Optimal Timing of Pavement Preventive Maintenance Treatment Applications (2004)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

65 APPENDIXES A THROUGH E UNPUBLISHED CONTRACTOR’S MATERIAL Appendixes A, B, C, and E submitted by the research agency are not published herein. Titles of available appendixes are as follows: APPENDIX A Summary of Agency Experiences APPENDIX B Historical Optimization-Based Approaches Used for Transportation-Related Problems APPENDIX C User’s Guide for the Optimal Preventive Maintenance Timing Analytical Tool (OPTime) APPENDIX D Plan for Constructing and Monitoring Preventive Maintenance Test Sections APPENDIX E Example Illustrating the Inclusion of Different Cost Types Appendixes C and E are accessible on the web at http://trb.org/news/blurb_detail.asp?id=4306. The OPTime software in Appendix C can be copied on a CD-ROM for use. For a limited time, copies of Appendixes A and B will be available on a loan basis from the NCHRP. Appendix D is provided on the following pages.

67 APPENDIX D PLAN FOR CONSTRUCTING AND MONITORING PREVENTIVE MAINTENANCE TEST SECTIONS INTRODUCTION The underlying premise of preventive maintenance is that the application of treatments to a pavement in “good” con- dition will provide some benefit above and beyond the per- formance of the untreated pavement. It is further assumed that the benefit will vary, depending on the type of treatment, when it is applied, and the condition of the pavement at the time of application. However, because only a few agencies have had long-term experience with preventive maintenance practices, there is little evidence of these benefits. Attempts to track these benefits after the fact—by examining historical pavement performance data, for example—are problematic because of the absence of critical data, such as the condition of the pavement at the time of treatment application, the qual- ity of construction, and periodic observations of performance. Contributing to the difficulties in documenting the bene- fits of preventive maintenance is the lack of a strong connec- tion between the commonly used methods of monitoring pavement performance and the types of benefits provided by preventive maintenance treatments. Preventive maintenance is often aimed at maintaining or improving functional per- formance while most condition surveys focus on a pavement’s structural performance. Perhaps the best way to evaluate preventive maintenance effectiveness—and show when it is most effective—is through monitoring specially constructed test sections. Properly designed, constructed, and monitored test sections would gen- erate data appropriate for the pavement types, traffic loadings, environmental conditions, and maintenance treatments that are typical of an agency’s practices and conditions. This appendix outlines the steps involved in creating a plan for establishing preventive maintenance test sections that can be used to generate the information needed to implement a successful preventive maintenance program. Results from the experiment would be used to determine the benefits (or effec- tiveness) of specific preventive maintenance treatments based on the age and condition of the pavement. An analysis using the methodology described in Chapter 3 can then be made to identify the optimal time to perform preventive maintenance. The steps in developing the plan include the following: • Identify objectives • Complete experiment design • Construct experiment • Monitor performance • Analyze results OBJECTIVES The first step in developing an experiment is to identify the objectives or goals of the preventive maintenance program to help establish a link between the treatments selected for study, the measures used to monitor performance, and the agency’s expectations. Goals might address pavement smoothness, noise mitigation, accident reduction, and pavement life exten- sion, for example. While the overall objective of the exper- iment is to identify the best time to apply preventive main- tenance, the objective is inextricably linked to preventive maintenance performance objectives. The types of treatments of asphalt and concrete pave- ments that might be evaluated to achieve specific objectives are listed in Table D-1. While almost any treatment could extend pavement life, certain performance objectives would best be achieved with the application of specific treatments. EXPERIMENT DESIGN Perhaps the most important part of the plan is the design of the experiment. An effective design ensures that the objec- tives of the experiment are fully met. Test section sites are selected to meet the immediate and long-term needs of the experiment. While concerns about constructibility and the availability of local support for placing the test sections are recognized. Long-term needs include monitoring and data collection, and subsequent analyses of data. An underlying consideration in locating a test section is to avoid possible confounding factors, such as variability in the pavement con- dition that could impact the interpretation of the results. Some of the key items in the design are discussed in the following sections. Site Selection There are two major issues to consider in site selection: (1) limiting or avoiding confounding factors and (2) ensur- ing that applicable and useful results are obtained from the site. Confounding factors refers to variations in site condi- tions that might later complicate the analysis of the data. Among such factors are non-uniform traffic volumes, cross sections, and support conditions. The key to site selection is to consider the analyses that will be performed and control as many of the factors that will affect them as possible. Under- standing how the findings will eventually be used is another factor in obtaining useful results. For example, if the agency

maintains pavements in different environments, then produc- ing broadly acceptable results from sites constructed in one type of environment should be carefully considered. Similar consideration should be given to other relevant aspects of the site, such as pavement type, design, condition and age, and traffic level. Pavement Type Hot-mix asphalt (HMA), portland cement concrete (PCC), and other bituminous-surfaced pavements are all candidates for inclusion in a preventive maintenance experiment. Test sections should be constructed on the types of pavement for which preventive maintenance treatment applications will be evaluated. It is recommended that initial efforts be kept fairly simple by limiting pavement type to bituminous-surfaced or PCC pavements that have not been rehabilitated or received any other blanket maintenance treatment. While it could be argued that an overlaid pavement will provide similar results for some of the objectives (such as noise mitigation or improved surface friction), the contribution of the original pavement to 68 the performance of the overlay, and to that of the maintenance treatment, cannot be fully isolated. Pavement Design The pavement should be of a uniform design over the length of the project. This means that all structural features (i.e., paving layers, materials, and thicknesses) and geomet- ric features (i.e., number of lanes) should be the same over the length of the project. The subgrade should also be fairly uniform over the project and free of swelling or frost sus- ceptible soils. Pavement Condition and Age Because pavement age is an indirect indicator of the pave- ment condition, a pavement that is fairly young (e.g., less than 5 years old) and still in good condition should be selected. It should not exhibit any signs of significant structural deterio- ration (such as rutting or fatigue cracking), and only small amounts of other types of distress (such as linear cracking or Pavement Surface Type Preventive Maintenance Objective Bituminous PCC Performance Measure Improve Ride (Reduce roughness) Slurry Seal Microsurfacing Ultrathin Friction Course Thin Overlay Diamond Grinding IRI PSI Noise Control Ultrathin Friction Course Slurry Seal Microsurfacing Diamond Grinding dB Increase Surface Friction Chip Seal Slurry Seal Ultrathin Friction Course Thin Overlay Diamond Grinding Skid Number Mean Texture Depth IFI Extend Pavement Life Crack Sealing Fog Seal Scrub Seal Chip Seal Slurry Seal Microsurfacing Thin Overlay Ultrathin Friction Course Joint and Crack Sealing Condition: Cracking Patching Rutting Raveling Faulting Pumping Spalling Potholes Patching Reduce Moisture Infiltration Crack Sealing Scrub Seal Chip Seal Slurry Seal Microsurfacing Thin Overlay Ultrathin Friction Course Joint and Crack Sealing Condition: Cracking Patching Rutting Raveling Faulting Pumping Spalling Potholes Patching IRI = International Roughness Index; IFI = International Friction Index; dB = decibel TABLE D-1 Relationship between performance objectives and preventive maintenance treatments

weathering/raveling) should be present. As with other factors, it is also desirable that the condition of the pavement be fairly uniform over the length of the project and that any sig- nificantly deteriorated areas not be included as part of the experiment. Traffic Levels The traffic levels should be uniform over the project to elim- inate the effect of traffic variability on treatment performance. Low to moderate traffic volumes (e.g., 1,000 to 5,000 vehicles per day) may be most appropriate for the experiment because they cover the conditions for which many treatments are used. While lower traffic volumes make it easier to monitor the performance, roadways with higher traffic volumes provide a more severe test for the treatments. Higher traffic volumes also make it harder to monitor performance and may cause prob- lems when the treatments fail and some form of rehabilitation is required. Also, it is important that adequate construction and performance records be kept not only to fully document the design of the project, but also to help assess the effects of the various treatments on key performance measures. Treatment Selection The selection of preventive maintenance treatments for evaluation in the project should be based on the specific goals of the agency’s preventive maintenance program. The agency must recognize that including different treatments will require a larger test site and will involve the collection of a large amount of data and the conduct of extensive data analysis. For each treatment included in the experiment, additional sections are needed for replicating, and multiple sections are needed for treatment applications at different times in the future. 69 Ultimately, the selected treatments should match the agency’s preventive maintenance objectives. For example, if an agency’s objective is to maintain high levels of surface friction, then treatments that enhance surface friction should be evaluated in the experiment. Table D-2 summarizes some of the primary benefits provided by the different preventive maintenance treatments; this information would help in select- ing treatments to support specific preventive maintenance objectives. Of course, several different treatments intended for different purposes cannot be studied in the same project. As part of the experiment, agencies may also include new materials or techniques or treatments with which they have little or no previous experience. Treatment Timing In the experiment, the timing of the treatment application will be varied so that the effect of treatment timing on per- formance (or effectiveness) can be evaluated. In this regard, two critical issues must be considered: determining when the first treatment should be applied and determining how often subsequent treatments should be applied. On a new pave- ment, a preventive maintenance treatment might be applied before the pavement is opened to traffic (e.g., a fog seal appli- cation to bituminous surfaces) or shortly after construction (e.g., 1 to 3 years). To evaluate timing issues, a number of untreated sections must initially be kept within the experi- ment so that treatments can be applied at different times in the life of the pavement. For example, if a chip seal is applied 2 years after construction, sufficient untreated test sections must be available to allow chip seal application later (e.g., 3, 4, 5, or 6 years). Applying the treatment at 1 year should also be considered to determine if more benefit is obtained from such an early application. Treatment Roughness Friction Noise Life Extension Moisture Reduction Bituminous-Surfaced Pavements Crack Sealing X ✓ Fog Seals X ✓ Scrub Seals ✓ ✓ Slurry Seals ✓ ✓ ✓ ✓ X Microsurfacing ✓ ✓ ✓ ✓ X Chip Seals ✓ ✓ ✓ X Ultrathin Friction Course ✓ ✓ ✓ ✓ ✓ Thin Overlays ✓ ✓ ✓ ✓ ✓ PCC Pavements Joint and Crack Sealing X ✓ Diamond Grinding ✓ ✓ ✓ ✓ ✓ = Major effect x = Minor effect TABLE D-2 Primary benefits of different maintenance treatments

The actual timing of the treatments depends on the type and purpose of the treatment. More substantial maintenance treatments (such as thin overlays) would require greater tim- ing cycles than a lesser treatment (such as a fog seal). Also, the timing cycles are influenced by other factors such as cli- mate and quality of construction; some general guidelines are provided for various preventive maintenance treatments in Table D-3. Site Layout One of the most important aspects of the site layout is its length, it must be long enough to accommodate all the treat- ments under consideration, including control (do-nothing) and replicate sections. The site must be long enough to allow adding treatments to bare sections in subsequent years in order to address the timing issue. Specific items relevant to the site layout are described in the following subsections. Project Length The project must be long enough to accommodate all test sections. As a general rule, the required length of the project can be computed as follows: (Eq. D-1) where: TPL = Total project length, m (or ft). TSL = Total section length, m (or ft) (457 m [1,500 ft] recommended). TPL TSL N TC 1 R= × ×( ) +( )[ ] × 70 N = Number of treatments to be evaluated. TC = Number of timing cycles per treatment. R = Number of sections incorporating each treatment. For example, if four treatments are to be evaluated at three timing cycles (3, 6, and 9 years), and there are to be two sec- tions per treatment (1 set of replicates), then the required proj- ect length is [457 × ((4 × 3) + 1)] × 2, or 11,882 m (38,980 ft). Some additional length may be needed for transitions between sections or to exclude certain areas (such as intersections or bridges) within a project. It can be seen from the example that a test site can become quite long rather quickly, so it is important that agencies care- fully select the number of treatments to evaluate. Of course, the replicate treatments could be placed in the opposing direc- tion which would help shorten the required project length (but potentially add a confounding factor because of different traf- fic levels). Section Length Each individual test section should be long enough not only to facilitate construction but also to provide a statistically valid sampling of performance. With many of the treatments using equipment that requires some start-up calibration, shorter sections could have areas at their beginning and/or end that are not uniform in performance. At the same time, the section should be short enough to help contain the phys- ical size and costs of the experiment. A minimum section length of 457 m (1,500 ft) appears to be reasonable, although longer sections may be warranted in some instances. How- ever, the evaluation length does not need to be as long as the section length; a section evaluation length of 150 m (500 ft) is appropriate. Replicate Sections The use of replicate sections as part of the design is strongly recommended. Replicates are identical sections that are con- structed to improve the statistical validity of the analysis and also to create “back-ups” if the original sections are taken out of service. On the other hand, while more replication improves the reliability of the results it also increases the cost of con- structing, monitoring, and analyzing the results. Although sev- eral replicates makes it easier to break out anomalous behav- ior and improve the statistical validity of the results, only one set of replicates is recommended (that is, a total of two sections for each treatment/timing combination) to reduce cost. It is also recommended that replicate sections be con- structed on the same roadway, end to end in the same lane, if possible, but placed randomly within the project. If site con- straints do not allow this layout, the replicates can be built in the opposing traffic lanes and placed randomly within the proj- ect. For multi-lane roadways, replicates can either be con- Treatment Recommended Year of Initial Treatment Treatment Timing Cycle Crack Sealing 1 to 3 Annually Fog Seals 0 to 3 Annually Scrub Seals 2 to 6 Annually Slurry Seals 2 to 6 Annually Microsurfacing 3 to 7 2 years Chip Seals 2 to 5 Annually to 2 years Ultrathin Friction Course 2 to 6 2 years B itu m in ou s- su rfa ce d Thin Overlays 5 to 8 2 years Joint and Crack Sealing1 4 to 102 2 years PC C- su rfa ce d Diamond Grinding 5 to 10 3 years 1 Refers to joint resealing and crack sealing. It is assumed that if optimal timing of joint resealing is being evaluated, any cracks will be kept sealed. 2 Timing is somewhat dependent on the occurrence of cracking and/or the need for resealing the joints. TABLE D-3 Suggested treatment timing cycles

structed at the end of the project or in the opposing lanes. For example, if chip seals, slurry seals, and thin overlays are con- structed as three separate sections in the northbound lanes of a roadway, replicate sections of the same three treatment types can be constructed in the southbound lanes of the same roadway. While it is possible to apply a treatment to both lanes in one direction of a multi-lane facility and use the sec- ond lane as a replicate, the different traffic level in the repli- cate lane will introduce a confounding factor in the analysis. Factorial Design Factorial designs are typically developed for such experi- ments. These designs are often presented in a tabular form to show what is being evaluated in the experiment in an easy- to-understand manner. A hypothetical example of a factorial design for a project that has been designed to last “n” years is shown in Table D-4. Factorial design tables are an effec- tive way for agencies to lay out their experiment and quickly get an indication of how sizeable it can become. Layout The order and layout of the test sections over the length of a project should be done as randomly as possible. However, given that different treatments will be constructed at differ- ent times, it is logical to construct all the treatments for a given timing cycle at one end of the project, and then proceed from that point for future construction of treatments at sub- sequent timings. An example layout of test sections on a multi-lane facility is shown in Figure D-1. For a two-lane facil- ity, test sections will have to be placed end-to-end. Duration of Experiment The required period of time for monitoring treatments varies depending on the type of treatments. Less substantial preventive maintenance treatments (such as fog seals or crack sealing) will require shorter evaluation periods then those required for treatments such as microsurfacings or thin over- lays. For most bituminous-surfaced sections, the monitoring period is expected to range between 6 and 15 years. 71 The duration of the experiment can be shortened, however, based on regular analysis of the test results. For example, new treatments do not need to be applied and maybe perfor- mance does not need to be further monitored if it is clear that the performance trend is declining. CONSTRUCTION The most important construction concern is ensuring that test sections are properly constructed. This is best accom- plished by following best practice, project specifications, and the material supplier’s recommendations. While this might seem like unnecessary guidance, there are any number of research efforts that have been compromised by construction problems. The following subsections describe specific areas where attention is needed to minimize or eliminate construc- tion problems. Time of Year Several of the bituminous-surfaced pavement treatments are affected by ambient conditions at the time of placement. In particular, the cold-applied thin surfacings do not perform well when placed at low air or pavement temperatures, and chip seals should never be placed on wet pavement when rainfall is expected. Also, joint and crack sealants cannot be placed on damp surfaces. While in practice, preventive maintenance treatments are not always placed during optimal environmen- tal conditions, it makes sense to try to construct the test sec- tions under favorable conditions. This is likely to mean a time of the year when daytime temperatures are 16 °C (60 °F) and rising, freezing is not expected within 24 hours, and rainfall can be avoided. Crack sealants and joint resealing materials are usually placed on a dry pavement when temperatures are mod- erate, such as during late spring or late fall. Quality Control/Quality Assurance The placement of a preventive maintenance treatment should not be approached any differently than other construc- tion undertaken by the agency. However, every effort should be made to ensure that treatments are properly constructed, Timing of Application Year 0 Year 1 Year 2 Year 3 Year 4 Year 5 … Year n Crack Sealing 2 sections 2 sections 2 sections 2 sections 2 sections 2 sections 2 sections Chip Seal 2 sections 2 sections 2 sections Fog Seal 2 sections 2 sections 2 sections 2 sections 2 sections 2 sections 2 sections 2 sections No Treatment (Control) 2 sections TABLE D-4 Example factorial design table

and that subsequent performance is related to the treatment’s capabilities and not to construction defects. It is recommended that the construction of all preventive maintenance treatments follow the agency’s standard speci- fications. In the absence of a standard specification, such as when an experimental material is being evaluated, the sup- plier’s or contractor’s specifications should be followed. The agency should provide inspection services during con- struction to monitor the placement of the test sections. The inspector should be familiar with the project specifications and note the aspects of the project that affect performance, including the following: • Surface preparation – Defects – Overall surface condition 72 – Treatments – Cleanliness • Materials – Constituents – Mix design – Properties • Environmental conditions – Temperature – Humidity – Rainfall • Equipment calibration and performance • Treatment application rates A standard form may be adapted to local conditions and used to record the results of the construction inspection. While the specifications may be fairly detailed, most preventive Two-Lane Roadways Section 1 Section 2 Section 3 Section 4 Section 5 Section 6 Section 7 WB Treatment B Timing X Control Treatment C Timing X Treatment A Timing X Treatment C Timing Y Treatment A Timing Y Treatment B Timing Y EB Treatment C Timing Y Section Treatment A Timing Y Treatment B Timing Y Treatment B Timing X Treatment C Timing X Treatment A Timing X Min 457 m Four-Lane Roadways Section 14 Section 13 Section 12 Section 11 Section 10 Section 9 Section 8 WB Treatment A Treatment C Treatment B Control Treatment A Treatment B Treatment C WB Timing X Timing X Timing X Section Timing Y Timing Y Timing Y Section 1 Section 2 Section 3 Section 4 Section 5 Section 6 Section 7 EB Treatment B Control Treatment C Treatment A Treatment C Treatment A Treatment B EB Timing X Section Timing X Timing X Timing Y Timing Y Timing Y Min 457 m Total Project Length Depending on Number of Treatment/Timing Combinations Total Project Length Depending on Number of Treatment/Timing Combinations Figure D-1. Example test section layouts for 2-lane and 4-lane roadways.

maintenance treatments are not complex, and required plans can typically be covered in a single 81/2 × 11 sheet of paper. Contractor Versus Agency Forces Some treatments are applied solely by specialty contrac- tors and equipment (such as microsurfacing, diamond grind- ing, and ultrathin surface treatments), while others may be applied by agency forces or contractors (e.g., fog seals and chip seals). There is no compelling reason to use contractors; perhaps the best practice to follow is to apply the treatments in the same manner that the agency would normally follow. Voiding Sections Sections that are improperly placed should be voided and removed from the experiment (i.e., not be further monitored). Signs of improper placement include failure of the treatment to “stick,” improper application rates, placement outside of the recommended environmental conditions (temperature and moisture, for example), and failures within the first year. Section Marking It is extremely important to be able to locate the various pavement sections for many years after construction in order to (1) place the treatments in the right locations in subsequent years and (2) perform the necessary performance evalua- tions. The use of permanent markers, such as surveying nails driven into the pavement, is preferred over paint, which can wear off over time and under traffic. Often test sections are marked and remarked on the shoulder, but this may not be possible with certain surface treated or granular shoulders. If the shoulders cannot be permanently marked with the test section limits, delineators should be placed adjacent to sec- tion limits at a safe distance to the side of the pavement. A map to the experimental section should also be devel- oped. The map should show the locations of permanent land- marks (e.g. culverts, intersections, etc.) and offsets to the var- ious test sections. The map should be updated whenever new sections are constructed. MONITORING ACTIVITIES Regular monitoring of the experimental pavement sections is needed to assess the effects of the treatment and timing combinations. There are a wide variety of monitoring activ- ities that can be carried out; data collection efforts should focus on collecting information that will facilitate the evalu- ation of the treatment objectives (see Table D-1). The types of information that could be monitored are described in the following subsections. 73 Manual Condition Surveys It is recommended that manual condition surveys be con- ducted on all experimental sections within a project on at least an annual basis. This can be done using an agency’s distress manual or any similar manual that provides uniform defini- tions of distress type and severity (such as the LTPP distress manual). Recommended distress types that should be col- lected are listed in Table D-5, but highway agencies may include additional distresses as appropriate. A 150-m (500-ft) segment located within the central part of each section and away from the transition areas at either end should be selected as the monitoring sample unit. Roughness Roughness should be measured on all experimental sec- tions within a project on an annual basis. The use of profil- ing equipment is recommended, and the results should be expressed in terms of an International Roughness Index (IRI). ASTM E1926, Standard Practice for Computing Interna- tional Roughness Index of Roads from Longitudinal Profile Measurements, and AASHTO PP 37-00, Standard Practice for Quantifying Roughness of Pavements, provide details on these measurements. Surface Friction If improving surface friction is a goal of placing the pre- ventive maintenance treatments, then surface friction should be monitored on an annual basis. Surface friction is generally measured using a locked-wheel skid trailer with either a ribbed or smooth tire; however, the smooth tire correlates better with surface texture and wet-weather accidents. The output of the surface friction is expressed as either a skid number (SN) or in terms of the International Friction Index (IFI). Applicable specifications include ASTM E1960 and ASTM E274. HMA and Bituminous Pavements PCC Pavements Block Cracking Corner Breaks Fatigue Cracking Linear Cracking Linear Cracking Joint Seal Damage Rutting Joint Spalling Bleeding Joint Faulting Raveling Pumping Weathering (Oxidation) Blowups Polished Aggregate Patching Potholes Patching TABLE D-5 Recommended distress types to be collected

Surface Texture Closely related to surface friction is surface texture, which refers to the variations in a pavement surface that contribute to wet-weather friction, tire-pavement noise, splash and spray, rolling resistance, and tire wear. If determined to be appro- priate for a project, it is recommended that surface texture be measured on an annual basis using either the sand patch test or an outflow meter; alternatively, the use of high-speed, laser- based profiling devices could be used if available. The pre- ferred method of reporting surface texture is the mean tex- ture depth (MTD); if automated equipment is used the output is an estimate of MTD, which is also acceptable. ASTM E965 is the relevant specification. Noise Noise produced by the tire-pavement interaction of vehi- cles may be a concern in urban areas. If controlling noise lev- els is one of the goals of an agency’s preventive maintenance treatments, it should be monitored for consideration in the analysis. Photo and Video Documentation As a final part of the data collection activities, it is recom- mended that each test section be photographed and perhaps videotaped during each annual inspection to provide a per- manent record of the treatment condition over time. These items are perceived to be the primary performance indicators to be collected for the experimental pavement sec- tions. There may be additional indicators that highway agen- cies may wish to include for specific treatments or to ensure compatibility with other performance monitoring that they may be conducting. Treating Failure Eventually, the test sections will fail. If all goes well, fail- ure will occur at the end of the life of the treatment. However, as part of the design, the agency must be prepared to address sections that fail either due to a construction problem or due to failure of the pavement. The following guidance may be used to formulate a response to various failures. It is based on the premise that it is treatment timing that is being inves- tigated in this experiment and not treatment performance. Therefore, certain treatment failures may be repaired to allow the section to stay in service. • Alligator cracking or other localized structural failures— Fix pavement failures and continue to monitor the treat- ment if possible. 74 • Bleeding—If localized, the section can continue to be monitored. If widespread, remove the section from the experiment. • Rutting—It is unlikely that rutting is related to the per- formance of a preventive maintenance treatment; the cause and extent of rutting should be evaluated. Rutting that is prevalent throughout the project indicates that the section may not have been a good candidate for preven- tive maintenance, but rutting that is isolated to different sections may indicate a performance difference. As sec- tions fail due to rutting they should be removed from the experiment, but localized areas may be left in service. • Raveling/Delamination—The cause of these distresses should be further investigated. If they occur in localized areas, then it may be possible to keep the section in ser- vice. Thin bituminous surfacings may be repaired if the problem is localized and repaired soon after it occurs. If the problem is widespread or it is likely that the treat- ment has failed, the section should be taken out of the experiment. • Faulting/Pumping—These are likely signs that the pave- ment was not a good candidate for a preventive mainte- nance experiment; the cause and extent of faulting or pumping should be evaluated. Faulting and rutting that are prevalent throughout the project indicate that the section was probably not a good candidate for preven- tive maintenance, but faulting and rutting that are iso- lated to different sections may indicate a performance difference. As sections fail due to faulting or pumping they should be removed from the experiment, but local- ized areas may be left in service. • Spalling—This should be further investigated to deter- mine whether the spalling is due to a materials/ construction problem or to failure of the sealant system. • Joint or Crack Sealant Failure—If the sealant can be fixed shortly after failure, it should be fixed and the sec- tion monitoring continued. If the sealant has failed and cannot be repaired rapidly, the section should be taken out of the experiment. Ongoing Maintenance Once all treatments are constructed and the pavement is opened to traffic, the issue of what maintenance is allowed needs to be addressed; the agency’s approach should be deter- mined ahead of time. It is recommended that the agency main- tain the serviceability of the experimental section by main- taining the project with the same level of crack sealing and patching that would normally apply to the pavement. DATA ANALYSIS After sufficient performance monitoring data is collected, the optimal timing of a given preventive maintenance treat-

ment can be estimated using the spreadsheet-based analysis tool. The analysis tool allows the calculation of the optimal time to apply a specific treatment by analyzing different treat- ment application ages (timing scenarios) through the compu- tation of a benefit-cost (B/C) ratio associated with each selected timing scenario. The timing scenario with the largest computed B/C ratio identifies the optimal timing of those application ages investigated. The primary reason for implementing a plan is to collect performance data that can be used to compute benefit values associated with different timing scenarios for a specific pre- ventive maintenance treatment. Benefit is defined as any observed influence (mostly positive, but it could also be nega- tive) on any one or more condition indicators resulting from the application of a preventive maintenance treatment. Using this definition, there could be many different types of benefit associated with a given application of the treatment (e.g., applying a chip seal could result in benefits in the form of improved friction, retarded oxidation, or reduced rutting). Benefit for a given condition indicator is determined by comparing the area associated with the condition indicator curve without the application of preventive maintenance (i.e., the do-nothing curve) with the area associated with the condi- tion indicator curve that is altered by the application of the pre- ventive maintenance treatment. For condition indicators that decrease over time (e.g., serviceability, friction, or a typical composite index), it is the area under the curve that defines benefit. For condition indicators that increase over time (e.g., roughness, cracking, rutting, faulting, and spalling), it is the area above the curve that defines benefit. Figure D-2 illustrates the resulting benefit area (AREABENEFIT) for a chosen condition indicator (e.g., serviceability [roughness]) when a treatment is applied at a pavement age of 12 years. 75 The treatment performance data collected as part of the plan is directly used to define these post-preventive mainte- nance relationships associated with each unique combination of condition indicator and treatment application timing. Database Development To effectively use the provided analysis tool, it is recom- mended that a database be built to store the performance and cost data that will be collected as part of the plan. The primary data types required by the analysis tool are the following: • Do-nothing expected condition indicator (perfor- mance)—Before the influence of a preventive mainte- nance application can be analyzed, the analysis requires a baseline performance curve (or curves). The baseline performance curve of interest for a particular condition indicator is that performance curve (condition indicator versus time) that the agency would expect if only routine maintenance were conducted on the pavement (such curves are referred to as do-nothing performance curves). The current methodology requires that the user define a do-nothing performance curve for each of the condition indicators that are included in the analysis. The best source for this information is existing pavement man- agement systems, although users without access to such curves can easily be walked through a process of approx- imation. Within the analysis tool, each do-nothing per- formance relationship may be defined as (1) a known equation (i.e., defined by an equation type and associ- ated coefficients) or (2) a series of performance versus age points through which a regression equation is fit. Se rv ic ea bi lit y (ro ug hn ess ) Age, yrs 4.0 5.0 3.0 2.0 1.0 5 10 15 20 AREABENEFIT Condition trigger level = 2.5 Do-nothing condition indicator curve Expected do-nothing service life = 20 yrs Condition indicator curve after PM application Application of PM activity Expected service life with PM = 22 yr Figure D-2. Illustration of benefit associated with the application of a preventive maintenance treatment.

• Treatment performance relationships—In order to compute the benefit associated with a given performance indicator, the performance monitoring data collected under the plan is used to define the pavement’s per- formance after a preventive maintenance treatment is applied. These relationships are used to compute the benefit associated with each unique combination of con- dition indicator and application age. As with the do- nothing curves, each post-preventive maintenance rela- tionship may be defined by either defining a known equation or by entering a series of performance versus age data. • Treatment cost data—As part of implementing the plan, detailed cost-related records should be kept during treat- ment construction to document the treatment-related costs incurred by the agency. Refining Data for Analysis Upon collecting and organizing all relative performance and cost data, the user must refine the data to facilitate use in the analysis tool. The goal is to get one performance-versus- time relationship for each unique combination of condition indicator and treatment application timing. Therefore, data from replicate experimental sections must be combined into one representative performance relationship. This may be accomplished by using engineering judgment or mathemati- cal techniques such as averaging expected condition values at each treatment age. Replicate cost data should be analyzed using similar methods. While statistical analyses such as t-tests are most appropriate for analyzing whether replicate data are representative of the mean values, it is unlikely that there will be enough replicates to apply such tests. Conducting the Analysis In addition to defining the many performance relationships (do-nothing and treatment-related) required by the analysis tool, many other project specific data elements must also be defined prior to conducting the analysis. The following are the primary steps involved in the analysis: • Condition indicator selection—Specification of one or more condition indicators used to define pavement performance. • Preventive maintenance treatment selection—Selection of one preventive maintenance treatment to be analyzed. • Selection of treatment application ages—Definition of more than one treatment application age that will be 76 compared in the analysis (the analysis will identify the most cost-effective application age from among those ages included in the analysis). • Definition of do-nothing performance curves—Do- nothing condition indicator relationships are entered (as defined equations or as data for regression analysis) to define the baseline pavement performance without pre- ventive maintenance. • Definition of post-preventive maintenance performance curves—Performance data/relationships collected as part of the plan are entered directly into the analysis tool. • Definition of costs—Inclusion of one of three costs types: (1) treatment construction costs, (2) work zone-related user delay costs, and (3) rehabilitation costs (applied at the end of a pavement’s expected service life). • Benefit ranking factors—If multiple condition indica- tors are selected, an individual benefit is calculated for each and ranking factors are assigned as a means for dif- ferentially weighting the individual benefits associated with the different condition indicators. The analysis will provide detailed information associated with each application age. For each considered application age, the output data include a detailed benefit summary (both individual benefit values as well as a total combined benefit), a detailed cost summary, the computed B/C ratio, and the computed Effectiveness Index (EI). The timing scenario with the largest computed B/C ratio (i.e., EI = 100) is the most cost-effective application age. SUMMARY The recommended plan describes an approach to help highway agencies collect the necessary data for determining the optimal time to apply preventive maintenance treatments. Successful implementation of this plan requires identifica- tion of the objectives of the preventive maintenance program and then selection of treatments and monitoring methods that match these objectives. Recommendations are provided for site selection, site lay- out, construction, and monitoring. The approach for analyz- ing the collected data is described in this report; analysis can be facilitated through the use of OPTime, the software tool developed in this project. It should be emphasized that this plan is intended to identify the optimal time to perform a spe- cific preventive maintenance treatment, not to identify the best preventive maintenance treatment.