Proposed AASHTO Guides for Bridge Preservation Actions (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.

19 The research team collected data from state DOTs on the use, performance, and costs of preservation actions. Datasets were collected from the state DOTs of California, Connecticut, Illinois, Michigan, New York, Ohio, Oregon, Pennsylvania, Wisconsin, Virginia, and Washington. Information collected from these state DOTs include the following: • Unit costs of actions for bridges • Records of completed bridge maintenance and repair work • Lettings bridge maintenance projects • Lists of bridge maintenance recommendations • Deck treatments reported in field trials Information on Bridge Actions from State DOTs The research team also collected information from 11 state DOTs on preservation actions, costs, performance, and records of actions or projects at bridges (see Table 16). State DOT Bridge Preservation Data Data are obtained in various forms. State DOT Advice State DOT advice on unit cost is cost computed by staff and used for in-house estimates of bridge project costs. Advice on unit cost was collected from the state DOTs of California, Michigan, New York, Pennsylvania, and Virginia. California unit costs were provided by California DOT SM&I. Michigan DOT prepared spreadsheets with unit costs for capital scheduled maintenance and capital preventive maintenance. New York State DOT provided unit costs for maintenance, repair, and replacement of AASHTO National Bridge Elements (NBEs) and AASHTO Bridge Management Elements (BMEs). Virginia provided loaded unit costs for work in bridge maintenance and repair. State DOT Average Bid Cost Average bid item costs were collected from the state DOTs of Connecticut and Ohio. Connecticut DOT bid item costs are from contracts for rentals of bridge repair units. Contracts are statewide, multiyear agreements for as-needed bridge repair and maintenance. Ohio DOT actions for joints, deck overlays, bridge painting, and bridge repairs are items in Ohio’s C H A P T E R 4 Context, Performance, and Cost of Bridge Preservation Actions

20 Proposed AASHTO Guides for Bridge Preservation Actions standard specification. Ohio DOT bid items are for completed work—all necessary materials, equipment, and labor. State DOT Historical Records Historical records include maintenance accomplishments and let projects. Historical records were collected from the state DOTs of California, Illinois, Michigan, New York, Ohio, Oregon, Pennsylvania, Washington, and Wisconsin. California DOT provided a database of mainte- nance and repair recommendations. Illinois DOT provided a list of decks, overlay types, and years of installation. Michigan DOT provided a list of let projects for bridges. Ohio DOT provided maintenance records of their crews. Oregon DOT provided a list of projects in its major bridge maintenance program. Pennsylvania DOT provided records of completed work from its bridge management system. Washington DOT provided a database of decks, deck treat- ments, and years of application of treatments. Wisconsin DOT provided records of projects from its highway structures information system (HSI). State DOT Data for Deck Treatments The research team collected reports published by state DOTs on field trials of treatments of bridge decks in service. Research studies that include field trials of treatments usually address the condition of decks achieved by treatment plus deck conditions during the early years of service after treatment. Longer, continuing studies of conditions after treatment are uncommon. Many decks and treatments from field trials remain in service today. The team took existing field trials and attached deck condition ratings from the year of treatment to the present day to extend previous trials into longer-term studies. State Source Year Range Action History Cost Actual Average California Structure Maintenance and Investigations (SM&I) database (12) 1961 to 2016 Y Y Y Connecticut Maintenance contracts (13, 14, 15, 16) 2005 to 2015 Y Illinois Deck database (26) 1976 to 2016 Y Michigan Ledger of bridge construction and repair projects (33), and bridge scoping cost estimate worksheets (32) 1867 to 2016 Y Y New York Element-level unit costs of repair, rehabilitation,and replacement (35); maintenance lettings (34) 2007 to 2016 Y Y Y Ohio Bid item average costsActions and costs for in-house bridge maintenance 1998 to 2014 Y Y Y Oregon Projects in major bridge maintenance (45) 2011 to 2016 Y Y Pennsylvania Unit costs of flex actions for bridges, Bridge Management System 2 (BMS2) database of actions at bridges (47) 1901 to 2016 Y Y Y Virginia Loaded cost calculation for bridge preservation work (61) 2016 Y Washington Deck database (62) 1979 to 2016 Y Wisconsin Project in highway structures information system (HSI) (64) 1994 to 2016 Y Y Y = Yes, information found. Table 16. Information on bridge actions from state DOTs.

Context, Performance, and Cost of Bridge Preservation Actions 21 Use of State Data on Bridge Actions Context of Bridge Actions Bridge owners apply actions when bridges reach a certain condition, age, or both. Bridge condition and bridge age together are the context for an action. Condition and age of bridges were determined by linking state DOT historical records of actions to NBI records. NBI records for the years 1992 through 2016 were used. Bridge age at action is the difference between the year of action and the year of bridge construction or reconstruction, whichever is later. Bridge condition at action is the GCR in the last inspection before the year of action. Both age and condition are averaged for all bridges having the same action/component/detail in historical records for a state DOT. Matching Preservation Actions to NBI Records To match bridges in state datasets to NBI records, NBI structure numbers (SNs) were used. For bridges in most states, the SN is sufficient to reconstruct condition histories using annual update files of the NBI. A few states have renumbered their bridges. Links from prior SNs to current SNs were made for these states and bridges. A combination of other NBI items, including year built, route number, and milepost, was used to link prior and current SNs. Missing SN in Report Some research reports of state DOT field trials do not include NBI SNs. Research reports often name the route carried and the feature intersected. These were used to search among NBI records to determine SNs. The search succeeded for about one out of every two bridges. Many searches using route carried and feature crossed yielded two or more bridges. If the source research report offered additional information, such as route direction or year built, a unique match was found. Performance of Actions The performance of a preservation action is quantified as the action’s effect and the action’s interval. The effect of an action is the change in condition of a bridge, bridge component, or bridge element as a result of the action. The interval of an action is the time period until a bridge, bridge component, or bridge element returns to the condition prior to the action. A history of GCRs for a bridge component with a preservation action is shown in Figure 1. A bridge component has a GCR before preservation action, GCRpre. The bridge component exhibits an increase in GCR as a result of the preservation action followed by a monotonic decrease in condition rating over time. GCRs are integer values, so a monotonic decrease is stepwise. Effect of the Action The effect of the action is the change in GCR. Eq. 4GCR GCR GCReffect post pre= − where: GCReffect is the change in GCR due to the action. GCRpost is the GCR after the action. GCRpre is the GCR before the action.

22 Proposed AASHTO Guides for Bridge Preservation Actions Time Intervals in Performance Intervals in time are collected for GCRs in time compared with GCR immediately after treat- ment. Intervals correspond to change (decrease) in GCR, dGCR. The decrease is as follows: Eq. 5dGCR GCR GCR tpost ( )= − where: GCR(t) is the GCR at year, t, after the action. Intervals in time are collected for the following: t(0), time interval for dGCR = 0. t(1), time interval for dGCR = 1. t(2), time interval for dGCR = 2. Condition histories for individual actions may not exhibit all time intervals. Histories may end before some time intervals are observed. Rates of Change of General Condition Rating The average rate of change of GCR is computed after preservation action. The rate is computed from the action to the end of the relevant condition history. The end of the relevant condition history may be due to the end of NBI data (the year 2016 for NCHRP Project 14-36), or the absence of the bridge’s SN from later NBI update files, or improvement to GCR in later years, which may be due to other preservation actions. Eq. 6Det Rate GCR GCR T post end end ( )= − where: Det Rate is the time rate of change in GCR. Tend is the time in years from an action to the end of the relevant condition history. Action Interval and Service Life For each instance of an action, the interval for the action and the service life after the action are computed. The interval for an action is the time interval to return to the condition prior to the action. Figure 1. Condition history of bridge component with preservation action.

Context, Performance, and Cost of Bridge Preservation Actions 23 Eq. 7 Action Interval GCR GCR Det Rate Action Interval GCR Det Rate post pre effect ( ) ( ) = − = Service life is the interval in time from the action to poor condition. Eq. 8Service Life GCR GCR Det Rate post poor( )= − Notes on Use of NBI General Condition Ratings The Deck GCR was used for actions affecting deck, railing, and joints. The Superstructure GCR was used for actions affecting superstructure and bearings. The Substructure GCR was used for actions affecting substructures and approaches. The Channel GCR was used for actions related to scour. The research team used NBI update files posted by FHWA for the years 1992 through 2016— a range of 25 years. This range constrains the measures of performance of actions. • GCReffect can be computed when both GCRpre and GCRpost are found. Most actions applied between the years 1993 and 2015 were evaluated for GCReffect. For actions in earlier or later years, GCReffect could not be computed. • Intervals for persistence of condition, t(r), can be evaluated if GCRpost was found. Actions applied in the year 1992 or later were evaluated. • Intervals for t(r) were usually observed for r values equal to 0, less often observed for r values equal to 1, and not often observed for r values equal to 2. Greater values of r require longer histories of condition ratings. • Det Rate was computed for condition histories that exhibited a decrease in GCR after action. • Action Interval and Service Life were computed for actions that yielded a Det Rate. Computing Measures of Performance of Preservation Actions Performance was computed for each instance of an action (i.e., each bridge that has a known action at a known date and has been matched to a history of NBI GCRs). Each instance of an action yielded a set of measures of performance. Average values of performance measures were computed for populations of actions. For GCRpre, GCRpost, GCReffect, and Det Rate, arith- metic averages were reported. For time intervals t(r), mean values of Weibull distributions were reported. Evaluations of context, performance, and cost of preservation actions using state DOT data are listed in Table 24, Table 25, and Table 26, respectively, of Appendix C. State DOTs and Data on Preservation Actions California DOT Data California DOT provided unit costs used to estimate costs of maintenance for bridges (12). Actions include cleaning, coating, element repair, and element replacement. Costs of actions differ by type of delivery. Actions are delivered by California DOT crews, by maintenance con- tractors, or through California’s program for Structure Replacement and Improvement Needs (STRAIN). The California database has estimated total costs for actions completed at bridges.

24 Proposed AASHTO Guides for Bridge Preservation Actions Connecticut DOT Connecticut DOT provided copies of Bridge Rental Repair Unit (BRRU) contracts that are used in maintenance of bridges. Contracts are indefinite delivery, indefinite quantity. Illinois DOT Illinois DOT provided a database of state-owned bridges in District 1 listing deck overlays and year of application of overlay. The database has 303 reinforced concrete bridge decks. Overlays are latex-modified concrete, microsilica concrete, plasticized dense concrete, and PCC. The context and performance of decks overlays were computed. Michigan DOT Michigan DOT provided a list of estimated costs for actions in bridge maintenance. Costs are listed separately for projects in rehabilitation and replacement, for projects in capital preventive maintenance, and for projects in capital scheduled maintenance. Michigan provided a history of contracts for state-owned bridges. Context, performance, and costs of actions were computed. New York State DOT New York State DOT provided a list of average unit costs for element repair and rehabili- tation. Costs are presented for AASHTO NBEs/BMEs and for groups of elements used by New York DOT. New York DOT also provided a list of maintenance projects completed by contract for the years 2007 through 2016. Bridge age and condition in the year of letting are listed for each project. Ohio DOT Ohio DOT provided records from its transportation management system (TMS) of mainte- nance accomplishments by DOT crews for the years 1998 through 2015. TMS records identify action, bridge, date, and cost. Actions by Ohio crews include bridge cleaning, channel cleaning, deck repair, joint repair, painting, railing repair, sealing of concrete surfaces, sealing abutment seats and backwalls, substructure repair, and superstructure repair. Ohio provided average bid item costs for preservation-related items for the years 2011 to 2015. Oregon DOT Oregon DOT provided a history of projects in its major bridge maintenance program and its transportation improvement program for the years 2011 through 2016. These projects provide approach paving, approach mudjacking, bearing repair, bearing replacement, channel clearing, deck overlay, deck repair, deck sealing, joint repair, joint replacement, joint seals, painting, railing repair, railing retrofit, scour repair, substructure repair, and superstructure repair. Pennsylvania DOT Pennsylvania DOT provided unit costs for flex actions—repair and maintenance actions for bridges tracked in the Pennsylvania DOT bridge management system. Pennsylvania also provided a set of records of completed work supporting the unit costs.

Context, Performance, and Cost of Bridge Preservation Actions 25 Virginia DOT Virginia DOT provided a list of average costs for preservation actions together with the details of Virginia’s method for computing average costs. Washington DOT Washington DOT provided a database of bridge decks and treatments. The database has SN, type of treatment, year of treatment, and cost of treatment. Context, performance, and cost of treatments are computed. Wisconsin DOT Wisconsin DOT provided access to its HSI. HSI contains inventory data, construction history (including maintenance work), and pending recommendations for maintenance. Queries for preservation actions yielded lists of SNs, actions, years, and costs. The research team linked these records to NBI information to evaluate action context, performance, and cost. Highway Bridge Deck Treatments Deck Populations The main populations of deck treatments were as follows: • All treatments • All overlays using PCC • All overlays using polymer concrete • All asphalt wearing surfaces Deck Sub-Populations Main populations were subdivided. For overlays using PCC, the sub-populations were as follows: • Overlays using unmodified PCC • Overlays using latex-modified PCC • Overlays using fly ash added to PCC • Overlays using silica fume added to PCC For overlays using polymer concrete, the sub-populations were as follows: • Overlays using epoxy polymer • Overlays using polyester polymer For asphalt wearing surfaces, the sub-populations were as follows: • Asphalt wearing surface without waterproofing membrane • Asphalt wearing surface with waterproofing membrane Deck Treatments and Probability Distributions for Time Intervals Weibull distributions were fitted to populations of time intervals t(r) and service life. Weibull distributions were formed using both uncensored and censored time intervals. A time

26 Proposed AASHTO Guides for Bridge Preservation Actions interval is uncensored when its start and its end are both observed (i.e., when a history of deck GCRs includes both the condition rating at the date of treatment that starts an interval and the transition to a lower condition rating that marks the end of an interval). A time interval is censored when either its start or its end is not observed. The start of a time interval is not observed if the deck treatment was performed before the year 1992. The end of a time interval is not observed if the range of NBI update files is exhausted, if the SN is no longer found in NBI update files, or if a transition to a higher condition rating is observed. Transition to a higher condition rating may indicate additional treatment or repair of a bridge deck. The remaining portion of the history of condition ratings cannot be attributed solely to the known treatment. Performance of Treatments of Bridge Decks Measures of performance for the main populations of deck treatments are listed in Table 17. Deck treatment increases deck GCRs by about 1.5. The indicated interval for renewal of treatments is about 15 years, midway between the intervals for t(1) and t(2). A slightly longer interval is provided by asphalt wearing surfaces, and a slightly shorter interval is provided by polymer concrete overlays. Performance of Treatments for Sub-Populations Measures of performance for sub-populations of overlays using PCC are listed in Table 18. Modified PCC overlays offer longer intervals at improved deck concrete ratings than unmodified concrete. GCR before treatment was lower for use of modified concrete. This is not an attri- bute of the material; rather, it reflects the use of modified concretes by bridge owners. Some intervals are not listed. For these, there are insufficient observations. Measures of performance for sub-populations of overlays using polymer concretes are listed in Table 19. Intervals for t(0) and t(1) are similar for epoxy polymer overlays and polyester overlays. Measures of performance of asphalt wearing surfaces are listed in Table 20. Waterproofing membranes are used on decks at a somewhat higher condition rating, and waterproofing membranes offer a longer interval at increased condition rating. Treatment (), years Service Life, years0 –1 –2 All 5.4 1.3 11.7 21.4 26.5 33.1 Overlay, PCC 5.5 1.5 10.2 20.6 26.4 32.1 Overlay, Polymer concrete 5.1 1.6 10.0 17.4 21.5 20.7 Wearing surface, Asphalt 6.3 0.2 13.8 26.3 36.3 39.7 Table 17. Measures of performance of deck treatments. Treatment (), years Service Life, years0 –1 –2 Overlay, PCC, all 5.5 1.5 10.2 20.6 26.4 32.1 Overlay, PCC, unmodified 5.7 1.4 7.5 15.6 21.4 38.6 Overlay, PCC, latex-modified 4.9 1.9 21.4 - - 41.1 Overlay, PCC, fly ash 4.8 1.9 23.4 - - - Overlay, PCC, silica fume 5.1 1.6 23.6 - - 51.9 Table 18. Measures of performance of PCC overlays.

Context, Performance, and Cost of Bridge Preservation Actions 27 Treatment (), years Service Life, years0 –1 –2 Overlay, Polymer concrete, all 5.1 1.6 10.0 17.4 21.5 20.7 Overlay, Polymer concrete, epoxy 6.7 0.2 9.3 15.1 16.5 23.8 Overlay, Polymer concrete, polyester 4.9 1.7 10.0 17.5 - - Table 19. Measures of performance of polymer concrete overlays. Young Deck Versus Mature Deck Performance measures are compared for treatments applied to decks in the first 5 years of service and for treatments applied later in service. For measures of performance, young decks are decks that are not more than 5 years old at the time of treatment. Mature decks are older than 5 years at treatment. Performance measures are listed in Table 21. Time intervals are similar for treatments on young decks and on mature decks. Unit Costs of Preservation Actions The research team collected unit costs of preservation actions from state DOT documents and computed unit costs for preservation actions using state DOT datasets. Unit costs of preservation actions are listed in Table 26 of Appendix C. Treatment (), years Service Life, years0 –1 –2 Wearing surface, Asphalt, all 6.3 0.2 13.8 26.3 36.3 39.7 Wearing surface, Asphalt only 6.1 0.3 13.0 25.0 26.3 48.9 Wearing surface, Asphalt plus membrane 6.5 0.1 14.9 35.3 - - Table 20. Measures of performance of asphalt wearing surfaces. Treatment (), years Service Life, years0 –1 –2 All, Young deck 6.2 0.9 10.8 23.4 28.0 37.2 All, Mature deck 5.3 1.3 11.7 21.2 25.9 31.6 Overlay, PCC, Young deck 6.2 0.4 10.2 21.5 25.6 32.7 Overlay, PCC, Mature deck 5.5 1.5 10.2 21.1 26.6 31.6 Overlay, Polymer concrete, Young deck 6.0 1.0 12.3 - - - Overlay, Polymer concrete, Mature deck 5.1 1.6 10 17.3 21.3 19.9 Wearing surface, Asphalt, Young deck 7.1 0.5 15.5 26.0 - - Wearing surface, Asphalt, Mature deck 6.2 0.2 13.7 30.0 27.7 34.4 Table 21. Measures of performance of treatments, young versus mature decks.