Concrete Bridge Deck Performance (2004)

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.

25 CHAPTER FIVE STRUCTURAL DESIGN PRACTICES, CONSTRUCTION PRACTICES, SPECIFICATIONS, AND COSTS This chapter is concerned with aspects of structural design practices, construction practices, specifications, and costs that are related to concrete bridge deck performance. STRUCTURAL DESIGN PRACTICES General Responses to the questionnaire for this synthesis indicated that most agencies use a minimum deck thickness in the range of 190 to 230 mm (7.5 to 9 in.) Reinforcement bar sizes are typically 16 and 19 mm in diameter (No. 5 and No. 6), with a bar spacing not exceeding 305 mm (12 in.). Seventeen of the 38 U.S. respondents reported HS 20 as the design live load, 17 reported HS 25, and 14 reported HL 93. Twelve respondents reported using more than one design load. Cover to Reinforcement In 1970, the general recommendation for concrete cover was a minimum clear cover of 50 mm (2 in.) over the top- most steel (NCHRP Synthesis of Highway Practice 4 1970). NCHRP Synthesis of Highway Practice 57, pub- lished in 1979, reported that the specified concrete cover, until recently, was typically 38 mm (1.5 in.). Currently, the AASHTO Standard Specifications for Highway Bridges (2002) requires a minimum cover of 65 mm (2.5 in.) for top reinforcement in concrete deck slabs that have no posi- tive corrosion protection and are frequently exposed to deicing salts. Positive corrosion protection methods may include ep- oxy-coated reinforcement, special concrete overlays, and im- pervious membranes, or a combination of these methods. Reference is made to NCHRP Report 297 for additional in- formation (Babaie and Hawkins 1987). The AASHTO LRFD Bridge Design Specifications (2004) require a minimum cover of 65 mm (2.5 in.) for concrete that is ex- posed to deicing salts or on deck surfaces that are subject to stud or chain wear. The cover may be decreased to 40 mm (1.5 in.) when epoxy-coated reinforcement is used. In the survey for the Michigan DOT, the typical con- crete cover to the top layer of reinforcement was reported to range from 51 to 76 mm (2 to 3 in.), with 64 mm (2.5 in.) being the most common value (Aktan and Fu 2003). Most states also reported that the present require- ment for cover was larger than it had been in the past, with 38 mm (1.5 in.) being the most common previous value. Responses to the questionnaire for this synthesis indicated that 39 or 87% of the 45 responding agencies specified a minimum clear cover of 50 to 64 mm (2.0 to 2.5 in.) for the top layer of reinforcement and 42 or 93% specified 25 to 38 mm (1.0 to 1.5 in.) for the bottom layer. Several studies have identified that the depth of cover over the top reinforcing steel is the most significant factor contributing to the durability of the deck (Stark 1970; Crumpton and Bukovatz 1974; Clear 1976). In a Kansas DOT study, it was estimated that increasing the cover from 50 to 75 mm (2 to 3 in.) and decreasing the water–cement ratio of the concrete from 0.44 to 0.35 would triple the life of a deck (McCollum 1976). In NCHRP Report 57 (1979), it was pointed out that, if the cover distance had a standard deviation of 10 mm (0.375 in.), the specified cover must be approximately 67 mm (2.625 in.) for a minimum cover of 50 mm (2 in.) with a 95% compliance (Weed 1974; Van Daveer 1975). The authors of NCHRP Report 297 concluded that the effective service period for a bridge deck with 90-mm (3.5- in.) cover to the reinforcement may be 50 years or more when salt exposure is less than 3 Mg per lane-kilometer per year (5 tons per lane-mile per year) (Babaie and Hawkins 1987). For higher salt applications, the water–cement ratio of the concrete determines the service life. CONSTRUCTION PRACTICES Stay-in-Place Concrete Panels Precast concrete, stay-in-place deck panels are used exten- sively in several parts of North America to support the CIP concrete deck. After the concrete is placed, the panel be- comes an integral part of the composite deck to resist both transverse and longitudinal bending. Because the panels are not continuous for the complete length of the bridge or across the supporting beams, there is a tendency for cracks to occur in the CIP concrete above the discontinuities in the panels. This is often called reflective cracking. Concrete Temperature NCHRP Synthesis of Highway Practice 4 reported that concrete mix temperatures of 27°C to 32°C (80°F to 90°F)

26 were believed to play a major role in crack development, high water requirement, and strength loss (1970). Krauss and Rogalla (1996) reported that reducing placement and peak concrete temperatures relative to ambient tempera- tures can reduce deck cracking. They recommended that concrete temperature at time of casting be 5°C (10°F) cooler than ambient, except when temperatures are below 16°C (60°F), when the concrete temperature should be the same as ambient. Responses to the questionnaire for this synthesis indi- cated that the majority of the agencies specified a maxi- mum concrete temperature at time of placement of 32°C (90°F). However, very few respondents specified a maxi- mum temperature for the deck concrete during the curing period. Placement Procedures NCHRP Synthesis of Highway Practice 4 reported that ex- cessive surface manipulation lowers the surface scaling re- sistance especially if the manipulation occurs during the bleeding period (1970). The addition of water to the sur- face to facilitate finishing led to decreased scaling resis- tance (Malisch et al. 1966). Other experiments showed that concrete surfaces struck off immediately after casting with no further finishing operations showed greater resistance to surface scaling compared with surfaces given a second and final finish (Klieger 1955). Schmitt and Darwin (1995) in an investigation of 40 bridge decks in northeast Kansas could not identify any re- lationship between cracking and placement length for monolithic bridge decks. However, cracking clearly in- creased as placement length increased for bridge deck overlays. Krause and Rogalla (1996) reported that place- ment sequence is important, but that sequence is not a pri- mary cause of transverse deck cracking. In NCHRP Synthesis of Highway Practice 57 (1979), it was reported that insufficient bridge deck slope makes construction without localized depressions difficult. This accelerates the ingress of chlorides and promotes scaling as water containing deicing salts collects in these areas. Dete- rioration in gutter areas is common on flat bridges. Curing Practices The AASHTO Standard Specifications for Highway Bridges (2002) and the AASHTO LRFD Construction Specifications (1998) require that all newly placed concrete be cured for 7 days, except that the curing period shall be 10 days when pozzolans in excess of 10% by mass of the portland cement are used. The alternative curing methods that may be used are the water method, the liquid mem- brane curing compound method, and the waterproof cover method. For bridge decks, the specifications require that a combination of the liquid membrane curing compound method and the water method be used. The curing com- pound shall be applied immediately after the finishing op- erations on each portion of the deck are complete. The wa- ter cure shall be applied not later than 4 h after completion of deck finishing. For portions of the deck on which finish- ing is completed after normal working hours the water cure shall be applied not later than the following morning. Responses to the Michigan DOT survey indicated that 90% of the 31 responding states have a continuous wet cure with a duration of 5 to 14 days (Aktan and Fu 2003). Nineteen of the states allow a burlap cover and 16 allow the use of a curing compound. No states reported using air curing. When asked about the probable causes of early age deck cracking, most states responded “substandard curing.” In the questionnaire for this synthesis, each agency was asked to identify the type of curing that they specify. The responses are summarized in Figure 5. Forty of the 45 re- sponding agencies specify a water-saturated cover, al- though 27 or 60% of the respondents specify more than one method. Thirty-seven or 82% of the respondents re- ported that they specify that curing must begin immedi- ately after finishing any portion of the deck. Thirty-two or 71% of the respondents specify a 7-day curing period. The advantages of using a longer curing period include a lower permeability, increased hydration of the cement so that less free water is available to produce shrinkage, and higher tensile strength when the concrete begins to shrink. All of these factors contribute to a more durable bridge deck. The disadvantage of a longer curing period is that it extends the construction time. However, extending the cur- ing period on most projects represents only a minor exten- sion of the total schedule. Improper curing is thought to significantly contribute to cracking (Durability of Concrete Bridge Decks 1970; Poppe 1981; Kochanski et al. 1990). According to Krauss and Rogalla (1996), the most significant construction- related factors affecting transverse deck cracking involved weather and curing. They reported that transportation agencies observed more cracking when concrete was placed during lower humidities and higher evaporation rates. They recommended immediate water fogging or ap- plication of evaporation-retarding films regardless of evaporation rates or temperature. Early wet curing was rec- ommended to reduce evaporation of mix water and to cool the concrete. With HPCs, application of water curing immediately af- ter concrete finishing, as illustrated in Figure 6, is ex-

27 FIGURE 5 Survey results of bridge deck curing methods. FIGURE 6 Application of wet burlap immediately after concrete curing (Schell and Konecny 2001). (Courtesy: HPC Bridge Views published by FHWA and NCBC.) tremely important because these concretes have less bleed water and the likelihood of plastic shrinkage cracking is greater (Khaleghi and Weigel 2001; Praul 2001; Schell and Konecny 2001). Whiting and Detwiler (1998) emphasized the importance of curing silica fume concrete. The lack of bleeding means that water lost from the surface as a result of evaporation cannot be readily replaced. Consequently, Whiting and Detwiler recommended the following precau- tions: • Strict adherence to specifications regarding evapora- tion rates and cessation of concrete placement if rela- tive humidities are low and temperatures and wind speeds are high; 0 10 20 30 40 50 60 70 80 90 100 Water Ponding Water- Saturated Cover Fog Spray Waterproof Cover Liquid Membrane None Type of Curing P er ce nt ag e of 4 5 R es po nd en ts No. of respondents is shown for each option 5 40 18 10 16

28 FIGURE 7 Nozzles attached to the finishing equipment (Schell and Konecny 2001). (Courtesy: HPC Bridge Views published by FHWA and NCBC.) • Expeditious finishing of concrete and use of fog sprays during finishing; • Use of evaporation-retarding agents during and im- mediately after finishing; and • Initiation of wet curing as soon as possible after fin- ishing. Use of these techniques has, in general, reduced the in- cidence of plastic shrinkage cracking and allowed for the successful placement of many hundreds of silica fume con- crete overlays (Whiting and Detwiler 1998). In their study of silica fume bridge deck overlays, Miller and Darwin (2000) concluded that improved curing reduced cracking. The use of fogging equipment to reduce evaporation rates is shown in Figure 7. Traffic-Induced Vibrations For replacement of existing bridge decks, it is frequently necessary to undertake the construction in several phases so that part of the bridge can remain open to traffic. As a result, the fresh concrete in the new bridge deck may be exposed to vibrations from traffic on an adjacent structure. Based on laboratory tests using simulated traffic- induced vibrations, Harsh and Darwin (1986) concluded that traffic-induced vibrations have no detrimental effect on either bond strength or compressive strength of concrete in bridge deck repairs, if high-quality, low-slump concrete is used. As slump increased, the vibrations resulted in lower bond and compressive strengths. Slumps in the range of 100 to 130 mm (4 to 5 in.) could be detrimental and slumps of 175 to 200 mm (7 to 8 in.) were found to de- crease the bond and compressive strengths by 5% to 10%. In their study on transverse cracking in newly con- structed bridge decks, Krauss and Rogalla (1996) reported that other research showed that traffic-induced vibrations before or after concrete hardening do not cause cracking. They reported that deflections associated with the vibra- tions are too small to damage the concrete. More informa- tion on traffic-induced vibrations is available in NCHRP Synthesis of Highway Practice 86 (Manning 1981). Maintenance In the survey for this synthesis, 15 or 33% of the 45 re- spondents indicated that they repair cracks in bridge decks, 9 or 20% indicated that they did not repair cracks, and 17 or 38% indicated that they repair cracks sometimes. “Sometimes” depended on the severity of the cracking. The more frequently listed crack repair methods were epoxy in- jection and the use of methacrylates or other sealants. Of these, epoxy injection and methacrylates were identified as the most effective in prolonging bridge deck life.

29 The survey respondents were asked to identify what method they use to repair freeze-thaw damage. Most re- sponded that they removed the damaged concrete and re- paired with a deck patching material or overlay. Overlays were identified as the most effective surface repair method in prolonging bridge deck life. SPECIFICATIONS Prescriptive Versus Performance Specifications The traditional approach to achieving a durable concrete bridge deck has been a prescriptive one, where certain pa- rameters of the concrete mix proportions are specified. These typically include a maximum w/cm, a minimum cementitious materials content, and a percentage of supplementary cemen- titious materials. For bridge decks exposed to freezing and thawing cycles, a range of air contents is specified. The pa- rameters are selected in anticipation that they will result in a concrete with a low permeability and high freeze-thaw re- sistance. In some instances, testing is performed to verify that the desired properties will be achieved. With the FHWA initiative to implement the use of HPC in bridges, at least 16 states moved in the direction of per- formance-based specifications (High-Performance Con- crete 2003). Subsequently, many other states have imple- mented HPC (Triandafilou 2004). In this approach, the end performance characteristic is specified. The range of char- acteristics includes freeze-thaw resistance, deicer scaling resistance, chloride permeability, abrasion resistance, al- kali-aggregate reactivity, and sulfate resistance. It is then the contractor’s responsibility to conduct the necessary tests to prove that the proposed concrete mix proportions will satisfy the specified performance characteristics. This approach is similar to that used for concrete compressive strength. However, its application for durability character- istics presents new challenges that the industry may not be ready to handle at this time. For early HPC bridge projects, the Texas DOT (TxDOT) did not specify how the contractor was to obtain durable concrete other than requiring adherence to the specifica- tions for the project. Contractors were alerted that the bridges were part of a research program and that concrete mix designs would be developed by TxDOT and the re- search team to meet strength and durability guidelines. A by-product of the research was an HPC specification for use on future projects. The specifications required that mix designs be formulated and verified to meet strength and durability requirements. “After several projects, it became apparent that the contractors, the concrete suppliers, and TxDOT lacked the experience necessary to efficiently de- sign concrete that would meet performance-based specifi- cation requirements for durability” (Cox and Pruski 2003). To gain experience and a better understanding of the role that concrete constituent materials have on permeabil- ity, TxDOT began and continues to use prescriptive speci- fications that require the use of supplementary cementi- tious materials at a prescribed rate. The contracting community has expressed minimal opposition to this approach even though some projects require the use of supplementary cementitious materials where they have not been used before. TxDOT is aware of concerns about prescriptively specifying the use of supplementary cementitious materials when the materials supplier and contractor are not experienced with the materials. To ad- dress these concerns, TxDOT requires contractors to de- velop strength versus time curves for the concrete during the mix design process. For verification of durability pa- rameters, additional concrete test specimens are supplied to the central laboratory for durability tests (Cox and Pruski 2003). A further example of the reluctance is provided in the FHWA HPC demonstration bridges. Of the four character- istics for durability—freeze-thaw resistance, scaling resis- tance, abrasion resistance, and chloride penetration—only chloride penetration was consistently specified (Russell et al. 2003). This reluctance may be the result of a lack of familiarity with the test method, a lack of in-house capability to perform the tests, impact of costs when ad- ditional performance requirements are specified, or in- creased time to perform the tests. Whereas performance- based specifications for durability seem to be highly de- sirable, a lot more experience is needed before they can be fully implemented. Warranties In the survey conducted for this synthesis, the Ohio DOT was the only U.S. transportation agency that reported the use of warranties as part of their specifications. In 1999, Ohio introduced a specification requiring contractors to warrant new bridge decks constructed with HPC (Schultz 2002). The contractor is required to warrant against alliga- tor and map cracking for 1 year and against scaling and spalling for 7 years. The deck is evaluated for alligator and map cracking at 1 year. Scaling and spalling are evaluated at 2 years and 1 month before the end of the warranty pe- riod. If any of the defects becomes evident during the war- ranty period, the contractor is required to make repairs at no cost to the state. Alligator and map cracks over 20% or less of the deck area are required to be sealed. If deck scal- ing occurs on 20% or less of the deck area and the depth is greater than 3 mm (1/8 in.), but not greater than 6 mm (1/4 in.), the defective areas are to be ground out. If the scaling is greater than 6 mm (1/4 in.) deep, the scaled area must be removed to a depth of 25 mm (1 in.) and replaced. If the area of map cracking or scaling is greater than 20%, the

30 top 25 mm (1 in.) of the whole deck must be removed and replaced with an overlay. Schultz (2002) reported that 6 of the 16 decks that received the 1-year review required cor- rective work for alligator or map cracking. The contractor is required to provide the Ohio DOT with a maintenance bond for the bridge deck for a period of 7 years. The amount of the bond is 50% of the total price bid for the HPC (Schultz 2002). Although unit prices for the HPC increased during the first year of the program, the prices in the second year were the same as those before the program was introduced. COSTS Methods for Predicting Life-Cycle Costs A key element in the prediction of life-cycle costs is ade- quately estimating the service life of the bridge deck. The AASHTO LRFD Bridge Design Specifications (2004) de- fines service life as the period of time that the bridge is ex- pected to be in operation. The end of the service life occurs when the bridge becomes functionally obsolete or accumu- lated damage in the bridge exceeds acceptable performance limits. However, service life is typically extended by per- forming periodic repairs to restore the serviceability of the structure. Responses to the Michigan DOT survey in 2002 indicated that most respondents believed that their rein- forced concrete bridge decks will last 30 to 40 years (Ak- tan and Fu 2003). Bhidé (2002) identified some of the service life predic- tion models available in 2002 as follows: • Life-365—Computer software developed by M.D.A. Thomas and E.C. Bentz that addresses time-dependent diffusion of chlorides and predicts service life and life- cycle costs for various protection strategies. • CIKS—Computer-Integrated Knowledge System de- veloped by D. Bentz that predicts chloride ion diffu- sivity coefficients and time to initiation of corrosion. • Duramodel—Model developed by W.R. Grace that uses effective diffusion coefficients to account for mechanisms other than pure diffusion. • ConFlux—Personal computer-based Multimechanistic Chloride Transport Model developed by A. Boddy, E.C. Bentz, M.D.A. Thomas, and R.D. Hooton that ac- counts for diffusion, permeability, chloride binding, and wicking. • ClinConc—Chloride penetration model developed by L. Tang, based on mass balance and genuine flux equations to predict chloride profiles in submerged parts of structures. • HETEK Model—Ten-step spreadsheet calculation for service life developed by AEC Laboratory, Den- mark, and applicable to marine structures and salt water splash zones. It should be noted that all of these programs are based on uncracked concrete and do not include the effects of cracking on service life predictions. Additional information about prediction of service life is being developed in NCHRP Project 18-06A, Service Life of Corrosion- Damaged Reinforced Concrete Bridge Elements. The ob- jective of the project is to develop a manual that provides step-by-step procedures for (1) assessing the condition of reinforced concrete bridge superstructure elements sub- jected to corrosion-induced deterioration, (2) predicting the remaining service life of such elements, and (3) quantify- ing the service-life extension expected from alternative maintenance and repair options. Service-Life Costs Babaie and Hawkins (1987) compared lifetime costs for several different bridge deck protection strategies, includ- ing increased concrete cover from 38 to 89 mm (1.5 to 3.5 in.), epoxy-coated top layer of reinforcement, special con- crete overlays, and interlayer membranes. They also in- cluded three double protection strategies of epoxy-coated top and bottom layers of reinforcement, epoxy-coated top layer of reinforcement with special concrete overlay, and epoxy-coated top layer of reinforcement with interlayer membrane. An annual interest rate of 10% and an annual inflation rate of 5% were assumed in the calculation for 50- year lifetime costs. For the singly protected decks, the least expensive strat- egy was the provision of a concrete cover of at least 89 mm (3.5 in.) over the uppermost bar. The other strategies in or- der of increasing costs were epoxy-coated top layer, inter- layer membrane with asphaltic concrete, and a low- permeability concrete overlay of either low-slump dense concrete or latex-modified concrete. The least expensive of the double protection strategies was epoxy-coated top and bottom layers of reinforcement. In 1999, the National Institute of Standards and Tech- nology published software to help bridge designers deter- mine the cost-effectiveness of new alternative construction materials based on a life-cycle costing methodology (Ehlen 1999). A sample analysis compares a bridge with conven- tional strength precast, prestressed concrete girders and a normal permeability concrete deck to one that has high- strength precast, prestressed concrete girders and a low- permeability concrete deck. The use of high-strength and HPCs was the more cost-effective solution. Kepler et al. (2000) compared the present value of costs for 33 corrosion protection methods assuming discount

31 rates of 2%, 4%, and 6%. The total present value was cal- culated by adding the initial cost to the present values of costs for repair and replacement, maintenance, and opera- tion. A 75-year service life was selected as the basis of comparison. A 230-mm (9-in.)-thick reinforced concrete bridge deck with 50 mm (2 in.) cover over the top layer of reinforcement was generally used. However, a 205-mm (8-in.)-thick deck was used for bridges with epoxy-coated reinforcement as the only corrosion protection method. Based on their analysis, the system with the lowest present value consisted of stainless steel clad reinforcement. The cost did not change with discount rate because it was as- sumed that repairs or maintenance would not be necessary. At the 2% discount rate, solid stainless steel reinforcement was a cost-effective option. At the 4% rate, hot rubberized asphalt membranes and calcium nitrite as a corrosion inhibitor were cost-effective. At the 6% rate, calcium nitrite was cost-effective. CONCLUSIONS ABOUT STRUCTURAL DESIGN PRACTICES, CONSTRUCTION PRACTICES, SPECIFICATIONS, AND COSTS The most important structural design practice to reduce corrosion of reinforcement in uncracked concrete bridge decks is to provide a minimum cover to the top layer of re- inforcement of 64 mm (2.5 in.). The most important con- struction practices to achieve a low-permeability, un- cracked bridge deck with adequate freeze-thaw resistance is to initiate wet curing of the concrete immediately after finishing any portion of the concrete surface and maintain- ing wet curing for a minimum of 7 days. Other practices that are beneficial include moderate concrete temperatures at time of placement, minimum finishing operations con- sistent with achieving the desired concrete surface, gradual development of performance specifications, and warran- ties.