Control of Concrete Cracking in Bridges (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

54 One of the intents of providing concrete cover to steel reinforcement is to protect the reinforcement from direct contact with materials that will cause corrosion, such as saltwater and deicing chemicals. If cracks occur, that protection is lost, and this is the concern that owners have, particularly for bridge decks and substructures. However, the following factors determine if corrosion actually occurs: • Is the concrete surface exposed to chemicals that will cause the reinforcement to deteriorate? • Can the chemicals reach the reinforcement? • If the chemicals reach the reinforcement, will they cause deterioration? If a crack is present and exposed to harmful chemicals, the chemicals will reach the level of the reinforcement much earlier than if a crack is not present. Miller and Darwin (2000) reported signifi- cantly higher chloride contents at the locations of cracks at the level of the transverse reinforcement. The chloride contents exceeded the threshold level of corrosion in as little as 1,000 days. Lindquist et al. (2005) reported that chloride concentrations taken at the level of the top transverse reinforce- ment at crack locations can exceed the corrosion threshold in as little as 9 months and most bridge decks exceeded the threshold level in 24 months. In contrast, the chloride levels away from the cracks rarely exceeded even the most conservative estimates of the corrosion threshold for conventional reinforcement. According to ACI Committee 222 (2001), the role of cracks in the corrosion of reinforcing steel is a matter of controversy. One viewpoint is that cracks permit deeper and rapid penetration of carbon- ation, chloride ions, moisture, and oxygen. The other viewpoint is that chloride ions eventually pen- etrate uncracked concrete, and the resulting corrosion of the reinforcement is more widespread. So, after a few years, there is little difference between the amount of corrosion in cracked and uncracked concrete. However, with the lower-permeability concrete being used today, the length of time for chlorides to penetrate uncracked concrete is greater. Cracks perpendicular to reinforcing bars hasten corrosion of the reinforcement at the crack location by facilitating the ingress of moisture, oxygen, and chloride ions (Russell 2004). Studies have shown that crack widths of less than 0.01 in. have little effect on the overall corrosion of the reinforcing steel (Houston et al. 1972; Ryell and Richardson 1972). Although wider cracks may accelerate the onset of corrosion over several years, crack width has little effect on the rate of corrosion (Beeby 1983). Cracks that follow the line of a reinforcing bar are more serious because the length of the bar exposed to the ingress of moisture, oxygen, and chlorides is equal to the length of the crack. This crack- ing can initiate as settlement cracking, or the bars can create a weakened plane. In addition, the presence of the cracks reduces the resistance of the concrete to spalling if the reinforcement corrodes. Beeby (1983) reported that no consensus exists regarding the levels of cover, concrete quality, and permissible crack width that should be specified. Beeby concluded that crack widths have little influence on corrosion, and many design recommendations require unnecessary detailed calculations for crack control as a corrosion control method. Fanous et al. (2000) collected concrete cores from cracked and uncracked areas of bridge decks in Iowa to determine the extent to which the epoxy-coated reinforcement had deteriorated at the location chapter six Influence of crackIng on long-Term BrIdge Performance

55 of cracks and evaluate the impact of cracking on service life. No delaminations or spalls were found in bridge decks constructed with epoxy-coated reinforcement. The oldest bridge deck was 20 years old. All of the reinforcing bars extracted from uncracked locations showed no evidence of corrosion. Most of the corrosion on the epoxy-coated bars was on bars extracted at cracked locations. For a corrosion threshold range of 3.6 to 7.2 lb/yd3 for epoxy-coated reinforcement, the predicted service life for Iowa bridge decks with epoxy-coated reinforcement was more than 50 years. Rodriguez and Hooton (2003) investigated the influence of crack widths and crack surface rough- ness on chloride ingress into concrete. Smooth and rough crack surfaces with widths ranging from 0.003 to 0.027 in. were exposed to a chloride bulk diffusion test. They concluded that chloride diffu- sion was independent of crack width or crack wall roughness. The cracks behaved like free concrete surfaces, greatly contributing to lateral chloride diffusion. In Pennsylvania, concrete cores were extracted from 19 bridge decks and evaluated in the labora- tory to determine the chloride content at the reinforcement level (Manafpour et al. 2016). For each deck, an attempt was made to take one core sample at a crack location and one core off a crack location. The chloride content at the reinforcement level for the on-crack locations was found to be as much as 10 times greater than that for the samples from the off-crack locations in the same deck. The chloride content was used to calculate the deck’s effective diffusion coefficient according to the Fick’s second law, accounting for age and surface concentration of chlorides. On-crack effective dif- fusion coefficients typically were as much as four times greater than off-crack values. Brown et al. (2003) investigated the size and length of cracks in Virginia bridge decks to assess the frequency and severity of the cracks. Correlation of cracks with chloride penetration was used to characterize the influence of cracking on deck deterioration. Cracks influenced the rate of chloride penetration, but the frequency and width distributions of cracks indicated that cracks are not likely to shorten the overall service life of most bridge decks in Virginia. It is reasonably well established that a chloride content of approximately 1.0 to 1.5 lb/yd3 at the level of the reinforcement will initiate corrosion of uncoated steel reinforcement (ACI Committee 222 2001). Although some data exist, it is not well established what level of chlorides can exist before coated steel reinforcement begins to corrode and what levels solid corrosion-resistant bars can tolerate; it is always assumed that a limit exists. Sim (2014) compared the performance of different types of steel reinforcement using macrocell specimens with cracked concrete specimens and ranked 10 types of steel reinforcement in order of corrosion performance. In general, stainless steels showed the most effective performance. McDonald et al. (1998) conducted laboratory exposure tests on concrete specimens reinforced with uncoated reinforcement, epoxy-coated reinforcement, stainless steel reinforcement, copper-clad reinforcement, galvanized reinforcement, and spray metallic-clad reinforcement. The corrosion rates of epoxy-coated bars were less than those of uncoated bars. The authors concluded that Type 316 stainless steel reinforcement should be considered as a means for achieving a 75- to 100-year ser- vice life. Michigan DOT (MDOT) compared the deterioration trends of bridge decks containing epoxy- coated reinforcement, stainless steel reinforcement, and FRP reinforcement using the National Bridge Inventory condition rating scale (Valentine 2015). The condition rating scale ranges from 0 (failed) to 9 (excellent). The study yielded the following conclusions: • The service life of a bridge deck containing epoxy-coated reinforcement is estimated to be approximately 86 years. • The trend for bridge decks with stainless steel reinforcement is slightly better in the early stages than the trend for bridge decks with epoxy-coated reinforcement. • The trend for bridge decks with FRP reinforcement is not as good in the early years as the trend for bridge decks with epoxy-coated reinforcement. This is attributed to the lower modulus

56 of elasticity of the FRP reinforcement, which may be resulting in increased cracking of the bridge deck surface. The few FRP decks included in this study used different materials and design. aaSHTo lrfd SPecIfIcaTIonS for duraBIlITy Article 5.14 of the AASHTO LRFD Bridge Design Specifications (AASHTO 2017) addresses durabil- ity of concrete structures. The principal aim of the AASHTO Specifications, with regard to durability, is the prevention of corrosion of the reinforcing steel. The commentary states that design consider- ations for durability include concrete quality, protective coatings, minimum cover, distribution and size of reinforcement, details, and crack widths or prestressing. Reinforcement that is susceptible to corrosion and used in concrete exposed to deicing salts or saltwater shall be protected by the use of low-permeability concrete and concrete cover to the reinforcement. Specific information is provided for concrete cover but not for permeability. The commentary also states that the effects of salt intrusion and depassivation caused by carbon- ation can be mitigated by using corrosion inhibitors, coated reinforcement, bimetallic reinforcement, stainless steel reinforcement, or nonmetallic reinforcement, such as FRP composites. Article 5.14 does not mention controlling shrinkage and permissible crack widths, although crack widths are a design consideration in other articles. PermISSIBle crack WIdTHS Although the research described indicates that crack width may not be that significant when it comes to corrosion, recommendations for limiting crack widths have existed for many years and are still in use today. For example, Tadros et al. (2010) summarized permissible maximum crack widths developed between 1935 and 1970 and generally used to control flexural cracking in beams. Values ranged from 0.001 to 0.080 in., depending on the application and exposure conditions. Their analysis of the recommendations indicated that most flexural crack widths in beams, at a 40-ksi tensile stress in the reinforcement, ranged from 0.005 to 0.010 in. ACI Committee 224 (2008) provides a table of reasonable crack widths in reinforced concrete under service loads, as shown in Table 7. A footnote to the table states that a portion of the cracks in a structure will exceed these values, and with time, a significant portion can exceed these values. These crack widths are not always a reliable indication of the corrosion deterioration to be expected. A larger cover, which will lead to wider surface crack widths, may be preferable for corrosion control in certain environments (ACI Committee 224 2008). Article 5.6.7 (formerly 5.7.3.4) of the AASHTO LRFD Bridge Design Specifications, which addresses control of cracking by distribution of reinforcement, is indirectly based on crack widths of either 0.017 or 0.013 in., depending on the selected exposure condition. The commentary states that “Previous research indicates that there appears to be little or no correlation between crack width and corrosion.” Exposure Condition Crack Width (in.) Dry air or protective membrane 0.016 Humidity, moist air, soil 0.012 Deicing chemicals 0.007 Seawater, and seawater spray, wetting and drying 0.006 Water-retaining structures 0.004 Source: ACI Committee 224 (2008). TABLe 7 ReASONABLe CRACk WIDTHS

57 deTermInaTIon of Bar SPacIng To conTrol crack WIdTHS Over the years, different equations have been developed for the calculation of crack widths in concrete components (Modjeski and Masters et al. 2015). Most of these equations are based on an analysis of crack widths measured on laboratory test specimens loaded in flexure. The first equation relating to bar spacing and crack width to be used in U.S. bridge specifica- tions was the Gergely-Lutz equation (Gergely and Lutz 1968). It was based on a statistical analysis of experimental data. The original equation for predicting crack width, wc, was w f d Ac s s c0.076 (15)3= b where wc = maximum probable crack width at the tension face (in.); bs = ratio of flexural strain at the extreme tension face to the strain at the centroid of the reinforcement; fs = stress in steel reinforcement (ksi); dc = thickness of concrete cover measured from extreme tension fiber to center of the flexural reinforcement located closest thereto (in.); and A = average effective concrete area per bar of the flexural tension reinforcement (in.2). For a single layer of reinforcement of constant spacing, the term, A, simplifies to 2dcs, where s = bar spacing. h kd d kds (16)b = − − where h = overall thickness or depth of the beam (in.); k = distance from neutral axis to compression face divided by the effective depth of the beam (in.); and d = effective depth of the beam (in.). The AASHTO Standard Specifications (AASHTO 2002) and the subsequent LRFD Specifica- tions (AASHTO 1994) included the Gergely-Lutz equation in a slightly rearranged form. The crack width variable and the bs factor were consolidated into a single Z-factor, and the equation was written in terms of allowable stress. Using an approximate limiting crack width of 0.016 in. and an average bs factor of 1.2 resulted in f Z d A sa c (17)1 3( )= where fsa = allowable reinforcement stress (ksi); Z = factor = 170 for moderate exposure conditions = 130 for severe exposure conditions, with the remaining terms as defined previously. Based on physical phenomenon, Frosch (1999) developed the following equation to predict crack widths: w f E d scu s s s c2 2 (18)2 2( )( )= b +

58 where wcu = maximum crack width for uncoated reinforcement (in.); fs = stress in steel reinforcement (ksi); and Es = modulus of elasticity of reinforcing bars (ksi), with the remaining terms as defined previously. The equation can be rewritten to solve for maximum permitted reinforcement bar spacing as follows: s w E f d cu s s s c2 2 (19) 2 2( )= b     − where ds c1.0 0.08 . (20)b = + Frosch compared crack widths calculated by his equation with existing test data for reinforcement stress levels ranging from 20 to 50 ksi. Based on crack widths between 0.016 and 0.021 in., Frosch proposed the following simplified design equation: s d s c s s12 2 3 12 (21)= a − a     ≤ a where fs s c 36 (22)a = g as = reinforcement factor; and gc = reinforcement coating factor: 1.0 for uncoated reinforcement and 0.5 for epoxy-coated rein- forcement, unless test data can justify a higher value. Based on a review of past research, parametric studies, and various crack width predictive methods, DeStefano et al. (2003) proposed the following equation: f s d fss e r s c y 700 2 0.8 (23)( )= g g b + ≤ where fss = calculated tensile stress in steel reinforcement at the service limit state (ksi); ge = exposure factor; and gr = reinforcement factor = 0.75 for smooth welded wire reinforcement = 1.00 for all other types of reinforcement. d h ds c c 1 0.7 (24)( )b = + − The upper limit of 0.8fy on the allowable stress was proposed to provide a factor of safety against permanent yielding of the reinforcement under service loads. It is similar to that stipulated for steel flexural members in Article 6.10.4.4.2 of the AASHTO LRFD Specifications (AASHTO 2014). The paper (DeStefano et al. 2003) compares their equation with various design equations but does not compare the equation directly with test data.

59 A rearrangement of the equation 23 results in s f d e r s ss c 700 2 (25)= g gb − With gr = 1.0, this equation becomes equation 5.6.7-1 (formerly 5.7.3.4-1) in the 2017 AASHTO LRFD Specifications (AASHTO 2017): s f d e s ss c 700 2 5.6.7-1 (26)( )≤ gb − In certain situations involving higher-strength reinforcement or large concrete cover, the use of equation 5.6.7-1 can result in small or negative values for s. Therefore, the 2016 Interim Revisions (AASHTO 2016) introduced a limit that for calculation purposes dc need not be taken greater than 2 in. plus the bar radius and that s need not be less than 5 in. to control flexural cracking. equation 5.6.7-1 contains an exposure factor, ge, with suggested values of 1.0 or 0.75 depending on the exposure condition. This same factor could be used to address different types of corrosion-resistant reinforcement, or a new factor could be added to the equation similar to the gr in equation 25. The fac- tors could also vary depending on the desired service life of the structure. For this approach to be imple- mented, additional research and data are needed to determine the appropriate values for the factors. ServIce lIfe Service life for bridges is defined by AASHTO as the period of time the bridge is expected to be in operation (AASHTO 2017). At the present time, there are no U.S. standards or guidelines in place to establish performance criteria for service life design. However, this may change in the future as a result of an AASHTO and FHWA program to promote the use of service life design. The Second Strategic Highway Research Program produced a “Design Guide for Bridges for Service Life” (Azizinamini et al. 2014). The guide provides information, guidance, and procedures to systematically approach service life and durability for new and existing bridges. It is expected that the guide will be expanded, modified, and progressively embraced at different project and program levels by the bridge and structures community. Many factors can limit the service life of a bridge, but a major factor is corrosion of steel rein- forcement caused by deicing chemicals or saltwater. Prediction of service life based on corrosion of reinforcement generally is based on a two-part model: an initiation phase, during which chloride ions build up at the level of the reinforcement until a critical concentration is reached, and a propagation phase, during which the reinforcement corrodes (Bartholomew 2015). The duration of the initiation phase depends on the chloride concentration at the concrete surface, the rate at which the chloride ions penetrate the concrete, the distance that the chloride ions have to travel, and the critical chloride concentration for the initiation of corrosion. For most bridges, the only variable that is well defined is the distance that the chloride ions have to travel. Consequently, a deemed-to-satisfy approach is usually taken by specifying a minimum concrete cover and a maxi- mum permeability. All of this assumes that the concrete is not cracked. If service lives of 100 years are to be predicted with any degree of reliability, analytical models or design procedures that include the presence of cracks in the concrete are needed. Although there are controversial findings about the impact of crack width on corrosion rate, gen- eral agreement exists that cracking reduces the time to corrosion initiation (TRB 2006). The local- ized corrosion at the cracked areas leads to additional longitudinal surface cracking, delamination, and debonding, which ultimately result in a reduction in the strength and stiffness of the structure. Thus, it is desirable to control crack widths to an acceptable level even though there is no consensus on the size of allowable crack widths.

60 According to Nair et al. (2016a), early deterioration of concrete bridge decks has serious implica- tions both financially and with regard to public safety. A shortened service life of the bridge deck, higher maintenance costs, increased frequency of maintenance, and corrosion of reinforcing steel are some of the consequences of deck cracking. The dollar impact of corrosion on highway bridges is considerable. The average annual direct cost of corrosion for highway bridges was estimated to be $8.29 billion (Yunovich et al. 2005). Thus, the corrosion of the reinforcing steel that is attributed primarily to bridge deck transverse cracking and the application of deicing chemicals containing chloride is costly (McLeod et al. 2009). concluSIonS aBouT THe Influence of crackIng on long-Term BrIdge Performance Research has established that the presence of cracks in concrete allows chloride ions to reach the reinforcement in less time than in uncracked concrete. Consequently, corrosion of noncorrosion- resistant reinforcement will begin earlier in cracked concrete than in uncracked concrete. However, most analytical models assume uncracked concrete. The use of corrosion-resistant reinforcement prolongs the time to initiation of corrosion and thus increases service life. Stainless steel provides the longest life, although coated bars, such as epoxy- coated ones, can provide sufficient protection for many applications. Additional work is needed to be able to predict service life with cracked concrete and the different types of reinforcement available. Currently, no U.S. standards exist to predict long-term performance and service life.