using the same values of surface chloride concentration, chloride threshold, and corrosion rate, respectively.
4
Service life of concrete bridge decks
In this study, most life-cycle performance predictions were made using the deterministic and reliability-based analytical models developed by Lounis et al. (2006), and implemented in NRC’s SLAB-D software (Daigle & Lounis 2006). Furthermore, the effects of early-age cracking and internal curing on chloride penetration, which are not part of the original
SLAB-D software, were estimated using additional models presented in this paper.
Figure 7 presents the different stages of corrosion-induced damage developing in a typical reinforced concrete bridge deck, which is based on a two-phase (i.e. corrosion initiation and propagation) simplified model, first introduced by Tuutti (1982). With time, each stage develops into a higher level of damage. The different stages include: (i) early-age cracking of concrete due to restrained shrinkage (if any); (ii) initiation of reinforcement corrosion after a relatively long period of chloride diffusion through concrete; (iii) internal cracking around the reinforcing bars due to the build-up of corrosion products; (iv) surface cracking due to further progression of corrosion-induced cracks; (v) delamination or spalling of the concrete cover, depending on the slab reinforcement design; and finally (vi) failure of the concrete deck, depending on the amount of concrete damage that can be tolerated by the bridge owner before deck rehabilitation or replacement is considered.
Time Da ma ge l eve l Initial cracking Rebar corrosion Internal cracking Surface cracking Delamination or spalling Failure Service life
Corrosion initiation Propagation
Cl-Cl- Cl- Cl -Cl
-Shrinkage
cracking diffusionChloride stress generationCorrosion and Concretedamage
Time Da ma ge l eve l Initial cracking Rebar corrosion Internal cracking Surface cracking Delamination or spalling Failure Service life
Corrosion initiation Propagation
Cl-Cl- Cl- Cl -Cl
-Shrinkage
cracking diffusionChloride stress generationCorrosion and Concretedamage
4.1 Initial cracking due to restrained shrinkage
Early-age cracking is a common problem on concrete bridge decks (TRB 1996) as more than 100,000 bridges in the United States are reported to have developed transverse cracking shortly after construction. This type of through cracking is of great concern to engineers and bridge owners since it may lead to premature reinforcement corrosion and concrete deterioration due to accelerated moisture and salt ingress through the cracks. Rodriguez & Hooton (2003) studied the influence on chloride penetration of artificially created parallel cracks with widths ranging from 0.08 mm to 0.68 mm. They concluded that chloride diffusion in concrete was not affected by the width and wall roughness of individual cracks for the ranges studied. This finding allows the use of a simplified smeared approach proposed by Boulfiza et al. (2003) to estimate the effect of cracks on chloride ingress. It assumes that chloride ingress into cracked concrete can be approximated using Fick’s second law of diffusion, in which an apparent diffusion coefficient (Dapp) is first
calculated with this equation:
cr cr cr c app D s w D D = + (Equation 2)
where Dc is the chloride diffusion coefficient in uncracked concrete; wcr is the crack width;
scr is the crack spacing; and Dcr is the diffusion coefficient inside the crack, which is
assumed to be 5x10-10 m2/s (Boulfiza et al. 2003). In the present case study, the value of
wcr/scr is set to 0.0003, which corresponds to the 7-day value of shrinkage strain measured
for a concrete that was very similar to that selected for the HPC deck option (Cusson & Hoogeveen 2008). As a result, an average increase in the chloride diffusion coefficient of approximately 1.5x10-13 m2/s can be expected. For the decks made with NC, HPC-IC and VHPC-IC, the effect of cracking on chloride ingress was neglected (i.e. Dapp = Dc),
assuming that autogenous shrinkage in these concretes is not significant enough to result in concrete cracking.
4.2 Corrosion initiation due to chloride attack
In uncracked concrete, the chloride ingress was determined by using Crank’s solution of Fick’s second law of diffusion (Crank 1975):
⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ − = t D x erf C t x C S 2 1 ) , ( (Equation 3)
where C(x,t) is the chloride concentration at depth x after time of exposure t; Cs is the
chloride concentration at the concrete surface; and erf is the error function. Equation 3 can also be rearranged to predict the time of corrosion initiation (ti) by setting C(x,t) equal to a
chloride threshold value (Cth), at which the corrosion of the reinforcing steel is expected to
initiate, and x equal to the effective cover depth of the reinforcing steel (dc).
The chloride diffusion coefficients of the different concretes were estimated based on their water-cement ratios and the types of SCM selected for this case study. The empirical models of Boulfiza et al. (2003) developed from a large set of literature data were used:
0 . 14 ) / ( 2 . 7 ) / ( 9 . 3 2+ − − = w cm w cm D Log C 2
(with no SCM) (Eq. 4a) 7 . 13 ) / ( 4 . 5 ) / ( 0 . 3 + − − = w cm w cm D
where Dc is expressed in units of m2/s. Table 2 presents the resulting coefficients of
diffusion calculated for the different concretes considered, including the effect of initial cracking on the apparent chloride diffusion coefficient (Dapp).
Table 2 – Values of w/cm and chloride diffusion coefficients Deck option w/cm Dc Dapp
(10-13 m2/s)
NC 0.40 18 18
HPC 0.35 6.6 8.1
HPC-IC 0.35 6.6 6.6
VHPC-IC 0.30 4.4 4.4
It is important to note that the estimated values of Dc for the HPC-IC and VHPC-IC
concretes are considered to be conservative values, since internal curing can also enhance cement hydration, thus reducing concrete permeability and diffusivity. Indirect observations of reduced permeability/porosity due to internal curing through improved early-age strengths can be found in the literature (Bentz 2007). Since there is not enough test data available in the literature to model the effect of internal curing on chloride diffusivity of concrete, this beneficial effect of internal curing was not accounted for.
Figure 8 presents the chloride profiles calculated for the four different concrete bridge deck options after a 20-year exposure to de-icing salts. The critical zone corresponds to chloride concentrations at the reinforcement level (or deeper) that are higher than the chloride threshold of 0.7 kg/m3, over which initiation of steel corrosion is likely to occur. It can be seen that after 20 years, the chloride concentration in the normal concrete deck has reached the critical concentration at the 75-mm depth.
0 1 2 3 4 5 6 7 8 9 10 0 25 50 75 100 Depth (mm) C hl o ri de con te nt ( k g/ m 3) Rebar depth Chloride threshold NC HPC VHPC Critical zone 20 years
Figure 8 - Chloride profiles in concrete slab after 20 years
4.3 Corrosion-induced damage initiation and growth
Assuming that concrete in tension behaves like a homogeneous elastic material (with a tensile strength = f’t), and that corrosion products are equally distributed around the
perimeter of the reinforcing bars (Bažant 1979), stresses generated in the concrete deck by corrosion products can be estimated using the thick-wall cylinder model (Timoshenko 1956). This model allows the calculation of the increase in the rebar diameter (Δd) related
Corrosion products σ r+ dσr σr σ t σt σt=f’t (a) (b) Corrosion products σ r+ dσr σr σ t σt σt=f’t (a) Corrosion products σ r+ dσr σr σ t σt σt=f’t (a) (b)(b)
Figure 9 - Thick-wall cylinder model of corroding reinforced concrete: (a) Tensile stresses developed at crack initiation; (b) Propagation of internal cracks in thick-wall concrete cylinder. Adapted from Lounis et al. 2006
The corrosion propagation times (tp), corresponding to the onset of internal cracking,
surface cracking, and delamination/spalling, respectively, can be found as follows:
⎥ ⎦ ⎤ ⎢ ⎣ ⎡ − Δ = s r r p j S d d t ρ α ρ π 1 2 ) ( (Equation 5)
where d is the rebar diameter; S is the rebar spacing; jr is the rust production rate per unit
area (Bažant 1979); ρr is the density of corrosion products (assumed at 3600 kg/m3 for
Fe(OH)3); ρs is the density of steel (7860 kg/m3); and α is the molecular weight ratio of
metal iron to the corrosion product (assumed at 0.52). The total time to reach a given corrosion-induced damage level is the sum of the corrosion initiation time (ti) and the
individual corrosion propagation times (tp) up to that level.
Work by Allan & Cherry (1992) showed that there may be a porous zone around the steel reinforcement at the steel/concrete interface, in which a specific quantity of corrosion products can accumulate before tensile stresses actually develop in concrete, thus delaying the damage initiation of the concrete cover. The thickness of the porous zone may be in the order of 12.5 μm (Liu & Weyers 1998); however, the thickness and porosity of the porous zone depend on many factors, such as: w/cm, presence and type of SCM, and hydration, which in turn influence the overall concrete porosity.
Internal curing may affect the porous zone in two competing ways: (i) by decreasing the porosity of the cement paste due to an improved cement hydration; and (ii) by increasing the porosity of the zone due to the relatively large and empty pores left in the LWA after the internal curing water has migrated into the cement paste. Due to the lack of data in the literature on the effect of internal curing and w/cm ratio on the thickness and porosity of the porous zone, this possible beneficial effect has been neglected in the analysis. Preliminary calculations indicate that neglecting the effect of the porous zone would provide a more conservative estimate of service life by at most 2 or 3 years only.
4.4 Estimated
service
life
In this case study, the service life of a bridge deck is defined as the time to initiate delamination or spalling (which ever comes first). Figure 10 illustrates the average service life estimated for each concrete deck option, in which the times to reach lower levels of concrete damage are also illustrated. Based on the above hypotheses, the service life increased due to the use of a lower water-cement ratio (reduced chloride diffusion), in general, and the reduction of early-age cracking due to internal curing. The service lives were 21 years for the 0.40 w/cm NC deck; 41 years for the 0.35 w/cm HPC deck; 49 years for the 0.35 w/cm HPC-IC deck; and 71 years for the 0.30 w/cm VHPC-IC deck. For simplicity, other factors such as live loads, dead loads and thermal effects were not considered in this comparative analysis.
0 10 20 30 40 50 60 70 80
NC deck HPC deck HPC-IC deck VHPC-IC deck
T im e ( year s) Onset of spalling Onset of surface cracking Onset of internal cracking Onset of corrosion
21 years
41 years
49 years
71 years
Figure 10 - Service life predictions from deterministic analysis
The service life predictions presented in Figure 10 were obtained through a deterministic analysis, i.e. an analysis that do not consider the variability and uncertainty associated with the main parameters governing the different mechanisms of deterioration in reinforced concrete decks. Deterministic analyses are based on mean or characteristic values of the variables and can only predict the times to reach the different stages of corrosion initiation and corrosion-induced damage caused by an average condition.
The variability and uncertainty of the main parameters (e.g. material properties, structure dimensions, reinforcement, initial conditions, and environment) that govern service life are the sources of the observed variable performance of concrete decks in the field, which is better described with a reliability-based approach. In this paper, the advanced first-order reliability method is used to determine the Hasofer-Lind reliability index and the corresponding probability of failure for the different nonlinear performance functions that describe the different failure modes at different points in time (Lounis et al. 2006), which was implemented in the SLAB-D software.
The four concrete deck options were also analyzed using the reliability method, considering the variability of six key parameters (Table 3) expressed in terms of their averages and coefficients of variation (COV).
Table 3 - Average and COV values of main parameters
Parameter Average COV
(% ) Surface chlorides (kg/m3) 9 30 Diffusion coef. (10-13 m2/s) - NC - HPC - HPC-IC - VHPC-IC 18 8.1 6.6 4.4 30 30 30 30 Concrete cover depth (mm) 75 30
Chloride threshold (kg/m3) 0.7 30 Corrosion rate (μA/cm2
) 0.5 30
Bar spacing (mm) 200 5
Figure 11 presents the results from the reliability analysis for the four deck options, in which the probability of spalling increased over time. The investigated governing parameters exhibit considerable spatial variability that may require the use of random fields instead of random variables for their modeling (Lounis et al. 2006). The use of random fields, however, requires more input data such as the correlation function and scale of fluctuation, and adds considerable analytical and computational complexity to the analysis. If the variability of the parameters can be assumed to be completely random, i.e. no systematic variability and no spatial correlation, then the use of random variables for modeling the above parameters is adequate (Lounis et al. 2006). Considering that the bridge deck can be divided into a very large number of small surface areas with same probability (P%) of spalling after a number of years (t), it can be approximated that a certain percentage (P%) of them are spalled. The 10%, 25% and 50% probabilities of spalling are highlighted in Figure 11. These probabilities can be used to make initial and approximate comparisons to the five condition states for concrete deck condition assessment (Table 4) defined in the AASHTO (1998) Guide, which is used in the bridge management systems of several states in the USA and some Canadian provinces.
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0 10 20 30 40 50 60 7 Time (years) P ro b ab ili ty o f sp allin 0 g NC HPC HPC-IC VHPC-IC 21 yrs 41 yrs 49 yrs
71 yrs
Figure 11 - Time-dependent probability of spalling
Table 4 - AASHTO description of bridge deck condition states
Condition state Description
1 The surface of the deck has no patched area and no spall in the deck surface. 2 The combined distress area (i.e. existing patches, delamination and spalling)
of the deck is less than 10%.
3 The combined distress area of the deck is between 10% and 25%. 4 The combined distress area of the deck is between 25% and 50%. 5 The combined distress area of the deck is more than 50%.
For the normal concrete deck, the 10%, 25% and 50% probabilities of spalling are reached after 9, 14, and 21 years. For the HPC, HPC-IC and VHPC-IC decks, these same probabilities are reached after 16, 26 and 41 years; 19, 31 and 49 years; and 26, 44, and 71 years, respectively. As the apparent coefficients of diffusion of the four investigated deck options decreased, the times to reach critical condition states increased accordingly. Depending on the bridge owner policy and the estimated costs and benefits of repairs, the optimum recommended action to take for each deck condition state can be different, ranging from the do-nothing approach, to delamination and spalling repairs, and ultimately to deck replacement. In addition to the extended service life provided by reduced initial cracking and reduced chloride diffusion coefficients, the reliability-based analysis also shows that the use of lower w/cm concretes with internal curing results in longer periods of time before major repair or rehabilitation activities become necessary, as indicated by the times needed to reach critical condition states.
5
Life cycle cost of HPC bridge decks
A life cycle cost analysis (LCCA) is conducted when there is a need to evaluate and compare the economic performance of competing design and maintenance alternatives. A LCCA was conducted to assess the total life-cycle costs of the four alternative bridge decks, in which many different types of activities may be scheduled at different points in time, for example: periodic inspections, required maintenance and repairs, as well as deck replacement. This is usually achieved by calculating the present value life cycle costs (PVLCC) of the alternatives over a given time period (Grant et al. 1990; Hawk 2003):
∑
= + − + + = T t T v t i i i r R r C C PVLCC 1 0 ) 1 ( ) 1 ( (Equation 6)where C0 is the initial construction cost; Ci is the ith expenditure at time ti; r is the discount
rate; T is the analysis period; and Rv is the residual value of the alternative at the end of
the analysis period. In this study, the analysis period was set to 70 years for all cases, with an assumed discount rate of 3%. The direct costs incurred by the bridge owner usually include initial construction costs and other costs associated with the maintenance activities. In this case study, the construction costs included the cost of reinforcing steel ($2/kg) and the in-place cost of concrete (Table 1), thus considering formwork installation, detailing of reinforcing steel, placing and surface finishing of concrete, and form stripping. The in-place cost of concrete will depend on many factors, such as the type and quantity of cement, aggregates, supplementary cementing materials and admixtures used in the concrete mix, and their availability. The unit cost of HPC ($600/m3) was estimated to be 33% higher than that of normal concrete ($450/m3), mainly due to the increased quantity of cement in the mix. The unit cost of HPC-IC was set to that of HPC plus a 4% additional cost to account for the cost difference associated with the purchase and transportation of the lightweight aggregate sand (with a purchase cost of $75/ton) used to substitute a fraction (200 kg/m3) of the normal-density fine aggregate (with a purchase cost of $15/ton). Note that this small 4% cost increase could have been easily offset by using slightly less cement in the concrete mix, knowing that the more effective hydration due to internal curing would likely achieve similar concrete strength and stiffness. Finally, the unit cost of VHPC-IC was increased to $750/m3 in order to account for the increased quantity of cement in the concrete mix. In this case study, a set of various maintenance activities were assumed for the bridge decks over their life cycles. Figure 12a illustrates those scheduled for the NC deck. For instance, routine inspections were scheduled at two-year intervals, while non-destructive evaluation and protection activities were scheduled every five years.
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 Time (years)
Initial service life
70 2nd service life Replacement Patch repair Protection Non-destructive evaluation Routine inspection Time (years) Init. service life 2nd
service life 3rd service life 4th life Replacement Patch repair Protection Non-destructive evaluation Routine inspection
(a) NC deck (b) HPC deck
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60
Time (years) Initial service life
(c) HPC-IC deck (d) VHPC-IC deck Figure 12 – Optimised maintenance schedules for the different deck alternatives
While it was assumed that autogenous shrinkage cracking could be neglected for normal concrete, protection activities were scheduled to correct other problems such as drying shrinkage, freeze-thaw cycles, etc. Major patch repairs (incl. old concrete removal, surface preparation, patching, traffic control) were scheduled to take place approximately when 10% and 25% of the deck surface would be spalled (based on Figure 11). In this study, replacement was deemed necessary when 50% of the deck surface would be spalled. After replacement, it was assumed that the bridge deck would be rebuilt with a similar type of deck of a similar initial construction cost. Figures 12b, 12c and 12d show the scheduled activities for the decks using HPC, HPC-IC and VHPC-IC, respectively. The major difference with the NC deck is that the non-destructive evaluation and protection activities were scheduled to occur every 10 years, since high-performance concrete is expected to perform better than normal concrete, in general. Table 5 presents the agency costs associated with the difference maintenance activities. Figure 13 illustrates the present value life cycle cost incurred by the agency for each deck option.
Table 5: Maintenance costs to agency
Activity Unit cost
($/m2) Routine inspection 2 Non-destructive evaluation 20 Protection 20 Patch repair 200 Replacement (disposal and reconstruction)
350 + unit cost of reconstruction
Time (years) Initial service life
70 Replacement Patch repair Protection Non-destructive evaluation Routine inspection 2nd service life Replacement Patch repair Protection Non-destructive evaluation Routine inspection
324 380 471 870 0 100 200 300 400 500 600 700 800 900 1000
NC deck HPC deck HPC-IC deck VHPC-IC deck
U n it cost ( $ /m 2 ) 46% reduction 56% reduction 63% reduction
Figure 13 - Present value life-cycle costs of the different deck options
The PVLCC for the normal concrete deck is the highest at $870/m2, mainly due to the shorter service life and the more frequent maintenance and replacement activities. The HPC deck (no internal curing) reduced this cost by 46%, mainly due to the longer service life even though cracking was assumed to occur at early ages. The HPC-IC deck further reduced the PVLCC down to $380/m2, which is 56 % less expensive than the NC deck, or 20% less expensive than the HPC deck due to internal curing alone. When using very high-strength high performance concrete with internal curing, the PVLCC is the lowest at $324/m2, mainly due to the very long service life of 70 years obtained from the use of a very low w/cm and adequate internal curing.
Figure 14 shows the present value cumulative expenditures incurred by the bridge owner over the 70-year analysis period for the NC and HPC-IC deck alternatives. It is clearly shown that the higher initial investment in the HPC-IC deck, compared to the NC deck, can be offset in only 5 years due to the lower maintenance costs associated with HPC-IC, which represents only 10% of its 49-year initial service life.
0 100 200 300 400 500 600 700 800 900 1000 0 10 20 30 40 50 60 7 Time (Years) C o s t ($ /m 2 )
Normal concrete dec
0 k HPC deck with internal curing 5 years to breakeven Initial service life of NC deck
Initial service life of HPC-IC deck
6 Summary
and
Conclusions
The performance of HPC structures can be substantially improved by the use of pre-saturated structural-grade LWA for internal curing that can reduce autogenous shrinkage, tensile stress development and cracking of concrete. The life-cycle performance of four bridge deck alternatives made with different concretes were estimated using analytical predictive models for chloride diffusion, corrosion initiation, and corrosion propagation leading to cracking, delamination or spalling. The concretes selected for the bridge decks were: (i) normal concrete deck (w/cm=0.40); (ii) HPC deck (w/cm=0.35); (iii) HPC deck with internal curing (w/cm=0.35); and (iv) very high-strength HPC deck with internal curing (w/cm=0.30). From this study, these conclusions can be drawn:
• The use of internal curing can increase the service life of high performance concrete bridge decks by almost ten years, due to the prevention of initial cracking.
• This estimated life extension due to internal curing alone is considered to be on the conservative side, since the beneficial effect of internal curing on the reduction of concrete permeability to water and chloride ions was not accounted for in this analysis, due to lack of available data.
• The service life of concrete bridge decks can be extended by 50 years when considering the use of internally-cured very high-strength high performance concrete (w/cm=0.30) over normal concrete (w/cm=0.40).
• The life-cycle cost of a bridge deck can be considerably reduced when using high-performance concrete over normal concrete, especially with internal curing. This can be attributed to fewer maintenance activities and a longer service life;
• The use of internal curing compared to conventional curing in high-performance concrete decks resulted in a 20% reduction in life-cycle cost, even with a 4% higher in-place unit cost for the internally-cured HPC;
• The higher initial investment required for the construction of the HPC-IC deck can be offset in only 5 years when compared to that required for the NC deck. This is due to less frequent maintenance activities and a longer service life.
7 References
AASHTO, “Guide for Commonly Recognized (CoRe) Structural Elements”, American Association of State Highway and Transportation Officials, 1998 (Rev. 2002), 54 p. ACI 209 “Prediction of creep, shrinkage, and temperature effects in concrete structures”,
American Concrete Institute, ACI Report 209R-92, Farmington Hills, 1992, 47 p.
ACI 222, “Protection of Metals in Concrete Against Corrosion”, American Concrete Institute, ACI Report 222R-01, Farmington Hills, 2001, 41 p.
Allan, M.L. & Cherry, B.W., “Factors controlling the amount of corrosion for cracking in reinforced concrete”, Journal of Corrosion Engineering, 1992, p. 426-430.
Bažant, Z.P., “Physical model for steel corrosion in concrete sea structures – theory”.
Journal of the Structural Division, Vol. 105(ST6), 1979, p. 1137-1153.
Bentz, D.P., “Internal curing of high-performance blended cement mortars”, ACI Materials
Journal, V. 104, No. 4, July-August 2007, p. 408-414.
Bentz, D.P., Jensen, O.M., Hansen, K.K., Olesen, J.F., Stang, H. & Haecker, C.J., Influence of cement particle-size distribution on early-age autogenous strains and stresses in cement-based materials, Journal of the American Ceramic Society, Vol. 84, No. 1, 2001, p. 129-135.
Bentz, D.P., Lura, P. & Roberts, J.W., “Mixture proportioning for internal curing”, Concrete
Bentz, D.P., Sant, G., & Weiss, W.J., "Early-Age Properties of Cement-Based Materials: I. Influence of Cement Fineness," ASCE Journal of Materials in Civil Engineering, Vol. 20, No. 7, 2008, p. 502-508.
Boulfiza, M., Sakai, K., Banthia, N. & Yoshida, H., “Prediction of chloride ions ingress in uncracked and cracked concrete”, ACI Materials Journal, Vol. 100, No. 1, January-February 2003, p. 38-48.
CEB “CEB-FIP Model Code 1990”, Information Bulletin No. 213/214, Euro-International Concrete Committee, Lausanne, 1993, 437 p.
Crank, J., The mathematics of diffusion, 2nd edition, Oxford University Press, London, 1975, 414 p.
CSA, Canadian Highway Bridge Design Code, CAN/CSA-S6-06, Canadian Standard Association International, Rexdale, Ontario, 2006.
Cusson, D., “Early-age viscoelastic behavior of high performance concrete with internal curing”, 8th International Conference on Creep, Shrinkage and Durability of Concrete and Concrete Structures, Ise-Shima, Japan, Sept. 30–Oct. 2, 2008, 6 p.
Cusson, D. & Hoogeveen T., “Internal curing of high-performance concrete with pre-soaked fine lightweight aggregate for prevention of autogenous shrinkage cracking”,
Cement and Concrete Research, Vol. 38, No. 6, 2008, p. 757-765.
Daigle, L. & Lounis, Z., “SLAB-D Version 1.0 User’s Manual (and software) – Service Life Analysis of Bridge Decks”, National Research Council Canada, Institute for Research in Construction, 2006, 119 p.
Grant, E., Ireson, W. & Leavenworth, R., Principles of Engineering Economy, 8th ed., John Wiley & Son, New-York, 1990
Hawk, H, “Bridge Life-Cycle Cost analysis”, NCHRP Report 483, Transportation Research Board, 2003.
Jensen, O.M. & Hansen, P.F., “Water-entrained cement-based materials: 1. Principle and theoretical background”, Cement & Concrete Research, V.31, No.4, 2001, p. 647-654. Kovler, K. “Testing system for determining mechanical behaviour of early-age concrete
under restrained & free uniaxial shrinkage”, Materials & Structures, 27, 1994, 324-330. Liu, Y. & Weyers, R.E., “Modeling time-to-corrosion cracking in chloride contaminated
reinforced concrete structures”, ACI Materials Journal, Vol.95, No.6, 1998, p. 675-681. Lounis, Z., Martín-Pérez, B., Daigle, L. & Zhang, J., “Decision support tools for service life
prediction and rehabilitation of concrete bridge decks”, NRC-IRC Final Client Report
No. B5318.2, April 2006, 256 p.
Powers, T.C. & Brownyard, T.L., “Studies of the physical properties of hardened Portland cement paste”, Journal of the American Concrete Institute, Vol. 18, No. 3, November 1948, p. 249-336.
RILEM TC-196, “Internal curing of concrete”, State-of-the-art Report of RILEM Technical
Committee 196-ICC, Edited by K. Kovler and O.M. Jensen, RILEM Publications
S.A.R.L., Bagneux, France, 2007, 139 p.
Rodriguez, O.G. & Hooton, R.D., “Influence of cracks on chloride ingress into concrete”,
ACI Materials Journal, Vol. 100, No. 2, March 2003, p. 120-126.
Rodriguez, J., Ortega, L.M. & Garcia, A.M., “Assessment of structural elements with corroded reinforcement,” International Conference on Corrosion, Sheffield, U.K., July 1994, 16 p.
Timoshenko, S.P., Strength of Materials. Part II: Advanced Theory and Problems, Princeton, NJ, 1956.
TRB, “Transverse cracking in newly constructed bridge decks”, National Co-operative
Highway Research Program Report 380, Transportation Research Board, National
Academy Press, Washington, 1996, 126 p.
Tuutti. K., “Corrosion of Steel in Concrete”, Ph.D. Thesis, Swedish Cement and Concrete Research Institute, Stockholm, Sweden, 1982, 469 p.
Villarreal, V.H. & Crocker, D.A., “Better pavements through internal hydration”, Concrete
International, February 2007, p. 32-36.
Weber, S. & Reinhardt, H.W., “New generation of high performance concrete: concrete with autogenous curing”, Advanced Cement Based Materials, Vol.6, No.2, 1997, p. 59-68.
Weyers, R.E., “Service life model for concrete structure in chloride laden environments”,
ACI materials journal, Vol. 95, No. 4, July-August 1998, p. 445-453.
Zhutovsky, S., Kovler, K. & Bentur, A., “Influence of cement paste matrix properties on autogenous curing of high-performance concrete”, Cement and Concrete Composites, Vol. 26, 2004, p. 499-507.
