Bridges for Service Life Beyond 100 Years: Service Limit State Design (2014)

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.

41.1 Objectives of Project R19B The request for proposal for SHRP 2 Project R19B stated the following objectives: • Develop new design codes that incorporate a rational approach based on service limit states (SLSs) for durability and performance of bridge systems, subsystems, compo- nents, and details that are critical to reaching the expected service life and assuring an actual life beyond 100 years. Special focus should be given to problematic systems, sub- systems, components, and details. The proposed SLSs will include data sets related to durability, fatigue, fracture, and redundancy as integral issues of service life as reported in SHRP 2 Project R19A. • Develop performance measures incorporating predefined component classifications that will use full probability- based service life design criteria to maximize the actual life of the system. Consider material performance (including durability); structural performance of systems, subsystems, and components (optimum joints and bearings); and design practices leading to longer and more predictable service life. • Develop comprehensive design procedures, proposed speci- fication changes, and implementation tools that include durability design in addition to structural design. The development should also consider structural and material redundancy, and system, subsystem, and component per- formance measures that will use service life design criteria to maximize the actual life of the system. The adjustments to SLSs should not adversely affect ultimate or strength limit states (ULSs) and extreme event limit states. To best accomplish the project objectives, the project team first developed a list of the applicable SLSs for various com- ponents. A framework for calibration that accommodates aging and deterioration models, applicable loads, and other design parameters for the components was developed. Calibration was defined as the process of determining values of load and resistance factors so that the designed components will satisfy the selected reliability-based criterion (i.e., the reliability of the structure is close to the target value). Calibration involved the development of statistical models for load and resistance, selec- tion of the target reliability index, and reliability analysis. The products of this research, expected to be directly usable by the American Association of State Highway and Transporta- tion Officials (AASHTO) and departments of transportation (DOTs), include the following: • Provisions needed to implement SLSs and the associated load and resistance factors necessary to produce calibrated bridge components and systems expected to have a predict- able service life. When practical, the provisions are based on a 100-year life; if a component or system cannot reasonably be expected to last 100 years, the expected life is given. • Some detailed design and detailing provisions required to design and build the calibrated component or system. • Appendix F, which contains the databases used in the cali- bration, as well as instructions for a calibration spread- sheet for use by DOTs to track and adjust service-based reliability with time. It is expected that implementing own- ers will track deterioration and changes to load regimes with time and adjust built-in models and assumptions over time. It is assumed that the AASHTO LRFD Bridge Design Speci- fications (AASHTO LRFD) requirements are a package that has to be considered together. Owners who make exceptions to some AASHTO LRFD requirements will have to evaluate the findings of this research and decide their jurisdiction- specific requirements. The effect of the proposed specifica- tions revisions on specific types of components will be debated by AASHTO’s technical committees and Highway Subcom- mittee of Bridges and Structures (HSCOBS) when the revisions are considered. C h a P t e R 1 Purpose of Report and Relation to Scope

5 1.2 Scope 1.2.1 Original Scope As originally scoped, the project was broken into two phases containing the tasks described below. Phase 1 Task 1 Conduct a review of the literature to review and identify cur- rent practices of using SLS principles for determining struc- tural service life approaches. Task 2 Supplement the interim report provided by Project R19A at the beginning of this contract with a follow-up survey to identify successful systems, subsystems, components, and details that have lasted 100-plus years. Systems, subsystems, components, and details that have proven to or have the potential to solve common bridge durability and structural performance problems were of special interest for this task. In addition, identify problematic components and the nature and cause of failures that resulted in reduced service life, and document any available maintenance and rehabilitative costs. Compile any existing data regarding loadings and accelerated environmental testing results as documented by others for evaluating the performance of developed systems, sub- systems, components, and details. The research team was not aware of significant relevant information, so this type of data was not incorporated into the SLS calibration approach. Owner-supplied information can be used to supplement the calibration. Task 3 Develop an SLS approach that can be used to calibrate 100-plus-year service life. A benchmark in the calibration will evaluate the suitability of the existing 75-year load and resis- tance factor design (LRFD) approach presented in the cur- rent specifications. Task 4 Prepare an interim report documenting Tasks 1 through 3 and a detailed work plan for executing the SLS calibration. The research team proposed and the sponsor approved moving Task 5 into Phase 1. The work of an independent committee takes place in later tasks still left in Phase 2. Task 5 deals with identifying the members of the committee. Task 5 The research team proposed, for SHRP 2 approval, an inde- pendent national committee (INC) of experts to review and critique the suitability of the data set and the SLS approach. The INC included appropriate experts from AASHTO, the Federal Highway Administration (FHWA), state DOTs, industry, and academia (a minimum of seven volunteers is expected for this committee). Phase 2 Task 6 Conduct analytical trial runs as appropriate for evaluating the performance of systems, subsystems, components, and details developed under Project R19A, as well as existing sys- tems, subsystems, components, and details that are critical to reaching the expected service life beyond 100 years. Submit an interim report to SHRP 2 and to the approved INC. Task 7 Plan a working session with the INC to review and gather feedback on the interim report. Submit to SHRP 2 an updated interim report based on the input from the INC. Task 8 Incorporate the framework for the long-term bridge perfor- mance program (LTBPP), and use the LTBPP framework to validate performance expectations. Ensure that the ULSs and extreme event limit states are not compromised. The approach from the research team was to include recommendations for future bridge practitioners on how to adjust the SLSs to include semiprobabilistic assumptions developed under this project. Task 9 Develop a data set format that will be adaptable for future maintenance by the AASHTO Technical Committee on Loads and Loads Distribution (T-5). Task 10 Work with the Project R19A team and industry to develop recommendations for AASHTO-formatted LRFD design and load rating specifications and analysis methods, including detailed examples for bridge systems, subsystems, compo- nents, and details that incorporate results from Project R19A. Task 11 Develop an implementation plan suitable for adoption and maintenance by AASHTO that is based on the findings of this work. Task 12 Prepare a final report, including recommendations for future research. The first five tasks were developed in the Phase 1 report previously submitted and reviewed by the project expert technical group.

61.2.2 Revised Scope for Tasks 5 and 6 Task 5 was completed. Members of the INC were selected and approved. However, as the project unfolded, the expert technical group was augmented with additional experts in the areas of calibration and deterioration, and the value of the INC decreased accordingly. Copies of the Phase 1 report were provided to members of the INC in August 2011, but little response was received. Written comments received from one reviewer were very supportive of the approach outlined in the Phase 1 report, and those comments were submitted to the SHRP 2 staff. Task 6 of the research plan for Project R19B, as modified and submitted in November 2011, initially required the submission of an interim report documenting a proof-of- concept demonstration of the proposed calibration of SLSs. After consulting with SHRP 2 staff, the research team decided that the interim report served little purpose and that, due to the unforeseen difficulty in finding suitable data for calibration and other analytical work, the resources originally pro- grammed for the trial calibration runs, the interim report, and the working sessions with the INC (i.e., Tasks 6 and 7) could be better used to advance the other tasks. The specific require- ments for a revised Task 6 are described below. Revised Task 6 Conduct analytical trial runs as appropriate for evaluating the performance of systems, subsystems, components, and details developed under Project R19A, as well as existing sys- tems, subsystems, components, and details that are critical to reaching the expected service life beyond 100 years. Submit an interim report to SHRP 2 and to the approved INC. The research team developed serviceability provisions based on the findings and calibration approach outlined in Chapters 3 and 6. These provisions include improvements to the exist- ing service and fatigue limit states as shown below: • Load-induced fatigue of steel and concrete details and components; • Live load deflection; • Permanent deformation of compact steel components; • Cracking of reinforced-concrete components; • Tension in prestressed concrete components; • Settlement of foundations; • Horizontal movements of abutments; and • Slip of slip-critical bolted connections. Initially, the calibration was to proceed in two stages: a proof-of-concept stage involving a subset of the SLSs and a subset of parameters (random variables), followed by a pro- duction calibration involving all SLSs and a wider range of parameters. At the completion of a proof-of-concept partial calibration, the research team developed an interim report on findings. Once the calibration procedures were coded into spreadsheets, the value of trial runs or partial calibrations became insignificant. That report has been folded into the present report. A database of bridges was useful during the calibration process to assess current reliability versus the reliability resulting from proposed changes in design equations and methodologies, as well as selecting load and resistance factors. For this project, the database compiled under National Coop- erative Highway Research Program (NCHRP) Project 12-78 (Mlynarski et al. 2011) was selected as the source of sample bridges. The database contains information on over 18,000 bridges suitable for analysis using AASHTOWare’s Bridge Rat- ing analytical software. The NCHRP 12-78 database was sorted to select relatively modern bridges, potential candidate bridges were identified, and a partial list of candidate bridges was submitted to SHRP 2. As the calibration procedures were more fully developed and it was determined that insufficient data were available to fully calibrate some of the SLSs, the sample bridge population was used with three SLSs: cracking of prestressed concrete beams, settlement, and deflections. The sample bridge popula- tion was also used to investigate the ramifications of potential changes to two SLSs: cracking of prestressed concrete beams and overload of steel bridges. Subsets of the candidate bridge database used for these purposes are included in Appendix F. 1.3 Research team The organization and relationship of the primary team mem- bers are shown in the organizational chart in Figure 1.1. The functional lead responsibilities for leading individual tasks were as follows: • Task 1. The two universities. • Task 2. University of Nebraska–Lincoln (UNL) (Atorod Azizinamini took the lead). • Task 3. All components of the team participated in the work, but it was Modjeski and Masters’ role to see that the group reached a conclusion and a product. • Task 4. Modjeski and Masters, with help from the other team members. • Task 5. As with Task 3, this task was done by the entire group with Modjeski and Masters seeing that a successful outcome occurred. • Task 6. UNL took the lead. • Task 7. Modjeski and Masters took the lead. • Task 8. University of Delaware took the lead. • Task 9. University of Delaware took the lead, with assis- tance from UNL.

7 • Task 10. Modjeski and Masters took the lead, working pri- marily with the University of Delaware. • Task 11. Modjeski and Masters took the lead. • Task 12. Modjeski and Masters took the lead. 1.4 Relationship of Project R19B to Project R19a Projects R19A and R19B combined should have resulted in the development of AASHTO-formatted provisions for design of bridges capable of providing more than 100 years of ser- vice life. The provisions should address both existing and new bridges. The procedures have to be quantifiable for both existing and new bridges. One of the major tasks within R19A was identifying prom- ising systems, subsystems, components, details, and retrofit concepts capable of prolonging the service life of bridges at optimal total costs. R19A was to have developed details and subsystems requiring calibration or development (or both) of new limit state design provisions. The R19B work depended on R19A’s developing these details or subsystems. For selected ideas, R19A was also to have developed deterioration models. Incorporating these deterioration models into a general SLS design provision framework was to be a major undertaking within R19B. The choice of a general SLS design framework was an important issue that affected the research directions of both the R19B and R19A projects. As of this writing, no new details and subsystems have been recommended to R19B. One existing system, integral and semi-integral abutments, has been identified by R19A for calibration, but no limit states have been suggested or data- bases identified. 1.5 Relationship of Project R19B to NChRP Project 12-83 Several members of the research team were also involved with NCHRP Project 12-83, Calibration of LRFD Concrete Bridge Design Specifications for Serviceability. The goal of the NCHRP 12-83 project was to calibrate the concrete-related SLSs currently in the AASHTO LRFD (2012) and, as needed, to develop new calibrated concrete-related limit states for incor- poration into the AASHTO LRFD. Significant overlap exists between the SHRP 2 R19B and NCHRP 12-83 projects in the area of concrete structures. All aspects of the work under NCHRP 12-83 are fully applicable to SHRP 2 R19B. Most of the concrete-related aspects of this report were originally developed in NCHRP 12-83 and are incorporated here. 1.6 Special Challenges Related to SLSs The ULSs of the AASHTO LRFD are calibrated through structural-reliability theory to achieve a certain level of safety. They are intended to achieve similar component proportions to those of the Standard Specifications for High- way Bridges. These ULSs do not consider the integration of the daily, seasonal, and long-term service stresses that directly affect long-term bridge performance and subsequent service life. The current SLSs of the AASHTO LRFD are intended to ensure a serviceable bridge for the specified 75-year design life. These limit states are based on the traditional serviceability provisions of the Standard Specifications for Highway Bridges. The SLSs are not calibrated using reliability theory to truly SHRP 2 R19B TCC J. M. Kulicki, PhD, PE Principal Investigator D. R. Mertz, PhD, PE CO-P.I. A. S. Nowak, PhD CO-P.I. W. G. Wassef, PhD, PE CO-P.I. Management Limit States QA/QC Review Recommendations Limit States Calibration LTBPP Coordination Recommendations N. Samtani, PhD, PE Geotech Limit States Calibration A. Azizinamini, PhD, PE R19A Coordination Figure 1.1. Project R19B organizational chart.

8achieve a determined life with a specific level of certainty because the tools and data to accomplish this calibration were not available to the AASHTO LRFD code writers. The current AASHTO LRFD SLSs include limits on the following: • Live load deflection of structures; • Cracking of reinforced-concrete components; • Tensile stresses in prestressed concrete components; • Compressive stresses in prestressed concrete components; • Permanent deformations of compact steel components; • Slip of slip-critical friction bolted connections; and • Settlement of shallow and deep foundations, among others. The background of the current AASHTO LRFD (2012) SLSs is presented in Chapter 2. Some of these SLSs may relate to a specified design life; others do not. Many are presently deterministic, such as limiting the tensile stresses in pre- stressed concrete components to a level thought to result in a crack-free component. This SLS could be calibrated to achieve a certain probability of a crack-free component, but the calibration would include a service life only in determin- ing the live load the component must resist (e.g., a 75-year live load). To achieve the objective of developing the appropriate tools, candidate SLSs were evaluated against a set of criteria. This evaluation applied both to the retention of some of the existing SLSs in the AASHTO LRFD and any new limit states developed as part of this project and Project R19A. The crite- ria include the following: • Is the limit state quantitatively and qualitatively meaning- ful? Does it tell us something that we can use to maintain a structure in service and continue or extend its service life? • Can the limit state be calibrated? Can we develop limit state functions, such as indicated in Task 3, and develop a means either through the resources of Project R19B or by leveraging the results of Project R19A or the LTBPP to determine the data necessary to do a calibration? (When no such data existed, expert elicitation [Delphi process] was used to determine the range of data and the relative importance of certain characteristics in the data, including uncertainty, so that some calibration could proceed.) • Does a limit state really relate to the service life rather than to some other characteristic? For example, the Model Code for Service Life Design specifically states that it excludes fatigue as part of the SLSs (Fédération Internationale du Béton 2006). This exclusion may be in part because this document was developed primarily for concrete structures. The current AASHTO LRFD contains fatigue requirements under a separate limit state, the fatigue-and-fracture limit state. The assessment of fatigue life is very much related to the service life of steel structures. Should this limit state now be transferred to the SLSs? In many ways, fatigue is one of the more quantifiable and calibratable of the SLSs com- pared with those that may be developed dealing with dete- rioration of joints, bearings, coatings, and similar structural features. • Does it provide a method to evaluate the significance of interventions in extending the service life of the struc- ture component? Can the proposed limit states distinguish between interventions that slow deterioration and those that effectively halt deterioration for some period of time before it starts again? Can they respond to repairs that rein- state or increase load-carrying capacity? Consideration of SLSs requires different input data than ULSs require. In ULSs, the limit state function is defined with two variables: resistance (which is considered constant in time) and loads. For SLSs, a different approach is needed for the following reasons: • The definition of resistance is very difficult. • Acceptable performance can be subjective (full life-cycle analysis is required). • Resistance and load effects can be and often are correlated. • Load is considered as a function of time, described by mag- nitude and frequency of occurrence. • Resistance is strongly affected by quality of workmanship, operation procedures, and maintenance. • Resistance is subject to changes (mostly but not only dete- rioration) in time, with difficulty predicting initiation time and time-varying rates of deterioration (e.g., corrosion, accumulation of debris, cracking). • Resistance can depend on geographical location (climate, exposure to industrial pollution, exposure to salt as a deicing agent, or proximity to the ocean). In general, the consequences of exceeding SLSs are an order, or even several orders, of magnitude smaller than those associated with ULSs. Thus, an acceptable probability of exceeding an SLS is much higher than for a ULS. If the target reliability index (bT) for ULS is bT = 3.5 to 4.0, then for SLS, bT = 0.0 to 1.0 might be quite acceptable. The current AASHTO LRFD (2012) considers foundation settlement as an SLS. Foundation SLSs were probably the most difficult issue dealt with in R19B because of the wide range of physical parameters, numerous analytical solutions, and the regional nature of the practice of geotechnical engi- neering. Bridge foundations and other appurtenant struc- tures such as approach embankments should be designed so that their deformations will not damage the bridge super- structure or other structural elements or ancillary elements such as utilities, which are often attached to bridge structures.

9 Various aspects of deformations that should be considered in the design of bridges include • The effect of uneven settlement between various support elements; • The rotation and horizontal movements of the foundation system affecting movements at the bridge-seat level; and • Serviceability problems near a bridge abutment, in par- ticular the ubiquitous “bump at the end of the bridge” that affects joint serviceability and abutment performance. The cumulative effect of these deformations may generate uneven deformations and stresses across a bridge system and its subsystems. In the case of an irregular pattern of settle- ment, a reversal of stresses may occur in a bridge deck, result- ing in the deck cracking at various locations. Cracking allows moisture ingress, initiation of corrosion, and degradation of various bridge elements, resulting in reduced structural integrity. Thus, foundation deformations affect not only the quality of ride and the safety of the traveling public, but also the structural integrity of the bridge and its various compo- nents. In addition, such deformations often lead to costly maintenance and repair measures. The service life of a bridge structure, its components, or ancillary elements such as utili- ties attached to the bridge can be significantly affected by the deformation characteristics of the foundation system. In addition, it may be found that changes to material or construction specifications are a more effective way to deal with apparent serviceability issues than codified SLSs. This could be the case, for example, with deck cracking, for which changes to mix proportions or the use of curing practices designed to reduce shrinkage may be as effective as limit states based on strain calculations. The conservative nature of bridge engineering practice leads to one final special challenge for the development and calibra- tion of SLSs. This challenge is that the concern for public safety and the stewardship of public funds often results in a long institutional memory of past unsatisfactory experience. It is often a slow process to recognize when advances in technology or codification have addressed a past problem. In the case of SLSs in particular, which are often subjective, it is difficult to ascertain whether changes to design provisions have resulted in the desired improvement. It is analogous to the axiom that one cannot prove a negative. For example, several years ago major changes were made to the provisions for the design of modular expansion joints, particularly in regard to fatigue. Have these changes solved the problem so that the service life of these joints has been increased to the point that they need not be considered in this project? Has enough experience been gained to know? How much good experience is needed to alter any lingering perceptions based on earlier designs? To the research team, these issues imply that when the results of R19B lead to a reduced design requirement rather than a new or more strin- gent requirement, there is an enhanced need to thoroughly and prudently evaluate the design implications. 1.7 Serviceability Versus Deterioration Various researchers have considered deterioration of highway bridges and tried to track change over time for various types of bridge and service conditions (i.e., type of roadway) by using National Bridge Inventory (NBI) condition numbers or a similar state-specific index. Others have tried to relate dete- rioration to bridge type as the primary variable. Although the general deterioration of the bridge inventory is important from an administrative point of view, the specific impact on load-carrying capacity that might reduce the service life is a microlevel consideration. The various deterioration models are of limited value in that context. Nevertheless, they are part of the current state of the art and can inform an owner’s effort to account for the effects on resistance over time. With that in mind, several deterioration models found in the litera- ture that use information currently available to owners are reviewed in Chapter 4 of this report. As discussed in Chapter 2, a survey of bridge owners was conducted to identify which bridge components required sufficient periodic maintenance to be a significant factor in their maintenance budgets. The number of times 23 compo- nents were cited is shown in Chapter 2, Figure 2.6. 1.8 Durability Design Guide for Bridges for Service Life (Azizinamini et al. 2013), a product of SHRP 2 Project R19A, contains guid- ance for selecting system, subsystems, and components of bridges believed to promote long life. That information is not repeated here. Producing more durable bridges is best achieved through a holistic approach starting with type and location decisions through the entire bridge life to decommissioning. A study for the Alabama DOT that addressed virtually every aspect of the bridge delivery and maintenance system identified 57 fac- tors needed to provide more durable bridges (Ramey and Wright 1994). The DuraCrete report (2000) describes how the ability to quantify the durability and service life of a bridge changes during the design phase, construction phase ending with transfer (handing over), subsequent inspection and assess- ment phase, and possible repurposing phase. It is pointed out that the designer has the least accurate information about environmental loads, material properties, and quality of the constructed facility than at any other time in the service life of the facility.

10 Rostam (2005) describes two overarching strategies for addressing durability: • Strategy A, avoidance (such as use of corrosion-resistant rebar); and • Strategy B, selection of materials and details to resist dete- rioration for a given time. Reliability modeling of deterioration is relevant only for Strategy B, and European researchers have developed some deterioration models and reliability applications for concrete components (DuraCrete 2000). In particular, models have been proposed for corrosion of rebar from salt intrusion and carbonation. Some of the necessary data have been accumu- lated in Europe. Freyermuth (2009) lists the following options for achiev- ing extended service life of concrete bridges, although the extension is not quantified: • Use of high-performance concrete to decrease permeability; • Use of prestressing to reduce or control cracking; • Use of jointless bridges, or bridge segments, and integral bridges; • Use of integral deck overlays on precast concrete segmental bridges in aggressive environments; and • Selective use of stainless steel reinforcing. These important strategies may be regarded as high-level decisions that should be made before the detailed numerical design proceeds. Use of noncorrosive deicing and fixed anti- icing spray technology was also noted as an in-service strat- egy for enhancing concrete deck life, in particular. It is of interest to consider the number of railroad bridges that have served for over 100 years with minimal main- tenance. Although corrosion is often evident in railroad bridges, the severe attack of structural steel and reinforcing steel from deicing salt, in particular, is a major distinction in the deterioration of highway bridges, as is the pounding from truck traffic. Structures intended to provide at least a 100-year service life must have the following four attributes, which are dis- cussed in detail below: • Be conceived, sited, and designed to provide an acceptable level of reliability with respect to the natural environment and human-made loads. • Be properly constructed with suitable materials and details. • Be provided with adequate control of deck drainage, espe- cially in areas where deicing or environmental salt is applied. • Be treated with timely preventative maintenance of protec- tive coatings, drainage systems, joints, and bearings. 1. Acceptable level of reliability with respect to the natural environment and human-made loads. The particular mod- ification to the AASHTO LRFD proposed in Chapter 7 of this report relates to cracking of reinforced concrete, con- trol of stresses in prestressed concrete, control of fatigue cracking in steel and concrete construction, and settlement. This recommended modification, as well as the endorse- ment of current practice for limiting stresses due to over- loads, will contribute to reduced damage and hence extended service life. 2. Properly constructed with suitable materials. The benefits of quality construction are self-evident. Every DOT has con- struction and material specifications, as well as field and plant inspections, intended to ascertain that those require- ments are achieved in the completed project. It is outside the scope of this project to critique those processes. Gener- ally, concrete structures are adversely affected by ingress of salt, which leads to corrosion of embedded steel; chemical attack, such as alkali–silica reactivity and sulfate attack; and scaling, such as that associated with freeze–thaw cycles. Langley (1999) details some of the steps taken to address these issues on the Confederation Bridge. Mirza (2007) summarizes the concrete durability provisions of various Canadian Standards Association specifications. Some well-accepted durability-enhancing materials and processes are described below. The cost of these enhance- ments and the benefit achieved vary from state to state and even within a state. Environmental regulations and mainte- nance and protection of traffic can add tremendously to the total cost of maintenance operations, and these associated costs also vary widely. Therefore, no attempt has been made to quantify cost–benefit characteristics. • Salt intrusion is slowed by drainage control; providing suitable cover; use of dense, low-permeability concrete (such as high-performance concrete (HPC) and ultra- high-performance concrete); and control of cracking. • The effects of salt intrusion and depassivation due to carbonation can be mitigated by using corrosion inhib- itors, coated reinforcing, bimetallic reinforcement, stainless steel reinforcing, or nonmetallic reinforce- ment such as fiber-reinforced plastic composites. With- out citing cost–benefit specifics, it will generally be found that cost increases with each step in the reinforc- ing path above. • Aggregate reactivity issues such as alkali–silica reactivity are usually handled by prescreening possible sources by using laboratory tests to identify susceptibility. Most states have approved sources that largely eliminate aggregate reactivity. Use of low-alkali cement can also reduce sus- ceptibility of a concrete mix. • Sulfate attack is a result of the growth of minerals caused by reaction of chemicals in the cement with

11 sulfates in the mix; usually these sulfates are in the water, but they may be in the aggregate. Sulfate attack debonds the aggregate and creates expansive pressure leading to crack or delimitation. The causes and effects are similar to alkali–silica reactivity. Use of Type II, Type V, or blended cement is often indicated as well as use of approved material sources. Factors that reduce permeability are also usually helpful Detwiler (2008) states that “Maximum limits on the water-cementitious materials ratio, combined with good concreting practices—especially good curing— are even more important to sulfate resistance than the right cement.” ACI 201.2R-08 (ACI Committee 201 2008) provides recommended mix practices for vari- ous sulfate concentrations. • Freeze–thaw cycles can lead to scaling of the concrete surface due to pressure caused by the expansion of water in the concrete. Use of air-entraining, high- strength mixes and low permeability are effective countermeasures, although air entrainment can result in reduced strength and may not be compatible with HPC or high-strength concrete. Use of fly ash can be counterproductive if delayed strength gain exposes the concrete to freezing before sufficient strength has been developed. • Prestressing contributes to control of salt intrusion by reducing in-service cracking. Although some cracking may result from overloads, thermal gradients, and shrink- age, the cracks will generally close when the causative effect is reduced or eliminated. Beam ends exposed to salt-laden deck drainage have been found to be suscep- tible to corrosion damage resulting from water entering the beam via the strand ends. Tabatabai et al. (2004) documented the benefit of coating the end 2 ft with sealing materials and concluded that of four tested materials, a polymer resin coating was most effective and easiest to apply. Modern bridge steels are produced to tight tolerances of strength and element sizes. Toughness varies some- what more than other properties, but it usually exceeds the minimum specified values, sometimes substantially. Corrosion can have a significant effect on service life if not addressed. From a material point of view there are five general ranges of corrosion susceptibility provided by conventional steel, weathering steel, high-performance steel, 1035 steel (sometimes referred to as semistainless steel), and stainless steel. Improved corrosion resistance and cost increase with each step along the product line. Stainless steel has received little use in bridge construc- tion primarily due to cost. However, it has seen more use in recent years. Fatigue is an in-service design issue that is virtually unaffected by material choice and seems to be well addressed from the resistance side by current design criteria. Properly constructed, the key descriptor in the second attribute contributing to 100-year bridge service life, car- ries with it the requirement to provide sufficient field mon- itoring of construction to ensure that the work is executed within tolerances that are consistent with those assumed in the design. For example, the concrete cover is one of the major factors related to the rate of chloride intrusion and carbonation, yet it is difficult to control in the field unless suitable spacers are provided and the rebar cage, tendon ducts, and so forth are sufficiently tied to maintain their position during concrete placement. Proper consolidation, curing, and water control are critical, especially in regard to cracking and permeability. This requires vigilance by laborers, supervisors, and inspectors. 3. Providing adequate control of deck drainage. Damage caused by deck drainage, particularly salt-laden drainage, has been a major cause of deterioration in both steel and concrete bridges. The reduction in the number of deck joints through the use of continuous construction, combined with the widespread use of coated reinforcing, has reduced the impact of this problem. Although fully integral bridges have eliminated all deck joints, many bridges are still designed with some joints. In addition, most existing bridges contain joints, and they will be in use for a long time. 4. Timely preventative maintenance of protective coatings, drainage systems, joints, and bearings. Preservation of coat- ing systems is probably the most important step in the preservation of painted steel bridges and contributes to reduced permeability of concrete surfaces. Weathering steel bridges often have the area under and adjacent to deck joints coated, in which case preservation of that coat- ing is as important as maintaining the coating system on painted steel bridges. Maintenance of joints, troughs, and drainage hardware helps to control the flow of deck drainage to reduce dete- rioration of bearings, girders pier caps, and abutments. The use of continuity and integral and semi-integral abut- ments has been found to be effective in drainage control. Washing of bridges is usually thought to be a cost- effective means of bridge maintenance. However, Klaiber et al. (2004) found that for bridges on secondary roads, after 10 years deck washing did not produce significant improvement in deck durability. The effects of degradation were not included in the reported calibrations. It is assumed that maintenance will take place before deterioration significantly affects service load response. Further, the deterioration that might affect service response (other than appearance issues, which

12 could be considered an SLS for some bridges) could be quite different from that which affects deflection, vibra- tions, concrete cracking, and so forth. The condition of the bridge can be included as a change to resistance at a given point in time, and reliability indices can be recalcu- lated on that basis. Although details, materials, and techniques that are antici- pated to increase service life can be identified, the quantifica- tion of that increased life, or the change in reliability, is not generally possible at this time. An exception may be the rates of chloride ingress and carbonization for uncracked concrete under conditions similar to those in laboratory testing. For example, Fick’s Second Law of Diffusion has been used to estimate the time until chlorides reach a threshold value at reinforcement in the Confederation Bridge given cover, a dif- fusion constant, and a chloride content. Rostam (2005) lists the following parameters required to determine mix design qualities to provide a target service life: • The design surface chloride concentration; • The background chloride concentration foreseen in the concrete mix; • The chloride diffusivity; • The critical chloride concentration triggering corrosion of the reinforcement (the threshold value); and • The aging factor, represented by a decreasing diffusion coefficient with increasing age. Procedures are available (DuraCrete 2000) to develop a distribution of time to reach the threshold for a given cover from which the criteria exceedances, and hence a reliability index, can be found. To date this approach has not been widely used. Develop- ment of various parameters for regional or local material sources and concrete mixes would probably be needed for wide application. 1.9 Initial Coordination with FhWa Long-term Bridge Performance Program As indicated above, very little usable data have been found for use in developing and calibrating SLSs. FHWA recently initi- ated the long-term bridge performance program (LTBPP), which is intended to measure response factors for in-service bridges for as long as 20 years. This project could collect data needed for future development and improved calibration of SLSs, possibly even a full probabilistic approach. Task 8 in Phase 2 of this project requires that the R19B research team establish a dialogue with the LTBPP research team. SHRP 2 staff asked that this dialogue be started earlier, and a joint project coordination meeting was held in the autumn of 2009. Both teams recognized the benefit that could result from an open sharing of information and data needs. An initial list of worthwhile types of data that the LTBPP team might consider measuring in the bridges they will be instrumenting was presented to the LTBPP team for their consideration: • Put survey targets on substructures (piers and abutments), preferably starting with a bridge under construction, using some sort of laser monitoring to determine displacements and rotations with time. Possible foundation monitoring points were discussed with the LTBPP team. • Try to measure the relative and absolute movement between substructure and superstructure. • Collect data on the rate of aging of joints and bearings, including debris collection and initiation of leaking. • Collect data on vehicular damage to joints. • Try to collect data on traffic patterns, including convoying and lane usage. • Measure relative movements at and across joints, similar to what bridge inspection teams sometimes measure (some- times hard to relate to temperature). • Try to monitor longitudinal forces in structures. This could apply to the design of joints and bearings, as well as col- umns and foundations. • Try to determine if there is any in-service way to monitor change in friction with age and wear of expansion joints and bearings. • Try to monitor pavement growth and effectiveness of cycle control joints. • Assess corrosion loss or other elements of resistance change. • Monitor coating deterioration. • Try to measure something on jointless bridges involving the potential for pier damage or movements of integral and semi-integral abutments, pressure behind abutments, and movements and stresses in the piles of integral and semi-integral abutments. • Start to assemble data on the variability of prestress camber. • Monitor possible development of cracks in prestressed concrete beams and relate to overloads and environmental factors. • Monitor regional thermal gradients in superstructures, as well as large exposed box members such as tie girders and ribs. Verify Imbsen’s NCHRP study and extend to steel. • Monitor salt ingress regionally and relate to application rate and structural parameters. In order to remain in active contact with the LTBPP team, the R19B principal investigator accepted a position on the Transportation Research Board’s Long-Term Bridge

13 Performance Committee. In addition, one of the coprincipal investigators has been associated with LTBPP virtually throughout the project and is now technical director of the project. 1.10 Dialogue with aaShtO hSCOBS and Others Numerous presentations were made to keep the bridge com- munity apprised of issues related to Project R19B, to seek information for the project, and to gauge reactions of owners to potential new design requirements. Venues included the following: • The AASHTO Technical Committees: T-5 (Loads and Load Distribution) in 2011 and 2012; T-10 (Concrete Design) in 2012 and 2013; T-14 (Steel Design) in 2010, 2011, 2012, and 2013; and T-15 (Foundations) in 2012, as well as the full HSCOBS meeting in 2009, 2010, 2011, and 2012 • The 2012 Annual Meeting of the National Transportation Board in Washington, D.C. • The 2010 Annual Meeting of the Prestressed Concrete Institute in Philadelphia, Pennsylvania • The International Association for Bridge Maintenance and Safety meeting, 2012 • The U.S.–China Seminar on Highway Technology in Beijing, 2012 • The SHRP 2–Forum of European National Highway Research Laboratories Joint Symposium in Brussels, Bel- gium, 2010 • The poster session at the 2009 Annual Meeting of the National Transportation Board in Washington, D.C.