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.

61 This chapter contains case examples describing how four states have looked for solutions for pre- venting or reducing cracking in concrete bridge decks. California Department of transportation The following case example is extracted from an article by Maggenti et al. (2013) and is published with permission of the American Concrete Institute. Between 2001 and 2003, six new bridges with CIP concrete decks on spliced precast concrete bulb-tee girders were constructed on I-80 near Truckee, California. The concrete had a w/cm of 0.36, cementitious materials content of 752 lb/yd3 with 25% fly ash, and 6% air content. After the first few decks were constructed, multiple transverse cracks at about 2-ft centers were visible. Beginning in late summer of 2002, deck mixtures were modified to include SRAs. This caused a dramatic reduction in cracking. The following construction season, the remaining structures were constructed using an SRA in the deck concrete. The decks constructed using SRAs remained free of visible cracking for several years. Following this simple, yet effective, adjustment to the mixture design, specifications were written for the deck concrete for the Angeles Crest Bridge on SR-2 in Los Angeles County, in the mountains northeast of Los Angeles, California. This is a 208-ft, single-span bridge with the deck cast on six 8-ft-deep spliced precast, prestressed concrete bulb-tee girders spaced at 6.5-ft centers. The specifi- cations called for a 28-day concrete compressive strength of 5.0 ksi and a 6% air content. Construc- tion was completed in 2008 using a deck mixture with 767 lb/yd3 of cementitious materials and an SRA. Upon later inspection, only a few hairline shrinkage cracks were observed at the westerly end. The same result was achieved on the 2007 emergency replacement of the fire-destroyed bridge spans at the MacArthur Maze in Oakland, California. This replacement was completed in a mere 26 days using a deck mixture made up of SRAs, 800 lb/yd3 of cementitious material, water reducers, and a Type C accelerating admixture. The deck was cast on steel girders with headed studs to achieve composite action. By 2013, no cracking had been reported on this replacement deck span, whereas transverse cracks have been noted every few feet on all adjacent deck spans, which used a cementi- tious materials content of 564 lb/yd3 and a 1.5-in. maximum nominal aggregate size. These projects demonstrated that SRAs could eliminate the need to specify low-strength concrete, long curing times, a low w/cm ratio, or large aggregates. In 2011, a 5-in.-thick concrete deck was placed on precast concrete box beams over Craig Creek on SR-99 near Red Bluff, California. The concrete mixture was designed to develop a 3-day strength of 4.0 ksi, using 705 lb/yd3 of portland cement, a water-cement ratio of 0.39, an SRA, and synthetic macrofibers at 3.0 lb/yd3. The project was used to study accelerated bridge construction through HPC and only 3 days for moist curing. No visible cracking was noted during inspection after 14 months of service. It was concluded that a high-quality, durable deck can be successfully and rapidly constructed. The Caltrans Standard Specifications now require limiting the shrinkage of deck concrete to a maximum of 0.032% (320 millionths) after 28 days of drying, a minimum amount of an SRA, a dose chapter seven Case examples

62 of polymer fibers, and stricter curing requirements. In 2013, a Caltrans fact sheet estimated that Caltrans was spending $50 million annually on sealing cracks in concrete bridge decks. The increase in concrete costs after implementing the new requirements is estimated to be about $2 million annually. In 2012, the calculated increase in costs was $0.89/ft2. Kansas Department of transportation In 2003, Kansas DOT (KDOT) became the lead state of a pooled fund study with 18 other states and the FHWA. The goal of the study was to construct low-cracking, high-performance concrete (LC-HPC) bridge decks (Browning and Darwin 2007; Browning et al. 2009). By 2012, 16 bridges in northeast Kansas and seven in other participating states had been constructed using some or all of the new specifications (Darwin et al. 2012). Most LC-HPC bridge decks had a companion non–LC-HPC bridge deck so comparisons could be made. Evaluation of more than 150 bridge decks, most supported by steel beams, has demonstrated that crack density increases with time. The bridge decks that performed well in the first 3 years performed well at later ages. Monolithic decks tended to perform better than did those constructed with over- lays because cracks in the subdeck reflected through the overlays, and shrinkage of the overlay was restrained by the subdeck (Darwin 2014). Other conclusions from the research indicated the following: • Concrete mixes with greater paste content exhibited greater drying shrinkage. • The restraint to drying shrinkage was greater for steel beam bridges than for bridges with precast, prestressed concrete beams. • Increased concrete slump leads to increased settlement cracking. • Increased air content reduces cracking. • Increased concrete strength results in more cracking. • Higher creep is beneficial in reducing cracking. • Extra finishing leads to increased plastic and shrinkage cracking. • Rapid evaporation of bleed water increased plastic shrinkage cracking. • Large temperature differences between the deck concrete and the beams at time of placement result in thermal cracking. • The use of stay-in-place forms doubles the moisture gradient. Based on the previous information, specifications for LC-HPC for bridge decks were developed, including the following requirements: • Cementitious materials content of 500 to 540 lb/yd.3 (This was later changed to a maximum of 540 lb/yd3.) • Use of only Type I/II cement. • Maximum w/cm ratio of 0.42. (This was later revised to 0.44 to 0.45.) • Paste content (total volume of water and cement) less than 25%. • Maximum aggregate size of 1 in. • Combined aggregate gradation optimized for uniform size distribution. • Maximum aggregate absorption of 0.7%. • Designated air content of 7.0% to 9.0% (with 6.5% to 9.5% acceptable). • Compressive strength at 28 days of 3.5 to 5.5 ksi. • Designated slump of 1.5 to 3.0 in. (with a maximum of 3.5 in. acceptable). • Concrete placement temperature of 55°F to 70°F. • Evaporation rate less than 0.2 lb/ft2/h. • Wet curing with one layer of presoaked burlap starting within 10 minutes of concrete strike off followed by a second layer of burlap within 5 minutes, as shown in Figure 17. • Fourteen days of wet curing followed by application of a curing compound to be unmarred for 7 days. • A qualification slab with dimensions equal to the bridge width, full slab depth, and 33 ft long to be cast before bridge deck placement to demonstrate the contractor’s capabilities.

63 Curing during cold weather requires extra precautions to reduce cracking caused by thermal stresses. To prevent cracking, the difference between the concrete and steel girder temperatures must be kept within a tolerable range. During the first 72 hours of the curing period, if the ambient air temperature falls below 40°F, protective measures must be taken to maintain the temperature of the concrete and the girders between 55°F and 75°F. This may include using straw or extra burlap on the concrete, and wrapping the girders with plastic and using propane heaters. Similar protec- tive measures must be taken when the ambient air temperature is expected to drop more than 25°F below the placement temperature of the concrete during the first 24 hours after placement. After the first 72 hours, the contractor has the option of maintaining the temperature throughout the cur- ing period or extending the curing period to account for periods during which the air temperature drops below 40°F. Two of the primary lessons learned were that these concrete specifications can be implemented at a reasonable cost and that the low-paste concrete mix is workable, placeable, and finishable in the field (Browning et al. 2009). Concrete strengths were about 4.0 ksi. The establishment of a good working relationship among owners, inspectors, contractors, and concrete suppliers was of prime importance for the successful construction of an LC-HPC bridge deck. All participants must clearly communicate their expectations and successfully meet the specifications (Browning et al. 2009). Bridge decks constructed with LC-HPC had less than 10% of the cracking found in traditional bridge decks. Sixteen LC-HPC bridge decks and 13 control decks were inspected in 2014 and 2015 as part of the continuing program to monitor cracking in these bridge decks (Alhmood et al. 2015). Based on the results of these inspections, the following conclusions were developed: • LC-HPC bridge decks exhibit less cracking than do the matching control decks in most cases. • Only one bridge deck had a higher overall crack density than its control deck. • The most common crack type is transverse cracking. Cracks of this type appear to run parallel to the top layer of the deck’s reinforcement. • Near the abutments, cracks usually propagate perpendicular to the abutments. As part of the continuing LC-HPC program, the effects of SRAs, SCMs, prewetted LWA, slag cement, silica fume, and air entraining admixtures on concrete properties were evaluated in laboratory mixes (Reynolds et al. 2009; Pendergrass and Darwin 2014; Yuan et al. 2015). Shrinkage-reducing admixtures reduced early-age and long-term shrinkage, with the reductions in shrinkage concentrated FIGURE 17 Application of precut, rolled, wet burlap within 10 min- utes after concrete finishing (Courtesy: David Darwin, University of Kansas and Concrete Bridge Views, published jointly by FHWA and the National Concrete Bridge Council).

64 within the first 90 days. Prewetted LWA reduced early-age and long-term shrinkage (Browning et al. 2011). Shrinkage was also reduced when slag cement was used at 30% by volume of the cementitious materials and as silica fume was used at a nominal 3% by volume in conjunction with LWA and slag cement. pennsylvania Department of transportation In 1995, Purvis et al. (1995) reported that increased cracking had been observed in newly constructed concrete highway bridge decks in Pennsylvania. Consequently, PennDOT initiated a research project to identify factors that may cause premature concrete cracking, relate these factors to PennDOT bridge deck cracking, and recommend changes. Based on field observations and laboratory tests, the authors concluded that the main cause of transverse cracking was a combination of thermal shortening and drying shrinkage. The following recommendations were made: • Control the temperature difference between the deck and the girders to less than 22°F for at least 24 hours after the concrete is placed. • Require that deck concrete have a shrinkage not greater than 700 millionths after 4 months of drying. To comply with the temperature requirement in hot weather, the authors recommended the use of a retarder and to cover the deck concrete with wet burlap within 30 minutes after placing to minimize heat buildup from the sun. It was also preferable to place concrete at night. To comply with the temperature difference in cold weather, the authors recommended that the air underneath the deck be heated to 55°F to 75°F to raise the temperature of the beams and reduce the temperature difference with the deck concrete. At the same time, the top surface of the deck is to be insulated to maintain the same temperature range. To comply with the shrinkage requirement, the authors suggested the use of hard aggregates, such as quartz, dolomite, and limestone; maximum coarse aggregate absorption of 0.5%; maximum fine aggregate absorption of 1.5%; a lower water content in combination with a water-reducing admix- ture; a lower cement content; and a Type II cement. At that time, the specifications allowed a water content of 323 lb/yd3 and a maximum cement content of 752 lb/yd3. The Kernville viaducts in Pennsylvania consist of two 2,700-ft, 27-span, curved, continuous steel girder bridges with an 8-in.-thick reinforced concrete deck. The westbound lanes were built in 2001. A crack survey of the decks after construction of the westbound lanes showed that there were 237 cracks at an average spacing of 6.4 ft in the positive moment regions and 227 cracks at an average spacing of 5.1 ft in the negative moment regions (Spangler and Tikalsky 2006). Cracks had a width generally greater than 0.04 in., with numerous cracks being larger than 0.12 in. Below placement of the concrete deck for the eastbound lanes, the following changes to the con- crete mixtures were developed: • Increase the w/cm ratio from 0.40 to 0.43. • Decrease the cementitious materials content from 650 to 588 lb/yd3. • Decrease the percentage of ground granulated blast-furnace slag from 50% to 42% of the total cementitious materials content. • Decrease the target slump from 6.0 to 4.5 in. • Reduce the maximum concrete temperature at time of placement from 80°F to 70°F. The following construction changes were also implemented: • Place positive moment regions successively on one day, followed by placement of negative moment regions 3 or more days later. • Apply moist curing immediately after concrete finishing and maintain for 10 days. • Apply a pigmented curing compound at the end of the moist curing period. • Increase vigilance in quality control and quality assurance operations.

65 After construction of the eastbound lanes, crack surveys made immediately after curing and in subsequent months showed a dramatic reduction in the frequency and width of cracks. Early-age cracking immediately after curing was nearly eliminated by the changes in mixture design and con- struction practices. At 7 months, 174 total cracks were recorded on the eastbound deck, compared with 464 cracks on the westbound deck. Most of the cracks in the eastbound lanes were narrower than 0.005 in. during cold weather conditions. In another research project, Manafpour et al. (2016) reported data from field investigations of 203 bridge decks in Pennsylvania with ages to 50 years. The data were used to identify factors that contribute to early-age cracking and assess the effect of cracks on the long-term durability of bridge decks. The following main conclusions were drawn: • Higher concrete compressive strengths correlated with higher crack densities in the deck. The authors advised a maximum limit on the concrete compressive strength at 7 or 28 days (e.g., 4.0 or 5.0 ksi, respectively). • Lower total cementitious materials content and higher SCMs content resulted in less cracking. The authors advised a maximum limit on the cementitious materials content (e.g., 620 lb/yd3) and encouraged the use of SCMs to reduce heat of hydration, increase electrical resistivity, and prevent alkali–silica reaction. It was acknowledged that high SCM contents could pose poten- tial challenges with strength gain in cooler seasons. • Decks constructed with half-width procedures to allow one half of the bridge to remain open cracked four times more than did decks constructed with the full width at one time and the use of traffic detours. Washington state Department of transportation In a survey for NCHRP Synthesis 441 (Russell 2013), Washington State DOT (WSDOT) reported that the performance of HPC in CIP bridge decks was worse than that of conventional concrete. This was based on an observed increase in the amount of cracking, which appeared to be associated with the required use of fly ash. However, the cracking may have been caused by the use of girders with wide top flanges. These girders provide more restraint to the differential shrinkage between the deck concrete and the girders. WSDOT reported that no strategy employed was effective in minimizing cracking. The use of evaporation retardants was found to be the least effective. At that time, the WSDOT Standard Specifications for bridge deck concrete required a minimum cementitious materials content of 660 lb/yd3 with a fly ash content between 10% and 20% or a slag cement content between 10% and 30% of the total cementitious material. If both fly ash and slag cement were included, the maximum allowable content increased to 40%. The use of a water- reducing admixture and a retarding admixture was required. The use of a high-range, water-reducing admixture was permitted (Russell 2013). In the 2010 Standard Specifications, WSDOT increased the nominal maximum aggregate size from ¾ in. to 1 in. In 2011, WSDOT introduced a performance-based specification for bridge decks to eliminate or reduce early-age restraint cracking in bridge decks (Ferluga and Glassford 2015). Bridge decks constructed with this revised concrete specification are referred to as “Performance Based Bridge Decks.” The revised specification no longer had a minimum cementitious materials content and did not require the use of fly ash. Another significant change was to increase the nominal maximum aggre- gate size from 1 to 1½ in. The performance-based requirement for minimum concrete compressive strength remained in the specification at 4.0 ksi at 28 days. Performance limits on permeability, shrinkage, and surface scaling, with an optional requirement for freeze–thaw durability to reduce prescribed air content, were included. In addition to the performance limits, modulus of elasticity and density were required to be measured, but no limits were specified. In addition to revisions to the mix design, changes were made to the placement, finishing, and texturing portions of the specification. The goal of these revisions was to begin adequate wet curing as soon as possible. The original specifications for placing and texturing typically resulted in a delay

66 in the application of wet burlap to the surface of the bridge deck. This delay occurred because the texturing was done by tining transverse grooves with a metal comb and could not occur until the concrete was sufficiently stiff. After the bridge deck was tined, curing compound was applied. Fol- lowing the initial setting of the deck concrete, the presoaked burlap and soaker hoses were applied and kept in place for 14 consecutive days. Revisions to the curing portion of the specification now require fogging the deck immediately after the finishing machine passes. Tining of the bridge deck was eliminated and presoaked burlap applied almost immediately “without damaging the finish, other than minor marring of the concrete surface.” The use of curing compound was explicitly forbidden. Fogging must continue until the con- crete has achieved initial set when soaker hoses are added. The wet burlap and soaker hoses remain in place for 14 consecutive days. Another change to the specification required that the concrete temperature at the time of place- ment be between 55°F and 75°F. The original specification limited concrete temperature at time of placement to between 55°F and 90°F. In the spring of 2015, the undersides of 27 bridge decks constructed since 2007 were visually inspected for cracks: 15 bridges had been constructed using the performance-based specification, and 12 had been constructed using the traditional WSDOT specification. Seven single-span prestressed concrete bridges, six two-span prestressed concrete bridges, 10 multispan prestressed concrete bridges, and four steel plate girder bridges were selected. All bridges used I-girders for the ability to inspect the underside of the decks between the girders. The gathered information was converted into “crack inten- sity” diagrams to illustrate the severity and location of cracking for each bridge deck. The cracks were identified from the underside of the decks. Cracks on the top of the deck were not included. As illustrated in Figure 18, the performance-based concrete specification generally resulted in fewer visible cracks in bridge decks than did the traditional concrete specification. A few of the tradi- tional bridge decks performed similarly to the performance-based bridge decks, but this appeared to be the exception, not the rule. Only one of the performance-based concrete decks had a high intensity of cracking. It was unclear what contributed to the poor performance of this particular bridge deck. The performance-based concrete specification resulted in fewer cracks in decks on precast concrete girders compared with decks on steel girders. The study revealed that cracking of bridge decks varied within the same bridge. In some cases, it appeared to vary within the same concrete placement. The authors concluded that there are many (a) (b) FIGURE 18 Cracking in Washington State bridge decks: (a) before and (b) after (Courtesy: Washington State DOT).

67 variables that affect the cracking performance of a bridge deck, and these may change during the construction of the bridge. A similar effect was observed by Mokarem et al. (2009). A secondary objective of the WSDOT bridge study was to identify trends or issues with the performance-based specification that could be improved. Mix design, test data, and temperature information were gathered for the performance-based bridge decks evaluated in the study. No cor- relation could be made between the data and crack intensity; however, improvements in data col- lection on future projects may provide better data to identify trends or issues. Ultimately, based on the WSDOT study, no significant changes to the bridge deck concrete specifications were needed (Ferluga and Glassford 2015). The performance-based specification became part of the WSDOT Standard Specifications in 2015.