Guide Specification for Service Life Design of Highway Bridges (2020)

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.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 136 C H A P T E R 3 Development of the Guide Specification for Service Life Design of Highway Bridges This chapter describes the development of each section of the Guide Specification. For reference, the table of contents from the Guide Specification is included in Appendix C. Service Life Design Approaches The available approaches for service life design are defined in Section 1 – Introduction of the Guide Specification. These include the full probabilistic method, partial factor method, deemed-to-satisfy method, and avoidance of deterioration. The definitions are adopted from fib (2006) and ISO (2012b), which are used for the design of concrete structures; however, the design methods are applicable to other materials as well. Section 1 also includes background information, objectives, approach, future improvements, and definitions and notation. Service Life Categories Section 2 – Classification contains guidance on service life categories based on desired levels of quality for permanent and renewable bridge elements. For permanent elements, three categories of service life are established: normal, enhanced, and maximum, with corresponding good-better-best levels of qualitative practice. Specific values for target service life are not presented within the specifications such that a false degree of accuracy is not implied to the designer. There are numerous uncertainties involved in predicting target service life for a bridge; therefore, it implies an unrealistic level of precision to assign “hard” values to service life. However, values for target service life had to be established in order to perform the calculations for the calibrated probabilistic based limit state for chloride-induced corrosion, discussed later in Section 2.4.2.3.7 of this report. These values are only presented within the commentary of the Guide Specification to be transparent regarding how any design requirements were established for the probabilistic based limit state. For renewable elements, separate service life categories are not established. Rather, the target service life for renewable elements is determined on a case-by-case basis. For joints and bearings, the target service life is first established based on the practical life span on these elements, with guidance provided in Section 7 of the Guide Specification. Additional factors for the designer to consider are then presented, such as the ADT on a joint system and the existence of a joint above a line of bearings that will eventually leak and accelerate bearing corrosion and deterioration. For barriers and railings, the target service life is established based on the connection to the deck or superstructure, as well as the type of material. For integral connections to the deck, the target service life should match that of the deck; while replaceable barriers and railings may have a shorter service life than that of the deck.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 137 Environmental Exposure From a service life standpoint, the environment surrounding a bridge and its components represents the load on the structure. Standards and specifications typically define the environmental exposure on an element in terms of exposure zones and exposure classes. The establishment of exposure zones and classes is included in Section 2 of the Guide Specification in addition to the service life categories. Exposure Zones Exposure zones classify a bridge based on local site conditions such as proximity to a coastline, temperature, and other climate conditions. These zones are termed macro environment exposure zones in the Guide Specification and include the following: • Rural/Mild/Nonaggressive: little to no exposure to airborne or applied (i.e., deicing) salts. Low pollution from sulfur dioxide, low humidity and precipitation, and no exposure to chemical fumes. Typically, an inland location (SSPC 1996, ISO 2012a, FHWA 2015, ISO 2017). • Industrial/Moderate: occasional exposure to airborne salts or deicing salt runoff. Non-coastal bridges with irregular deicing salt application. Industrial areas with airborne contaminants, polluted urban areas, areas with moderate to high humidity (SSPC 1996, ISO 2012a, FHWA 2015, ISO 2017). • Marine: coastal areas with exposure to airborne salts or direct contact with sea water or brackish water. Typically defined by a limiting distance from a coast that depends on wind and other weather conditions (AS 2004, ISO 2012a, UFGS 2012, Caltrans 2014, ODOT 2016, FDOT 2017, ISO 2017). A distance of 0.5 miles from a body of salt water is established in the Guide Specification for cases where governing local/regional/state specifications do not provide guidance. • Deicing: region where deicing salts are used on a regular basis during the winter. • Buried: permanently buried in soil, below the finished grade or mudline, determined after consideration of all applicable scour (Tomlinson and Woodward 2008, Hannigan et al. 2016). In addition, exposure zones are used to identify the surrounding environment for individual elements, such as whether an element is in contact with water, soil, or the atmosphere. The term micro environment exposure zone is used in the Guide Specification for these classifications and include the following definitions: • Buried zone: See definition above. Currently only one micro exposure zone is defined for buried elements. It is anticipated that additional micro zones will be added in the future. • Submerged zone: permanently submerged in water, below the tidal or water level zone. Portion of the structure below Mean Lower Low Water (MLLW) level. For areas with minimal tides, portion of the structure below Mean Sea Level (MSL) (Tomlinson and Woodward 2008, UFGS 2012, Hannigan et al. 2016). • Tidal zone/Water Level zone: not permanently submerged in water, subject to wet-dry cycles (e.g., due to wave action). Any portion of the structure between MLLW and Mean Higher High Water (MHHW). For areas with minimal tides, portion of the structure between MSL and Mean High Water (MHW). For structures not in a tidal environment but where water level can vary, the water level zone is the portion between low water level and high water level. A sub-zone within the tidal zone, the Low Water Zone, is defined as the zone between lowest low water and highest low water (see Figure 51) and is specifically used in the estimation of corrosion for steel piles (West et al. 1999, Tomlinson and Woodward 2008, UFGS 2012, Caltrans 2014, Hannigan et al. 2016). • Direct Deicing Salts zone: directly exposed to the use of deicing salts. • Indirect Deicing Salts zone/Splash zone/Spray zone: indirectly exposed to deicing salt whether through roadway spray and/or roadway splash. For roadway splash/spray, the zone extends horizontally a distance x from the edge of the roadway and vertically a distance y from the roadway surface, as shown

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 138 in Figure 52. The splash and spray zone also exists above the tidal zone and is subject to splash and spray from waves. • Atmospheric zone: not exposed to soil, water, or deicing salts. Also, the region above the splash/spray zone when located in a tidal environment. • Interior: not exposed to the exterior environment. This zone was defined in order to be consistent with Section 5 of AASHTO LRFD. • Other Exterior: for exterior exposure conditions that do not fit into any of the above definitions. This zone was defined in order to be consistent with Section 5 of AASHTO LRFD. Figures are included within the article to help the designer assign exposure zones. The figures are included as examples only and should be modified as necessary based on the local site conditions. The first figure (Figure 51 below) shows the micro environment exposure zones with associated limits for each zone where applicable. The figure was developed from information provided in Morley and Bruce (1983), West et al. (1999), Tomlinson and Woodward (2008), Caltrans (2010), UFGS (2012), and Hannigan et al. (2016). Notes: 1For unprotected locations, the 20 feet area above the tidal zone (UFGS 2012, Caltrans 2014). For locations protected by seawalls or otherwise sheltered from open-ocean waves, 6 feet area above tidal zone (UFGS 2012). 2If subject to splash/spray/runoff due to joint failure. Figure 51. Micro Environment Exposure Zones [Guide Specification Figure 2.2.1.2-1].

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 139 The second figure (Figure 52 below) in the article on exposure zones defines the roadway splash/spray zone limits, x and y. The values used for x and y are adopted from EN 1992-2 (2005) and CSA (2014). EN 1992-2 recommends values of 6 meters for x and 6 meters for y. CSA recommends 10 meters for x and 3 meters for y. The recommended values of 30 feet and 20 feet for x and y, respectively, represent an envelope of the EN 1992-2 and CSA recommendations (with unit conversion and rounding tolerances) and are conservative. Notes: 1x may be taken as 30 feet in lieu of other guidance 2y may be taken as 20 feet in lieu of other guidance Figure 52. Roadway splash/spray zone [Guide Specification Figure 2.2.1.2-2]. Exposure Classes Exposure classes are defined for concrete and steel structures and are grouped by the deterioration mechanisms associated with each material. For exposure classes specific to foundations and retaining walls, the designer is directed to Section 6. Concrete Structures For concrete, the following exposure classes are defined based on common deterioration mechanism: • Corrosion (C) • Freezing and Thawing (FT) • Sulfate (S) • In Contact with Water (W) The exposure classes are presented in a tabular format with associated environmental conditions (i.e., exposure zones) and descriptions of each class, duplicated in Table 63. The classes for Exposure Class C are defined for use in the calibrated deemed-to-satisfy provisions discussed later in this report. Four of the previously defined macro exposure zones are used to group the exposure classes: Nonaggressive (NA), Buried (B), Deicing (D), and Marine (M). Each macro zone is subdivided into the previously discussed micro exposure zones that consider the exposure for specific components of the structure. Exposure Class FT, S, and W were adopted from ACI-318 (2014). Provisions from CSA (2014) were also used in the definition of Exposure Class S. Exposure classes were defined to allow for future refinement and expansion. For example, currently only one corrosion exposure class exists for buried elements (C-B). It is anticipated that additional C-B exposure classes will be created when sufficient data becomes available on the parameters that govern the corrosion of buried reinforced concrete. Edge of Roadway y (see Note 2) x (see Note 1) Centerline of Roadway

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 140 Table 63. Concrete Exposure Classes [Guide Specification Table 2.2.2.1-1]. Category Class Condition Description Corrosion (C) Nonaggressive (NA) Environment C-NA1 Interior exposure Sheltered surfaces, mostly dry conditions. C-NA2 Other exterior exposure Buried (B) Environment C-B Buried Buried elements (deep and shallow foundations, retaining structures, etc.) below finished grade (mudline in case of submergence). Deicing (D) Environment C-D1 Atmospheric in a deicing salts environment Surfaces exposed to airborne chlorides. Examples: abutments, piers, pile caps, girders, underside of decks. C-D2 Indirect deicing salts Surfaces exposed or potentially exposed to drainage water containing deicing salts or roadway spray. Examples: surfaces below expansion joints or drains, deck fascia, substructure surfaces near a roadway. C-D3 Direct deicing salts – low Direct exposure to deicing salts with low to medium application rate of deicing salts. Examples: top of decks, curbs, sidewalks, barriers. C-D4 Direct deicing salts – high Direct exposure to deicing salts with high application rate of deicing salts. Examples: top of decks, curbs, sidewalks, barriers. Marine (M) Environment C-M1 Marine – atmospheric Surfaces exposed to airborne chlorides. Examples: superstructure. C-M2 Marine – submerged Permanently submerged with salt water present. Examples: substructures between mudline and tidal zone. C-M3 Marine – tidal or splash/spray zone Surfaces in contact with salt water in the tidal zone or splash/spray zone. Examples: substructures within tidal zone or splash/spray zone. Freezing and Thawing (FT)1 FT0 Not exposed to cycles of freezing and thawing Members in climates without freezing temperatures; foundations not exposed to freezing; members buried below frost line. FT1 Exposed to cycles of freezing and thawing, limited exposure to water but without exposure to chlorides Vertical surfaces exposed to water and freezing; walls and columns. FT2 Exposed to cycles of freezing and thawing, frequent exposure to water but without exposure to chlorides Horizontal surfaces exposed to water and freezing. FT3 Exposed to cycles of freezing and thawing, frequent exposure to water and chlorides Bridge decks exposed to deicing chemicals; surfaces exposed to direct spray containing deicing chemicals and freezing; splash zones of marine structures exposed to freezing.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 141 Table 63. Concrete Exposure Classes (continued) [Guide Specification Table 2.2.2.1-1]. Category Class Condition Description Sulfate (S)1, 2 Water-soluble sulfate (SO42-)3 Dissolved sulfate (SO42-) in water (ppm)4 In soil (% mass) In recycled aggregate (% mass) S0 SO42- < 0.10 SO42- < 0.20 SO42- < 150 Conditions where water-soluble sulfate concentration in contact with concrete is low and injurious sulfate attack is not a concern S1 0.10 ≤ SO4 2- < 0.20 0.20 ≤ SO42- < 0.60 150 ≤ SO42- < 1500 or seawater Moderate sulfate exposure; sea water exposure S2 0.20 ≤ SO4 2- < 2.00 0.60 ≤ SO42- < 2.00 1500 ≤ SO42- < 10,000 Severe sulfate exposure S3 SO42- > 2.00 SO42- > 2.00 SO42- > 10,000 Very severe sulfate exposure In contact with water (W)1 W0 Concrete dry in service; concrete in contact with water and low permeability is not required Dry in service or in contact with water but no specific requirements for low permeability W1 Concrete in contact with water and low permeability is required Need for concrete with low permeability to water; penetration of water into concrete might reduce the durability of the member Notes: 1Adopted from ACI (2014) 2Adopted from CSA (2009) 3Percent sulfate by mass in soil shall be determined by ASTM C1580 4Concentration of dissolved sulfates in water, in ppm, shall be determined by ASTM D516 of ASTM D4130 Steel Structures Two exposure classes are defined based on the most common deterioration mechanisms for steel: • Corrosion (C) • Fatigue (F) Similar to concrete, the exposure classes for steel are presented in a tabular format with guidance on environmental conditions and descriptions for each class, repeated in Table 64. Exposure Class C is adopted from ISO (2017) and categorizes the corrosion potential of the environment using five classes from “very low” to “very high marine”. For the fatigue Exposure Class F, the designer is pointed to AASHTO LRFD for fatigue design. While the research team recognizes the importance of designing for fatigue this area is well covered in AASHTO LRFD, and any guidance provided here would be insufficient and redundant.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 142 Table 64. Steel Exposure Classes [Guide Specification Table 2.2.2.2-1]. Category Class Condition Description Corrosion (C)1 C1 Very low C2 Low Atmospheres with low level of pollution, mostly rural areas C3 Medium Urban and industrial atmospheres, moderate sulfur dioxide pollution, coastal areas with low salinity C4 High Industrial areas and coastal areas with moderate salinity C5I Very High Industrial Industrial areas with high humidity and aggressive atmosphere C5M Very High Marine Offshore areas with high salinity and sub-tropical and tropical atmospheres Fatigue (F) See Article 6.6 of AASHTO LRFD Bridge Design Specifications Notes: 1Adopted from ISO (2017) Foundations and Retaining Walls For foundations and retaining walls, the designer is referenced to Section 6 of the Guide Specification. The testing methods and associated limits vary among types of foundation elements (e.g., limiting soil pH values for piles versus retaining walls). Therefore, the exposure classes for foundation elements are presented in Section 6 within each article. Ideally, all exposure class definitions would be included in Section 2, however due to the fact that testing methods and associated limits are different for buried structures, it was found that presenting this information in Section 6 will minimize confusion for the designer. It is understood that in theory a unified system of exposure classes should be able to be defined for all element types and all materials. The exposure classes represent the environmental loads on the structure and are independent of element and material. For example, the concentration of chlorides in sea water is not influenced whether in contact with a concrete drilled shaft or a steel pile. While the deterioration mechanisms vary by material type, the exposure classes not associated with any of the mechanisms for a particular material could simply be noted as “not applicable”. Protection Strategies The protection strategies employed to resist the applied environmental loads and reach the target service life are given in Sections 3 through 7 of the Guide Specification, each for a different material or type of bridge element. The majority of these sections provide design provisions in the form of deemed-to-satisfy and avoidance of deterioration. Probabilistic methods are used to calibrate deemed-to-satisfy provisions for the chloride-induced corrosion limit state of reinforced concrete. In addition to specific design requirements, articles on good detailing practices are included in select sections. General Design Guidelines Section 3 of the Guide Specification addresses general aspects of service life design. Guidance is provided related to the planning stage of a project, global bridge design, common durability problems, and good detailing practices. The article on planning provides the designer with a list of considerations for the initial stages of a project that can have a direct impact on the service life of a bridge, including environmental concerns, drainage, material selection, efficiency of the design, climate change, substructures and foundations, maintenance

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 143 and inspection, and construction oversight. There are a number of factors that affect service life that must be addressed by the designer at the planning stage. Some of these factors are fairly well established, such as traffic volumes, clearances, and the environment. Others, however, are often unknown, such as changes in functionality and future maintenance and preservation actions, and therefore may be more critical to the service life. While many of these factors are out of the control of the designer, it is important that they be recognized at the planning stage so that design measures can be implemented that account for their potential variability. In addition, the article on planning covers a number of decisions that often take place during the conceptual design phase that have significant consequences on service life. In particular, the reduction and elimination of joints is stressed throughout the article, as this is the most effective measure to extend service life. Aspects of global design are then presented in a subsequent article and focus on how the failure of one component in a service life sense can detrimentally affect the service life of a different component, how to plan for the replacement of renewable elements, and how to assess the total bridge service life as the sum of the service lives of individual components. The designer is then presented with common durability problems that should be avoided whenever possible. These problems are separated as follows, with a top down approach: • Drainage • Decks • Joints and Bearings • Concrete Structures • Structural Steel • Foundations and Retaining Walls • Utilities and Appurtenances Photographs showing examples of poor durability performance are included to help the designer avoid making similar mistakes in new design. This is different from the guidance provided in traditional design specifications in that the designer is guided on what not to do. However, the research team felt that these examples were important to include because many of the durability problems presented are associated with commonly used structural designs and details. A subsequent article gives general durability considerations that aide the designer in making decisions that will help to achieve the target service life. The following topics are included: • Drainage • Deck Detailing • Joints • Bearings • Utilities and Appurtenances • Wildlife • Access and Inspection • Construction and Preservation Many of the considerations in this article are related to small details that can have a large impact on the achieved service life. The effect of joints on the service life of other components and the use of jointless designs is purposefully stressed again in this article.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 144 An article on documentation is included at the end of Section 3, and provides guidance on the development of a service life design manual and a preservation plan. The purpose of the service life design manual is to document all assumptions made by the designer related to service life including: • design assumptions – exposure conditions, deterioration mechanisms, material properties, etc. • assumed service life of replaceable components • assumed service life of coatings and wearing surfaces • vulnerable components or areas of the bridge expected that may necessitate added attention during inspections • recommended non-destructive evaluation (NDE) or special inspection techniques • a schedule of preservation and maintenance actions expected over the service life of the bridge The preservation plan is intended to be used for major and complex bridges, particularly in relation to inspection frequency and maintenance and preservation actions. A sample preservation plan is provided in Appendix D. The designer is pointed to other sections of the Guide Specification for more detailed information, where applicable. Concrete Structures Section 4 contains design provisions for concrete structures. The section begins with several articles that discuss the common deterioration mechanisms of reinforced concrete and provide general mitigation approaches for each mechanism. This is followed by an article on material protection strategies specific to each deterioration mechanism, which focuses on design provisions for the concrete exposure classes defined in Section 2. General Requirements General requirements for all exposure classes, including maximum water-cement (W/CM) ratio and minimum compressive strength requirements, are presented in tabular format, replicated in Table 65 below. The requirements for W/CM ratio and compressive strength are based on design provisions in CSA (2009), ACI (2014), and BSI (2014). The designer is directed to other articles for exposure zone specific requirements. The minimum AASHTO concrete class is included in Table 65 for each exposure class in order to be consistent with the AASHTO LRFD Bridge Design Specifications (2017a) and the AASHTO LRFD Bridge Construction Specifications (2017b).

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 145 Table 65. Concrete Material Requirements (CSA 2009, ACI 2014, BSI 2014) [Guide Specification Table 4.2.1-1]. Exposure Class Maximum W/CM1 Minimum f′c (ksi) Additional Requirements Minimum Class of Concrete2 All Corrosion (C) Exposure Classes Article 4.2.4 N/A Article 4.2.4 N/A FT0 N/A 2.5 N/A B FT1 0.55 3.5 Article 4.2.2 B (AE) FT2 0.45 4.5 A (AE) FT3 0.40 5.0 A (HPC) S0 N/A 2.5 N/A N/A S1 0.50 4.0 Article 4.2.3 A S2 0.45 4.5 A (HPC) S3 0.40 4.5 A (HPC) W0 N/A 2.5 None B W1 0.50 4.0 None A Notes: 1For concrete on or over saltwater or exposed to deicing chemicals, maximum W/CM shall be 0.45 (AASHTO, 2017b) 2As defined in the AASHTO LRFD Bridge Construction Specifications (AASHTO, 2017b) Freeze-Thaw and Sulfate Attack Subsequent articles give specific requirements for freeze-thaw attack and sulfate attack. The design requirements for freeze-thaw and sulfate attack are adopted from ACI 318 (2014), supplemented by CSA (2014) provisions where applicable. For freeze-thaw attack, air content requirements are given for each exposure class. Similarly, for sulfate attack, requirements for minimum permissible cementitious materials types are given for each exposure class. Probabilistically Calibrated Deemed-to-Satisfy Provisions for Chloride-Induced Corrosion The requirements for corrosion are centered around calibrated deemed-to-satisfy provisions for chloride- induced corrosion, supplemented by discussion on the assumptions and limitations used to create the requirements. The provisions use the combination of concrete cover thickness, concrete materials and quality, and type of reinforcing steel to specify an adequate design, and are based on a probabilistic approach for chloride-induced corrosion in concrete structures. The approach used is based on fib Bulletin 34 (2006), which uses a chloride diffusion model to predict corrosion. Deterioration mechanisms other than chloride-induced corrosion are not considered in this model and were considered separately by the research team. It should be noted that the diffusion approach to service life modeling is not without deficiencies. In particular, the fib Bulletin 34 model does not account for cracking. The effect of cracking, its extent and magnitude, on the diffusion of chlorides is recognized but not well understood at this time. However, the concrete cover requirements (based on concrete quality) that result from using the fib model appear reasonable; it must be noted that there are few practical alternatives to this approach. The following sections explain the modeling and the probabilistic analyses used to develop the guideline provisions. The fib Bulletin 34 model is based on metric units hence metric units are presented in this report

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 146 with corresponding U.S. units when relevant. U.S. units are solely used in the Guide Specification in accordance with AASHTO publication requirements. Limit State For concrete structures, a two-phase service life model is used to represent the development over time of chloride-induced corrosion: • The Initiation Phase – During this phase no noticeable weakening of the material or the function of the structures occurs. Chlorides from the surrounding environment penetrate into the concrete and diffuse further inward, toward the reinforcement. Once a critical concentration is reached at the surface of the steel, chloride ions disrupt the passive layer around the steel reinforcement. The dissolution of the passive layer and the decrease of the concrete pH promote the initiation of corrosion. The length of time necessary for this process to occur can be highly variable depending on the concrete properties, depth of cover to the reinforcement, and the exposure conditions. • The Propagation Phase – At the start of this phase, the protective barrier on the steel surface is broken down and critical levels of chlorides are reached, such that during the propagation phase an active deterioration develops, and accumulation of damage commences. In many cases, corrosion develops at an increasing rate with time. The two-phase model of chloride-induced corrosion is illustrated in Figure 53. Figure 53. Two-phase Modeling approach to Deterioration Specific to Chloride-Induced Corrosion [Guide Specification Figure C4.2.4.2.1a-1]. For the purpose of the service life modeling used to develop the provisions in the Guide Specification, the nominal service life is taken as the corrosion initiation time which marks the beginning of the propagation phase. In reality, the structure does not become unserviceable at the end of the initiation phase. However, this definition of the limit state is consistent with the objective of having concrete structures with minimal maintenance requirements over the service life. The calculations for the time to corrosion initiation model uses a full probabilistic approach where the probabilistic nature of the input parameters (both the material resistances and the environmental stresses) and intrinsic model uncertainties are taken into account. In this way, final design values for these parameters can be established which will delay the onset of corrosion sufficiently to achieve the required service life. Initiation Phase = Service Life Propagation Phase Corrosion Initiation Time Deterioration Age

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 147 Based on guidance provided by fib Bulletin 34, a confidence level of 90% that corrosion will not be initiated within the targeted corrosion initiation time (which, in this case, is equal to the target service life) was used for modeling. This confidence level is slightly higher than the ranges used in the SHRP 2 R19B (Kulicki et al. 2015) calibration of service limit states, due to the irreversible nature of corrosion damage and the uncertainties that go into the deterioration model. Mathematical Model The calibration was performed using the full probabilistic model in fib Bulletin 34, which utilizes Fick’s 2nd law of diffusion to estimate the rate of chloride penetration through the concrete cover. The model predicts the time necessary for a sufficient concentration of chlorides to diffuse through the cover layer to depassify the rebar and allow corrosion to occur. The equation for the chloride concentration at the level of the top reinforcing bars is: ( )0 , 0 , ( ) 1 2 , a i p r p C c t S xC r xC a ax eC C f D t t C∆   − ∆  ⋅ −  ⋅  =  = + −  =  (26) where: Ccrit = critical chloride concentration C0 = initial chloride content of concrete, taken as a constant CS,Δx = chloride content at a depth a, random variable erf = error function a = concrete cover over reinforcing, random variable Δx = depth of convection zone, random variable Dapp,C = apparent chloride diffusion coefficient t = time Dapp,C is determined using the following expression: , ,0 ( )app C e RCMD k D A t= (27) where: ke = environmental variable to account for temperature DRCM,0 = chloride migration rate, determined by rapid laboratory test, random variable A(t) = aging function ke is calculated as follows: 1 1expe e ref real k b T T    = −       (28) where: be = temperature coefficient, random variable Tref = reference temperature, taken as a constant Treal = temperature of structure, random variable The aging function is determined as follows:

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 148 0( ) tA t t α  =     (29) where: α = aging exponent, random variable t0 = reference point of time (reference age) at which the rapid chloride migration test is performed The variables most easily controlled by the designer are the concrete material properties (chloride migration coefficient and age factor), concrete cover, and the critical chloride content (via the type of reinforcement). Changing the values associated with these variables results in different times to corrosion initiation, and hence service life. Final design values for these variables can be established which will delay the onset of corrosion sufficiently to achieve the required service life for a given exposure class. Treatment of Buried and Nonaggressive Environments For both buried and nonaggressive environments, there is a lack of accepted data for several of the input parameters needed to use the model defined by Equation 26. The majority of research related to chloride- induced corrosion has focused on aggressive environments, such as deicing and marine, as accounting for these types of environments in design is critical to the durability of a structure. In addition, the environmental conditions encountered in soil can be highly variable, both between different sites and between soil types and depths at the same site location. This has made quantification of required input parameters difficult without costly in situ measurements. Therefore, in order to provide calibrated design provisions for these two environments, assumptions for certain input parameters had to be made. The assumption was made that for buried and nonaggressive environments current practice is adequate to achieve a 75 year service life (equal to the design life defined in AASHTO LRFD), even for buried structures in marine environments. A parametric study indicated that the current state of practice could be emulated by varying several parameters such as the chloride migration coefficient, aging exponent, or surface chloride saturation concentration either alone or in various combinations. Within subsequent sections presenting the background theory behind the calibrated deemed-to-satisfy provisions, assumptions for the input parameters specific to buried and nonaggressive environments are explicitly discussed. Input Parameters – Exposure Conditions For each C Exposure Class from Table 63, the following input parameters are defined as shown in Table 66: ambient temperature (Treal), chloride surface concentration (CS and CS,Δx), and the depth of convection zone (Δx).

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 149 Table 66. Input Parameters for Exposure Conditions. Input Parameter Temperature, Treal Chloride Surface Concentration, Cs or CS,Δx Depth of the Convection Zone, Δx dist. mean, μ stdev, σ dist. mean, μ stdev, σ dist. mean, μ stdev, σ a b Exposure Zone (°F) (°F) (%)1 (%) (in) (in) (in) (in) Deicing Environment Underside of Deck Normal 50 3.6 Lognormal 1.0 0.5 Constant = 0 Atmospheric 2.0 1.0 Indirect Deicing 2.0 1.0 Beta 0.35 0.22 0.00 2.00 Direct Deicing-Low 3.0 1.5 Direct Deicing- High 4.0 2.0 Marine Environment Atmospheric Normal 642 3.6 Lognormal 2.0 1.0 Constant = 0 Submerged 3.0 1.5 Tidal or Spray 4.0 2.0 Beta 0.35 0.22 0.00 2.00 Buried Environment Buried Normal 64 3.6 Lognormal 0.6 0.3 Constant = 0 Nonaggressive Environment Interior Normal 64 3.6 Lognormal 0.7 0.3 Constant = 0 Other Exterior 1.0 0.5 Notes: 1 % are given by mass of total cementitious materials 2 taken as 77 for hot marine environments The chloride diffusion rate is dependent on the temperature of the concrete and is accounted for in Equation 26 through the environmental transfer variable ke, as defined in Equation 28. The higher the temperature the higher the diffusion rate. Therefore, the mean temperature for each exposure zone was chosen to be on the upper end of possible mean temperatures within the U.S. For deicing environments, a mean value of 50°F was selected, and is representative of a location where deicing salts are regularly used in the winter but with a warmer mean annual temperature in comparison to the northernmost states. For marine environments, a mean value of 64°F was chosen and represents typical coastal U.S. locations (e.g., the West Coast, the Gulf Coast). To accommodate select U.S locations in which the mean annual temperature approaches 80°F (e.g., Hawaii, south Florida), an additional calibration was performed for hot marine environments assuming a mean annual temperature of 77°F. For buried and nonaggressive environments, the mean annual temperature was assumed equal to that of the marine exposure zone (64°F). While many buried and nonaggressive environments experience temperatures much lower than 64°F, and therefore a lower diffusion rate, this assumption was chosen to encompass a majority of the U.S. and produces conservative calibrations. For all exposure zones, the standard deviation of the temperature distribution was assumed to be 3.6°F, and represents the deviation from the mean annual temperature over a long period of time (e.g., 20+ years). The chloride concentration at the surface CS or at the substrate surface CS,Δx represents the loading on the concrete structure from the environment. CS,Δx is the chloride concentration below the convection zone,

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 150 which extends a distance Δx below the concrete surface. CS,Δx and Δx account for the deviation of chloride diffusion from Fick’s second law within the convection zone due to wetting and drying cycles (Figure 54). Values for CS, CS,Δx, and Δx for deicing and marine environments were chosen based on guidance provided in fib Bulletin 34 (2006) and fib Bulletin 76 (2015). Source: fib (2015) Figure 54. Deviation from Fick’s Second Law within the Convection Zone [Guide Specification Figure CA.2.3.3-1]. As introduced previously, assumptions were made for the input parameters for buried and nonaggressive environments such that the calibrated design provisions matched the current state of practice (i.e., AASHTO LRFD) for the normal service life category. A parametric study indicated that varying only the surface chloride concentration could produce covers that were reasonably consistent with apparently successful past practice for the Normal service life category. Varying only one parameter at this time had the virtue of simplicity. In addition, varying the surface chloride concentration allowed the buried and nonaggressive exposure classes to be isolated from the other exposure classes; whereas is if parameters associated with material properties were varied, all exposure classes could be affected. When surface chloride saturation concentration data is lacking, as is the case for buried components, fib Bulletin 34 provides the steps in Figure 55 that are needed to determine this key parameter. Per fib (2006), the surface chloride concentration is dependent on: • material properties • component geometry • environment

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 151 Source: fib (2006) Figure 55. Information needed to determine the variables CS and CS,Δx [Guide Specification Figure CA.2.3.3-2]. The equivalent chloride concentration of the ambient solution Ceqv governs the environmental impact on the surface chloride content (fib 2006) and is determined by the environment surrounding the concrete component (e.g., marine, deicing, etc.). The following material properties are required in order to calculate the chloride saturation content CS,0 (fib 2006): • chloride adsorption isotherms for the selected cement type • concrete composition Through the correlation between Ceqv and the chloride-adsorption-isotherm, CS and CS,Δx can subsequently be determined. The generic steps for determination of CS,0 are: 1. Determine the water-soluble chloride content of the surrounding media (e.g., soil, water, air). 2. Develop the chloride-adsorption-isotherm (CAI) based on the specific concrete mix that will be specified for construction. An example CAI is shown in Figure 56. 3. From the CAI, determine the total chloride content corresponding to the chloride content determined in step 1. This chloride content is the value of chloride saturation concentration, Cs,0, that will be used as the environmental load in the fully probabilistic approach.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 152 Source: fib (2006) Figure 56. Surface chloride concentration CS,0 in dependency on Ceqv for a Portland cement concrete [Guide Specification Figure CA.2.3.3-3]. Given the discussion above, a surface chloride saturation, Cs,0 = 0.60 with a standard deviation of 0.30 and assuming a lognormal distribution were used to calculate the calibrated cover values for buried components. This value was developed by analysis that leads to a service life of 75 years for 3-in cover for concrete cast against earth as per Table 5.10.1-1 of AASHTO (2017a). Based on the CAI in Figure 56, which is for OPC concrete mix with cement content that fits a Class B concrete which is typically used for foundations as per AASHTO (2017a), a limiting chloride content of the surrounding geomaterial (“ambient solution”) was estimated to be approximately 1.8 g/L (1,800 ppm) corresponding to a Cs,0 of 0.60. For the nonaggressive environment, a less rigorous approach was taken. Cover values were calibrated by setting CS such that the calibrated cover values approximately matched the corresponding cover provisions in Section 5 of AASHTO LRFD for the normal service life category. It is anticipated that the calibrated cover provisions will be updated when data for Cs in nonaggressive environments becomes available. Input Parameters – Material Properties The following material properties have to be defined as shown in Table 67 and Table 68: the concrete chloride migration coefficient (DRCM), the age factor (α), the initial chloride concentration (C0), critical chloride concentration (Ccrit), and cover thickness (c). Guidance from fib Bulletin 34, fib Bulletin 76, and past project experience were used to define the input parameters. The following concrete mixes were considered: • OPC: Ordinary Portland Cement (ASTM C150 Type I or Type I-II) • OPCFA: Ordinary Portland Cement with 20%-50% fly ash Type F by mass of total cementitious • OPCFA+SF: Ordinary Portland Cement with 20%-50% fly ash Type F by mass of total cementitious and with 5%-8% silica fume by mass of total cementitious

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 153 • GGBS: Ordinary Portland Cement with 36%-65% ground granulated blastfurnace slag grade 100 or higher by mass of total cementitious • GGBS+SF: Ordinary Portland Cement with 36%-65% ground granulated blastfurnace slag grade 100 or higher by mass of total cementitious and with 5%-8% silica fume by mass of total cementitious. Table 67. Input Parameters for Materials. Input Parameter Symbol Units distribution mean, μ stdev, σ a b Chloride Migration Coefficient DRCM,0 x10-9 in2/s OPC Normal 19.38 3.88 -1 - OPCFA 12.4 2.48 - - OPCFA+SF 7.29 1.46 - - GGBS 7.75 1.55 - - GGBS+SF 3.57 0.71 - - Aging Exponent α - - Beta Table 6 Initial Chloride Content C0 wt.- %/CM - Constant = 0.1 Critical Chloride Concentration Ccrit wt.- %/CM Class A Beta 0.6 0.15 0.2 2.0 Class B 1.5 0.15 1.1 2.9 Class C 3.0 0.15 2.6 4.4 Class D 6.0 0.15 5.6 7.4 Cover c in - Normal 1 to 4 0.3 - - Notes: 1 indicates not applicable Table 68. Aging Exponent. Exposure Zone Concrete Mix distribution mean, μ stdev, σ a b Atmospheric OPC Beta 0.65 0.15 0 1 OPCFA 0.65 0.15 0 1 OPCFA+SF 0.65 0.15 0 1 GGBS 0.65 0.15 0 2 GGBS+SF 0.65 0.15 0 2 Splash / Spray / Submerged / Buried OPC Beta 0.30 0.12 0 1 OPCFA 0.60 0.15 0 1 OPCFA+SF 0.60 0.15 0 1 GGBS 0.40 0.15 0 2 GGBS+SF 0.40 0.15 0 2 The chloride migration coefficient is a measure of the diffusion rate through concrete and is controlled by the pore structure of the concrete (fib 2015). The chloride migration coefficient is measured using the test method defined in NT Build 492 (Nordtest1999), typically at 28 days. Service life calculations and resulting selection of cementitious material combinations used in the calibration are based on chloride migration coefficients deemed technically achievable based on past experience and guidance from fib Bulletin 34 and Bulletin 76. Maximum allowable mean chloride migration coefficients assumed in the design table for each type of concrete mix are shown in Table 67. A constant coefficient of variation of 0.2 was assumed for all concrete mixes based on fib Bulletin 34 and Bulletin 76 guidance, resulting in the standard deviations presented in Table 67.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 154 The aging exponent is a critical material property used in the chloride diffusion calculations and represents the decrease in the apparent diffusion coefficient over time (fib 2015). It is dependent on both the concrete mix and the environmental exposure conditions, as shown in Table 68. Guidance from fib Bulletin 34 and Bulletin 76, along with past project experience were used to select values for the aging exponent. The initial chloride content of the concrete, C0, is determined by the chloride content of the aggregate, cementitious materials, mix water, and admixtures. Many specifications place limits on the maximum permissible initial chloride content (ACI 2014, AASHTO 2017b). For the purpose of the calibrations, a constant value of 0.1 wt.-%/CM was assumed for C0. Four classes of reinforcement were considered in the calibration: • Class A – representative of uncoated, typical reinforcing steel (i.e., black bar). • Class B – reinforcement with improved corrosion resistance, which might be obtained by defect-free epoxy coating or improved steel material formulations (e.g., galvanized). • Class C – reinforcement with higher corrosion resistance than Class B, but not to the extent of Class D. • Class D – highly corrosion resistant materials (e.g., Type 316LN stainless steel). Class A (black) and Class D (stainless) bars represent a lower and upper bound for the critical chloride concentration, Ccrit, respectively. The two intermediate reinforcement classes with Ccrit values between that of Class A and Class D bars were included to represent the variety of other reinforcement types available to designers. The Ccrit, value in Table 67 for Class A reinforcing bars is based on guidance in fib Bulletin 34 and Bulletin 76 for black bars. For the other three reinforcement classes shown in Table 67, the mean values of Ccrit were set using multiples of the Class A mean Ccrit. The following multiples were assumed: • Class B: 2.5 times Ccrit of black bar • Class C: 5 times Ccrit of black bar • Class D: 10 times Ccrit of black bar It is recognized that there are a wide variety of reinforcement types used for reinforced concrete construction, each with various resistances to corrosion. Even for a typical reinforcing steel (i.e., black bar), reported Ccrit values are not consistent. This is attributed to the large number of factors that influence the critical chloride threshold, including the conditions of the steel-concrete interface, properties of the concrete, and the conditions of the surrounding environment (Bertolini et al. 2004, Böhni 2005). The various influences on the critical chloride threshold are listed in Table 69. In order to design with the reinforcement classes in Table 67, particularly those other than Class A, the Ccrit value of the class of reinforcement used in construction will have to meet or exceed the associated value in Table 67. Alternatively, language is included in the Guide Specification that allows owners to set the classification for a particular reinforcement type based on their experience. The intent of this classification approach is to allow the specifications to accommodate a wide variety of reinforcement types in spite of the very limited data available on chloride threshold values. Further discussion on the need for research in this subject is discussed in Chapter 5 of this report. As a consequence, owner input may be required to classify reinforcement other than black bar.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 155 Table 69. Influences on the Critical Chloride Threshold for Mild Steel Reinforcement (after Glass and Buenfeld 1995, Böhni 2005). Influence Effect1 Evidence Influence Effect1 Evidence Theoretical Concrete Cover Cl- binding ↑ Hypothetical Curing ‒ Experimental Cl- mobility ↓ Hypothetical w/c-ratio ↓ Experimental Oxidizing conditions ↓ Hypothetical Cement content ↑ Experimental Steel Interface Cover depth ↑ Experimental Voids ↓↓ Experimental External factors Pre-rusting ↓ Experimental Moisture at low level ↓ Experimental Steel composition Experimental Moisture at high level ↑ Hypothetical Half cell potential ↓ Experimental Moisture variations ↓ Hypothetical Surface roughness Experimental Oxygen concentration ↓ Hypothetical Binder External chloride source ↓ Hypothetical C3A content ↑ Hypothetical Chloride type ‒ Hypothetical pH ↑ Hypothetical Temperature ↓ Hypothetical FA ↓ Experimental GGBS ↓ Experimental SF ↓ Experimental Notes: 1 indicates no change in threshold with increase in influence, ↑ indicates increase in threshold with increase in influence, ↓ indicates decrease in threshold with increase in influence Cover dimensions were calibrated with the assumption that the minimum cover is 1 inch and the maximum permissible cover is 4 inches. Covers less than 1 inch are difficult to achieve from a construction tolerance perspective, while covers greater than 4 inches become impractical. For the calibrated covers presented in the Guide Specification, the cover is measured from the exterior face of the member to the outermost layer of reinforcement. This is a deviation from the cover provisions in AASHTO LRFD, in which cover is measured to the main reinforcement. For the purpose of the calibration, a normal distribution was assumed for the cover. The minimum cover value of 1 inch was sufficient to avoid a significant probability of the cover taking a negative value. The mean and standard deviation of the normal distribution were defined using the approach outlined in fib Bulletin 76, in which the nominal cover specified in design specifications cnom is the sum of the minimum cover cmin and an allowance for deviation cdev, as follows: minnom devc c c= + ∆ (30) Setting the mean cover equal to the nominal cover and taking the minimum cover as the 5% quantile of the normal distribution, the standard deviation can be defined in terms of Δcdev. 0.05 0.05U uµ σ= + ⋅ (31) where: U0.05 = 5% quantile of the normally distributed cover = cmin μ = mean value of the cover = cnom σ = standard deviation of the cover u0.05 = 5% quantile of the normal distribution = -1.64

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 156 Solving Equation 31 for the standard deviation and combining it with Equation 30 results in: min 0.05 1.64 nom devc c c u σ − ∆ = = (32) Equation 32 was used to define the standard deviation used in the calibration. The allowance for deviation cdev represents the construction tolerance and was taken as +/- 0.5 inch based provisions in AASHTO (2017b) and ACI (2010). Model Limitations The calibrated design provisions were developed to be applicable to most typical highway structures in the United States. A detailed assessment performed for a specific structure (i.e., materials) in a specific location (i.e., environmental exposure) will produce a refined design. The fib Bulletin 34 model (Equation 26) does not explicitly consider the influence of crack widths on the initiation of corrosion in reinforced concrete, but fib does recognize that the crack width should be kept below a characteristic value in order to ensure a sufficiently long service life (fib 2006, fib 2010). While a cracked structure may perform satisfactorily if the width and frequency of the cracks are within acceptable limits, research on the effect of cracking on chloride-induced corrosion, although extensive, shows significant differences and conflicts in the findings (Matthews 2014). This is probably due to the fact that multiple factors related to cracking influence corrosion initiation and propagation such as crack width, depth, frequency, orientation, and nature (active or dormant). Despite the lack of quantitative consistency on the effect of cracking on diffusivity, which makes mathematical specification premature, there is sufficient evidence that the increase in diffusivity can be significant (Djerbi et al. 2008; Balakumaran et al. 2018) depending on the width, depth, and plan extent of cracking. In addition, research has shown that the effect of cracking is more important for mixes with low diffusion coefficients due to the increase in the ratio of cracked to uncracked diffusion coefficient in comparison to OPC mixes (Djerbi et al. 2008). At this point in time the most practical course of action is requiring materials and construction procedures that local experience has shown to be effective in reducing cracking. Calibrated Design Provisions – Concrete Cover Using the methodology above, design alternatives for combinations of concrete cover, concrete mix design, and type of reinforcement were established as a function of the exposure conditions and the service life category. These design alternatives are presented in the form of a design table in Section 4 of the Guide Specification, replicated in Table 70. Cover dimensions are grouped by exposure class, concrete mix, and reinforcement class. The cover dimensions presented in Table 70 were initially calculated based on a 10% probability of failure (reliability index β = 1.3). For practical purposes, these values were rounded up to the next half inch unless the minimum cover (i.e., 1 inch) was sufficient to achieve a probability of failure less than 10%, in which case the minimum cover was used. See Appendix E for unrounded cover dimensions and associated reliability indices. Cover dimensions greater than the maximum cover (i.e., 4 inches) were excluded from the table. Where the concrete mixes previously defined in this report (see “Input Parameters – Material Properties”) all met the probability of failure criteria with the minimum cover dimension for the same exposure class, reinforcement class, and service life category, the table rows for those concrete mixes were condensed into one row and the concrete type was changed to “any”. In this sense, “any” means any of the five concrete mixes (OPC, OPCFA, OPCFA+SF, GGBS, GGBS+SF). Similarly, when multiple concrete mixes produced identical cover dimensions for the same exposure class, reinforcement class, and service life categories, the rows of the table for those mixes were combined to reduce the size of the table.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 157 While values for the target service life integers are not presented in Table 70 or in the identical table in the Guide Specification, it was necessary to select values in order to perform the calculations. The following assumptions were made for each service life category: • Normal = 75 years. This was chosen to match the probabilistically calibrated design life value in the AASHTO LRFD Bridge Design Specifications (2017a). • Enhanced = 100 years. This is a reasonable middle ground between the normal and maximum categories. • Maximum = 150 years. This is at the practical limit, and possibly beyond, that the service life of a bridge can be projected to. In addition, the societal needs for bridges will most likely change significantly within this length of time, causing replacement to be driven by functionality rather than deterioration.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 158 Table 70. Calibrated Design Provisions for the Limit State of Chloride-Induced Corrosion [Guide Specification Table 4.2.4.2.2-1]. Exposure Class Concrete Type Reinforcement Class Service Life Category Normal Enhanced Maximum Class Condition Description Examples Cover (in)1,2 C-D1 Atmospheric in a deicing salts environment Surfaces exposed to airborne chlorides Underside of Deck (Interior Bays) Any D 1.0 1.0 1.0 Any C 1.0 1.0 1.0 Any B 1.0 1.0 1.0 OPCFA+SF, GGBS+SF, GGBS A 1.0 1.0 1.0 OPCFA, OPC A 1.5 1.5 1.5 All Other Surfaces (abutments, piers, pile caps, girders, underside deck overhang) Any D 1.0 1.0 1.0 Any C 1.0 1.0 1.0 Any B 1.0 1.0 1.0 OPCFA+SF, GGBS A 1.5 1.5 1.5 OPCFA A 2.0 2.0 2.0 GGBS+SF A 1.0 1.0 1.5 OPC A 2.0 2.0 2.5 C-D2 Indirect Deicing Salts Surfaces exposed or potentially exposed to drainage water containing deicing salts or roadway spray Surfaces below expansion joints or drains, deck fascia, substructure surfaces near a roadway Any D 1.0 1.0 1.0 Any C 1.0 1.0 1.0 OPCFA+SF, OPCFA, GGBS+SF B 1.5 1.5 1.5 GGBS B 2.0 2.0 2.0 OPC B 3.0 3.5 3.5 OPCFA+SF A 2.0 2.0 2.5 OPCFA A 2.5 2.5 2.5 GGBS+SF A 2.5 2.5 3.0 GGBS A 3.0 3.5 4.0 C-D3 Direct Deicing Salts - Low Direct exposure to deicing salts with low to medium application rate of deicing salts Top of decks, curbs, sidewalks, barriers Any D 1.0 1.0 1.0 OPCFA+SF C 1.0 1.0 1.0 OPCFA, GGBS+SF C 1.0 1.5 1.5 GGBS C 1.5 1.5 1.5 OPC C 2.0 2.5 2.5 OPCFA+SF B 1.5 1.5 2.0 OPCFA, GGBS+SF B 2.0 2.0 2.0 GGBS B 2.5 2.5 3.0 OPC B 4.0 4.5 5.0 OPCFA+SF A 2.5 2.5 2.5 OPCFA A 2.5 3.0 3.0 GGBS+SF A 3.0 3.0 3.5 GGBS A 3.5 4.0 4.5 C-D4 Direct Deicing Salts - High Direct exposure to deicing salts with high application rate of deicing salts Top of decks, curbs, sidewalks, barriers Any D 1.5 1.5 1.5 OPCFA+SF, OPCFA, GGBS+SF C 1.5 1.5 1.5 GGBS C 2.0 2.0 2.0 OPC C 3.0 3.0 3.5 OPCFA+SF B 2.0 2.0 2.0 OPCFA B 2.0 2.0 2.5 GGBS+SF B 2.0 2.5 2.5 GGBS B 3.0 3.0 3.5 OPCFA+SF A 2.5 2.5 3.0 OPCFA, GGBS+SF A 3.0 3.0 3.5 GGBS A 4.0 4.5 5.0 Notes: 1Nominal minimum cover measured to the outermost layer of reinforcement. 2Cells shown blacked out are for design combinations that require cover greater than 4 inches, which is not permitted in the Guide Specification.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 159 Table 70. Calibrated design provisions for the limit state of chloride-induced corrosion (continued) [Guide Specification Table 4.2.4.2.2-1]. Exposure Class Concrete Type Reinforcement Class Service Life Category Normal Enhanced Maximum Class Condition Description Examples Cover (in)1,2 C-M1 Marine - Atmospheric Surfaces exposed to airborne chlorides Superstructure Any D 1.0 1.0 1.0 Any C 1.0 1.0 1.0 OPCFA+SF, OPCFA, GGBS+SF, GGBS B 1.0 1.0 1.0 OPC B 1.5 1.5 1.5 OPCFA+SF, GGBS A 1.5 2.0 2.0 OPCFA A 2.0 2.5 2.5 GGBS+SF A 1.5 1.5 1.5 OPC A 2.5 2.5 3.0 C-M2 Marine - Submerged Surfaces permanently submerged with salt water present Substructures between mudline and tidal zone Any D 1.0 1.0 1.0 OPCFA+SF, OPCFA, GGBS+SF C 1.0 1.0 1.0 GGBS C 1.0 1.5 1.5 OPC C 2.0 2.5 2.5 OPCFA+SF B 1.5 1.5 1.5 OPCFA B 1.5 2.0 2.0 GGBS+SF B 2.0 2.0 2.0 GGBS B 2.5 2.5 3.0 OPCFA+SF A 2.5 2.5 2.5 OPCFA, GGBS+SF A 3.0 3.0 3.5 GGBS A 4.0 4.5 5.0 C-M3 Marine - Tidal or Splash/Spray Zone Surfaces in contact with salt water either in the tidal zone or splash/spray zone Substructures within tidal zone or splash/spray zone Any D 1.0 1.0 1.0 OPCFA+SF C 1.5 1.5 1.5 OPCFA C 1.5 1.5 2.0 GGBS+SF C 1.5 2.0 2.0 GGBS C 2.0 2.0 2.5 OPC C 3.5 4.0 4.5 OPCFA+SF B 2.0 2.0 2.5 OPCFA, GGBS+SF B 2.5 2.5 3.0 GGBS B 3.5 3.5 4.0 OPCFA+SF A 3.0 3.0 3.5 OPCFA A 3.5 4.0 4.0 GGBS+SF A 3.5 4.0 4.5 C-B Buried Buried elements below finished grade or below mudline (in the case of submergence) Deep and shallow foundations, retaining structures, etc. Any D 1.0 1.0 1.0 Any C 1.0 1.0 1.0 Any B 1.0 1.0 1.0 OPCFA+SF, OPCFA A 1.0 1.0 1.0 GGBS+SF A 1.0 1.0 1.5 GGBS A 1.5 1.5 1.5 OPC A 2.5 3.0 3.5 Notes: 1Nominal minimum cover measured to the outermost layer of reinforcement. 2Cells shown blacked out are for design combinations that require cover greater than 4 inches, which is not permitted in the Guide Specification.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 160 Table 70. Calibrated design provisions for the limit state of chloride-induced corrosion (continued) [Guide Specification Table 4.2.4.2.2-1]. Exposure Class Concrete Type2 Reinforcement Class3 Service Life Category Normal Enhanced Maximum Class Condition Description Examples Cover (in)1,4 C-NA1 Interior exposure Sheltered surfaces, mostly dry conditions Any D 1.0 1.0 1.0 Any C 1.0 1.0 1.0 Any B 1.0 1.0 1.0 OPCFA+SF, OPCFA, GGBS+SF, GGBS A 1.0 1.0 1.0 OPC A 1.0 1.0 1.5 C-NA2 Other exterior exposure Any D 1.0 1.0 1.0 Any C 1.0 1.0 1.0 Any B 1.0 1.0 1.0 OPCFA+SF, GGBS A 1.0 1.5 1.5 OPCFA A 1.5 1.5 1.5 GGBS+SF A 1.0 1.0 1.0 OPC A 1.5 2.0 2.0 Notes: 1Nominal minimum cover measured to the outermost layer of reinforcement. 2Cells shown blacked out are for design combinations that require cover greater than 4 inches, which is not permitted in the Guide Specification. Calibrated Design Provisions – Points System A scoring system was previously developed during the Phase II research and presented in Interim Report No. 2 in which a minimum point value was set for the normal, enhanced, and maximum service life categories as a function of exposure class. The scoring system was developed using the same methodology that was used to create Table 70. Point values were given to different concrete mixes, reinforcement classes, and cover dimensions. Summation of the point values associated with concrete mix, reinforcement class, and cover, produced different combinations that could be used to meet the minimum point requirements. During the Phase III research, which included expansion of Table 70 to include additional reinforcement classes and concrete mixes, it became evident that the scoring system was no longer practical. The number of different design combinations in Table 70 and the independent behavior of each, in terms of service life modeling, prevented the provisions from being condensed into a simplified points system. Therefore, the focus of the corrosion provisions in Section 4 of the Guide Specification was placed solely on the calibrated cover dimensions from Table 70. Other Protection Strategies An article discussing other available protection strategies in addition to the previously presented material design provisions is included in Section 4. These other strategies include surface treatments (e.g., sealers), special provisions for decks (e.g., overlays and membranes), and detailing for durability. Due to the number of different types of surface treatments and deck overlay systems and a lack of data on their long-term performance, detailed design provisions (in terms of anticipated service life and relative performance of different treatments) are not provided. However, it is recognized that these protection strategies can be beneficial to service life and therefore general guidance on their use is provided. Provisions allowing wearing surface thickness to be included in the design cover dimension are provided. For decks, the importance of proper wet curing is stressed, with recommended practices for placement and length of curing included in the article. Provisions on preferred-detailing for service life include guidance on details for reinforcement, post-tensioning, and drainage. Several example details are included to provide the designer with additional guidance.

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 161 Steel Structures The section on steel structures, Section 5, leads with a discussion of the two most common deterioration mechanisms for steel bridge components: corrosion and fatigue. This is followed by an article on the protective measures for each mechanism. For corrosion, a table is presented that allows the designer to choose a steel protection strategy, in terms of steel type and coated or uncoated, depending on the exposure class and the desired service life category. The table was developed consistent with FHWA (2015) and Helsel and Lanterman (2018), as well as the four Environmental Exposure Classes from ISO (2012) selected for use in FHWA (2015). The suggested applications found within each Environmental Exposure Class were selected based on the least project cost to accomplish the level of protection required. Forewarnings are provided that any coating strategy employed assumes a certain level of maintenance and reapplication. Expected service lives of various coating systems are provided based on the latest work of Helsel and Lanterman (2018), which is included in the form of an appendix to Section 5. Considerations when employing weathering steel as a protection strategy are presented, as it is not suitable in certain environments and locations. For fatigue, the designer is directed to AASHTO LRFD with the caveat that the fatigue design provisions within AASHTO LRFD are based on an assumed design life of 75 years. When designing for a service life greater than 75 years, guidance exists within AASHTO LRFD on how to make adjustments to the fatigue design provisions. The final article in Section 5 provides detailing practices that should be followed in order to prevent corrosion and fatigue. Example details for corrosion prevention are provided to the designer as a guide. Foundations and Retaining Walls Section 6 of the Guide Specification addresses the service life design of foundations and retaining walls. Design methods that consider durability or service life have been developed for many foundation and retaining wall components, and are in general use. These methods have been developed on a component basis, often by the industries that produce or promote those components. As such, the established practice is not consistent across all components. In the development of the foundations and retaining walls section of the proposed Guide Specification, importance has been placed on incorporating past efforts and practices to the greatest extent possible. A system based on a protection index, PI, has been developed in order to achieve this goal. Rather than a “good-better-best” approach, or defining specific target service lives, the protection index approach divides design into three categories based on a number of factors, and the mitigation of corrosion processes are scaled appropriately. This is broadly in keeping with the three level approach of the proposed Guide Specification, and is effective at stitching together disparate requirements in a way that leverages existing practice effectively with a minimum of disruption. The first article establishes the protection index, as follows, based on the environment, consequences of adverse performance, location, and the type of facility carried: • PI = 0 for structures in nonaggressive environments, where the consequences of loss of serviceability are not serious, or for structures with low traffic volumes such as local and rural routes. • PI = 2 for structures in an aggressive environment, where consequences of loss of serviceability are serious, or for routes with high traffic volumes such as interstates and freeways. • PI = 1 for cases between PI = 0 and PI = 2 as determined by the owner in terms of the aggressiveness of the environment, the consequences of loss of serviceability, or routes with intermediate traffic volumes such as arterials, service roads, and state-owned routes. The remainder of Section 6 consist of design provisions for different types of foundation and retaining wall elements. For each element, considerations for the deterioration environment are given followed by design provisions to prevent said deterioration. The designer is directed to other sections of the Guide

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 162 Specification for additional guidance where appropriate, such as the provisions of Section 4 when designing a concrete foundation element. The following elements are included: • Spread footings • Driven piles • Micropiles • Drilled shafts • Nongravity cantilever walls • Anchored walls • Soil nails • MSE walls A brief discussion on durability considerations for approach embankments is also included at the end of Section 6. Renewable Elements The section on renewable elements, Section 7, includes design guidance for joints, bearings, railings and barriers, and roadway approaches. Different types of joint systems are presented along with qualitative information on their protection effectiveness, longevity potential, vulnerability to damage, initial cost, likeliness of being replaced, difficulty of replacement, and possible failure modes. An emphasis is placed on joint elimination through the use of jointless systems, as this is the preferred alternative over any joint system. For bearings, the different types and associated capabilities are presented, in terms of capacity ranges for load, translation, and rotation as well as relative cost information. Ranges of expected service life for different bearings types are given based on the work of Azizinamini et al. (2014); however, specific service life values are not presented due to a lack of data. Designing for replacement of both joints and bearings is included in Section 7. General service life design provisions for railings, barriers, and approach slabs are also provided. Life Cycle Cost Analysis Section 8, Life Cycle Cost Analysis (LCCA), provides a brief overview of LCCA as it applies to bridge design. The section is based on the guidance provided in NCHRP Report 483 (Hawk 2003). The following steps of a LCCA process are defined, and the subsequent articles of the section follow this process: • Define Design Alternatives • Set Planning Horizon • Schedule of Costs • Estimate of Costs • Net Present Value • Compare Alternatives The main types of costs associated with bridge LCCA, Agency and User Costs, are defined and common sources of each are presented. Relevant formulas for calculating present values for different expenditure types are included. Probabilistic Service life Design Framework A framework that explains the use of the probabilistic approach to service life design for chloride-induced corrosion of reinforced concrete is included in Appendix A of the Guide Specification. The background information on the probabilistic method from fib Bulletin 34 utilizing Fick’s Second Law is presented, as

NCHRP Web-Only Document 269: Guide Specification for Service Life Design of Highway Bridges 163 well as guidance on how to apply the method in design. This section focuses on the implementation of the probabilistic method for corrosion on a project specific level, such that a designer could use different assumptions than those presented in Section 4 (i.e., concrete mix type, cover dimensions, and type of reinforcement) and still produce an adequate design for service life. Guidance on how to define each of the input parameters in terms of the type of distribution and associated values is included. Concrete testing information is given, so that a State or owner is directed to the proper tests used to determine the necessary chloride diffusion parameters for their proprietary mix designs. The designer is directed to sources for additional information, often fib publications. Case Studies A series of case studies in the form of design examples that illustrate the implementation of the Guide Specification were developed. The goals of these design examples are to provide the following: 1. Detailed documentation of how the Guide Specification are intended to be implemented. 2. Documentation and summary of the differences that result from an implementation of the Guide Specification within the bridge design process. 3. Comparative analysis of how various locations (and their associated exposure classes) influence the design of a bridge. To accomplish this, the Research Team selected two bridges recently designed as per AASHTO LRFD (and the associated State bridge design specifications). Due to their prevalence in practice, one of the bridges is a steel multi-girder bridge and the other is a pre-stressed concrete multi-girder bridge. The design of each bridge is compared to the developed Guide Specification. All modifications to the original design resulting from the provisions included within the new Guide Specification are documented and summarized. In addition, a detailed documentation of the design process with direct references to all relevant provisions and design considerations is provided. Finally, the bridges are “relocated” in order to document how various exposure classes influence the design. For each bridge, two additional locations are examined, one milder and one harsher than the actual location of the bridge. At the completion of each case study, a table summarizing all of the differences that resulted from the various exposure classes is presented.