Implementation of the AASHTO Mechanistic-Empirical Pavement Design Guide and Software (2014)

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

29 INTRODUCTION During the development of the design procedure and with the release of the MEPDG and accompanying software, trans- portation agencies have been confronted with determining and defining data needs; determining applicability and use of the design procedure for the highway network; evaluating the sensitivity of the performance prediction based on material, traffic, and climatic inputs; and calibrating of the performance prediction models to local conditions. Much of this effort has been documented through various agency and national research studies. The level of activity is demonstrated by more than 600 documents directly associated with the MEPDG, and there are hundreds more dealing with materials properties, performance prediction, and traffic analysis. One challenge that agencies face is assimilating the infor- mation obtained from the various research studies, reports, and articles, and applying those results within their own agency. However, one benefit of a nationally developed pave- ment design approach is the ability for agencies to share infor- mation related to evaluation, implementation, and calibration. In this manner, agencies that have implemented the MEPDG can share lessons learned with agencies that are in the process of, are just beginning, or have yet to begin the evaluation and implementation process. Based on the survey responses, three agencies indicated that the MEPDG has been implemented: Indiana, Missouri, and Oregon. Table 24 provides an overview of the organi- zational information related to these agencies. Missouri and Oregon DOTs indicated that the pavement designs are con- ducted, reviewed, and approved by a central or headquarters office (centralized), whereas Indiana DOT is decentralized and pavement designs are conducted by district personnel. All agencies indicated that there was an MEPDG champion leading the implementation effort, and Indiana and Oregon both have an implementation oversight or review commit- tee. In addition, all agencies indicated there was consistent coordination (e.g., open discussion and access to data and information) across the entire agency. The three DOTs have implemented a number of the pave- ment types and rehabilitation treatments included in the MEPDG. Table 25 lists the predominant pavement types and preservation and rehabilitation treatments implemented by each of the three agencies. In an effort to provide information that may be useful for other agencies in the implementation of the MEPDG, this chapter describes the implementation processes used by these three agencies. The majority of each agency case example is based on the results of the survey conducted for this synthesis, and supplemented with follow-up questions and agency-provided documents and research reports (when applicable). INDIANA DEPARTMENT OF TRANSPORTATION Indiana DOT manages and maintains a highway network of 27,879 lane-miles that includes 5,146 lane-miles of interstate routes, 5,529 lane-miles of non-interstate National Highway System (NHS) routes, and 17,204 lane-miles of non-NHS routes (BTS 2011). Table 26 lists new construction, preser- vation, and rehabilitation pavement types used on the state highway system. Indiana DOT operates under a decentralized organiza- tional structure—pavement designs are conducted, reviewed, and approved at the district level. Major projects, such as warranty, alternative bid, and design-build, are designed and finalized by the Central Office. Indiana DOT also indicated that there is open discussion and access to data and informa- tion across all offices within the agency. Pavement Design Process As noted previously, pavement designs are conducted, reviewed, and approved by the Design Office within each of the six Indiana DOT districts. Pavement designs for new and overlaid JPCP and asphalt pavements are based on the MEPDG and are either conducted by a consultant or by an agency engi- neer. If conducted by a consultant, all required data inputs and calibration coefficients are provided in the final submittal doc- uments. Most pavement designs are finalized by the Indiana DOT Engineering Office. MEPDG Implementation Process Based on the recommendations of the Indiana DOT Pavement Steering Committee, with full commitment from the execu- tive staff, Indiana DOT began evaluation of the MEPDG in 2002 and fully implemented the design procedure in 2009. chapter five CASE EXAMPLES OF AGENCY IMPLEMENTATION

30 In general, the implementation plan included the following (Nantung et al. 2005): • Review current state of knowledge in pavement engi- neering and management. • Review and document hierarchical design input parame- ters for each level of design accuracy (document sensitiv- ity of design inputs to distress and smoothness prediction). • Review and document relevant data contained in the Indiana DOT and LTPP databases. • Review the readiness of laboratory and field equipment needed for quantifying higher-level MEPDG inputs. Acquire needed equipment and develop a testing program. • Develop and execute a plan to establish: – Local calibration and validation of distress prediction models, – Regions and segments for traffic input module, and – Software populated with additional climatic data. • Establish a “mini LTPP” program to more accurately calibrate the MEPDG performance prediction models. • Develop correlations and equations for soil resilient modulus, load spectra regions and segments based on existing WIM (weigh-in-motion) and AVC (automated vehicle classification) data, and a process to aid design- ers in easily migrating traffic data into the software. • Provide technology, knowledge transfer, and MEPDG training to other divisions, districts, local agencies, contractors, and consultants. • Revise Indiana DOT Design Manual Chapter 52, “Pavement and Underdrain Design Elements”. In addition, a design memorandum was established indi- cating that the MEPDG should be used in the development of a pavement design recommendation. The memorandum also includes the implementation plan (INDOT 2009). In response to the availability of data, Indiana DOT indi- cated that the existing pavement structure (both layer type and layer thickness), material properties, traffic data, and pavement condition data were available electronically for all state highways. Therefore, Indiana DOT focused much of its evaluation effort on material characterization, traffic, performance prediction, and in-house training of pavement design personnel. Agency Organizational Structure MEPDG Champion Oversight Committee Interagency Communication1 Indiana DOT Decentralized State pavement design engineer and research manager Yes Consistent coordination Missouri DOT Centralized Chief engineer No Consistent coordination Oregon DOT Centralized State pavement design engineer Yes Consistent coordination 1Consistent coordination—open discussion and access to data and information across all divisions. TABLE 24 SUMMARY OF AGENCY ORGANIZATION Agency New Construction Rehabilitation Indiana DOT Asphalt and JPCP Asphalt and JPCP overlays Missouri DOT Asphalt and JPCP Asphalt and JPCP overlays Oregon DOT Asphalt (high-volume roadways only), JPCP, and CRCP CRCP overlays TABLE 25 SUMMARY OF MEPDG IMPLEMENTED PAVEMENT TYPES New Construction Preservation and Rehabilitation Thin (<6 in.) asphalt over unbound aggregate Thick (>6 in.) asphalt over unbound aggregate Asphalt over subgrade/stabilized subgrade Asphalt over cementitious stabilized layers (e.g., lime, lime-fly ash, cement) Composite (new asphalt over new concrete) JPCP Asphalt overlays of existing asphalt, concrete, and fractured concrete pavements Concrete overlay of existing asphalt Bonded and unbonded JPCP concrete overlays HIR without an asphalt overlay Full-depth reclamation with an asphalt overlay Crack or break and seat with an asphalt overlay Rubblization with an asphalt overlay Dowel bar retrofit TABLE 26 PAVEMENT TYPES—INDIANA DOT

31 During 2009, Indiana DOT conducted an MEPDG analysis of more than 100 pavement sections, which included all new pavement designs and all existing pavement designs that had yet to be awarded for structural adequacy before construction (Nantung 2010). The pavement design types included in the Indiana DOT MEPDG implementation are shown in Table 27. The majority of the focus by Indiana DOT is in the imple- mentation of the MEPDG (Nantung et al. (2005). One unique characteristic of the Indiana DOTs implementation effort is that it was conducted using Indiana DOT staff. Materials Characterization A sensitivity analysis of material inputs was conducted on an atypical Indiana DOT pavement structure using input values similar to those used in the previous Indiana DOT design proce- dures (AASHTO 1993 Guide). The input values (using Level 3 inputs) for the pavement design were varied (one parameter at a time) and the predicted performance compared with those of the base design. For concrete pavements, Indiana DOT con- cluded that additional laboratory testing would be necessary to quantify concrete strength parameters, as well as other mix parameters (e.g., coefficient of thermal expansion) not cur- rently available. To achieve this, Indiana DOT worked with five local contractors who conducted concrete strength test- ing to evaluate the construction specification requirements. For asphalt pavements, Indiana DOT developed a database for dynamic modulus, creep compliance, and indirect tensile strength of asphalt mixtures commonly specified by Indiana DOT. The layer thicknesses for the Indiana DOT LTPP asphalt pavement sites were also re-evaluated using the MEPDG. The calibration coefficients for the asphalt performance prediction models were then adjusted so that the predicted performance more closely reflected the actual measured performance. Finally, asphalt models were further calibrated using data from the Indiana DOT accelerated pavement testing facility and “mini LTPP” sites. For unbound layers and subgrade soils, strengths are determined according to resilient modulus (AASHTO T307, Determining the Resilient Modulus of Soils and Aggregate Materials) obtained from triaxial testing. The agency has also generated a subgrade resilient modulus database, and devel- oped a simplified approach for resilient modulus testing and a predictive equation for estimating material coefficients k1, k2, and k3 using soil property tests (Kim and Siddiki 2006). Traffic To improve the ability of pavement designers to obtain traffic- specific data, Indiana DOT developed a software tool that provides visualization of WIM and AVC site locations and easy access for obtaining the required MEPDG input data. The software tool also allows the user to select relevant traf- fic data and directly export it to the MEPDG software. Indi- ana DOT also conducted a sensitivity study to determine the influence of traffic inputs on the predicted performance, gen- erated WIM and AVC geographic information system maps, and analyzed WIM and AVC data to generate axle-load spec- tra data. Indiana DOT also provides the truck weight road group (TWRG) database for Level 2 inputs. The TWRG is divided into four groups based on the average annual daily truck traffic (AADTT). Identifying Existing Conditions Indiana DOT annually collects automated pavement condi- tion and IRI data collection and video logging on the entire highway network. The data contained within the pavement management system also provides sufficient project-level information that was available for use during the calibration process. In addition to pavement distress and IRI data, the pavement management system also includes as-designed layer type and thickness, and the progression of pavement distress and IRI over time. Indiana DOT indicated that the distress definitions used by the agency match those in the Distress Identification Manual for smoothness and all asphalt and concrete distress types predicted in the MEPDG. Input Level The level of input selected for each of the MEPDG input values for Indiana DOT is shown in Table 28. Local Calibration Indiana DOT has conducted a local calibration of the MEPDG performance prediction models for IRI and asphalt alliga- tor cracking (Table 29). The DOT also determined that the MEPDG performance prediction models for asphalt thermal cracking, asphalt rutting, concrete transverse cracking, and concrete faulting models are applicable to local conditions and have therefore adopted these models without modifica- tion. Indiana DOT also indicated that the predicted distress from the asphalt longitudinal cracking, total rut depth, and asphalt thermal cracking prediction models are not used to determine the final pavement layer thicknesses. Table 30 lists the threshold limits for asphalt (new and overlays) and concrete (new and overlays) pavements. These values differ slightly from the default values included in the MEPDG and are based on pavement management data, Asphalt Concrete (JPCP) New construction New construction Overlay of existing asphalt Overlay of existing asphalt Overlay of existing JPCP Unbonded overlay Overlay of existing CRCP Overlay of fractured JPCP TABLE 27 MEPDG IMPLEMENTED PAVEMENT TYPES—INDIANA DOT

32 option at the start of the software program and then selects the needed climate, material, and traffic data for import into the analysis project. The pavement performance library, which contains agency-measured pavement performance data, was used for performance prediction verification. Additional Efforts In addition to personnel training and MEPDG libraries, the Indiana DOT has conducted additional efforts to enhance the implementation process. These include: • Software user manual—A software user manual was developed to supplement the help manual contained within the AASHTOWare Pavement ME Design™ software and provides Indiana DOT-specific pavement types, applicable performance prediction models, and input values to guide the pavement designer. • Concurrent designs—Concurrent designs were con- ducted in 2009 to compare the results of the AASHTO 1993 Guide with the MEPDG. • Design Manual revision—Chapter 52 of the Indiana DOT Design Manual was revised to include reference to the use of the MEPDG for the design of pavement structures. This chapter also provides values for the general inputs (e.g., design life, construction months, IRI and distress performance criteria, traffic, climate), and asphalt, con- crete, unbound aggregate, stabilized, and subgrade layers (INDOT 2013). • Model verification—Indiana DOT identified 108 pave- ment sections that were known to be constructed without any construction-related issues (based on construction records and the knowledge of the pavement manager). These pavement sections were located on three road classes (interstate, U.S. highway, and state route) and distributed across each of the six Indiana DOT districts. All pavement sections were evaluated using the MEPDG based on the inputs from the construction records and theoretical asphalt mix design results. The MEPDG pre- dicted performance was compared (based on a review by the agency MEPDG implementation committee) with the measured pavement condition for each of the 108 pave- ment sections. except for transverse cracking in asphalt pavements, which is based on department policy. Training Indiana DOT has developed an in-house program to train personnel in the basics of pavement engineering, pave- ment materials, fundamentals of ME design, application of the MEPDG to Indiana conditions, and software operation. Training modules include an overview of the MEPDG soft- ware, load spectra, material characterization (asphalt, con- crete, and unbound), principles of MEPDG design, traffic, materials, and climate inputs, and incorporate FWD data for pavement design and rehabilitation. Training, from basic pavement knowledge to ME theory, was noted as a key to the implementation of the MEPDG in Indiana. MEPDG Libraries The AASHTOWare Pavement ME Design™ software allows access to established agency databases for materials and traffic (the climatic database is included with the software license). Indiana DOT has created MEPDG-based databases for cli- mate, materials, traffic, and pavement performance data. In this regard, the pavement designer simply selects the database Feature Input Level All traffic1 Regional value All asphalt layers Regional value All concrete layers Regional value All chemically stabilized layers Regional value All sandwiched granular layers Regional value All non-stabilized base layers Poisson’s ratio Regional value Modulus Regional value Sieve analysis Regional value All subgrade layers Poisson’s ratio Regional value Modulus Regional value Sieve analysis Regional value All bedrock layers MEPDG default value 1AADTT is site-specific. TABLE 28 MEPDG INPUT LEVEL—INDIANA DOT Pavement Type Performance Indicator Selected Model All IRI Locally calibrated model Asphalt Longitudinal cracking 1 Alligator fatigue cracking Locally calibrated model Transverse cracking 1 Rut depth—asphalt layers MEPDG model Rut depth—total 1 Reflective cracking 1 JPCP Transverse cracking MEPDG model Joint faulting MEPDG model 1Distress criteria are not used to determine the recommended pavement structure. TABLE 29 PERFORMANCE PREDICTION MODEL SELECTION—INDIANA DOT

33 • Know the details of the agency standard specifications. Indiana DOT has based most of the material inputs for Level 2 and Level 3 on the standard specification requirements for construction quality control and qual- ity assurance. • Some input parameters will depend on agency internal policies. For example, the initial IRI depends on the agency acceptance of pavement smoothness after con- struction, while the terminal IRI depends on the factor of safety for the travelling public. The threshold value for the design reliability and performance criteria in the MEPDG should be based on the agency policy to assume risks. • Conduct the implementation and local calibration effort using agency personnel. The agency personnel know the policies and procedures best and therefore are more qualified to conduct the implementation and local cali- bration processes. • Local calibration is a plus, but should not need a require- ment for implementing the MEPDG. Indiana DOT used 18 LTPP test sections and agency research sections; however, this was not enough to fully calibrate the per- formance prediction models. However, obtaining data to meet all aspects of the calibration process may take years. Indiana DOT took the approach that it was better to conduct verification and validation of selected pave- ment sections with good pavement history, rather than attempt to obtain data on all agency-applicable pave- ment types. • Form an oversight committee to evaluate, guide, and direct the implementation process. • Coordinate and communicate with the materials office and the geotechnical engineering office early in the implementation process. • Provide training in ME fundamentals, MEPDG pro- cedures, and AASHTOWare Pavement ME Design™ software to agency and consulting pavement engineers. • Work together with the pavement associations in resolv- ing any implementation issues or concerns. • The MEPDG provides an improved pavement design procedure in comparison with the AASHTO 1993 Guide. Benefits Indiana DOT estimated that the MEPDG implementation has resulted in cost savings of approximately $10 million per year based on a comparison of resulting pavement structures from the AASHTO 1993 Guide and the MEPDG (Nantung 2010). In addition, although more difficult to quantify, the Indiana DOT survey response also noted that the MEPDG- based designs have improved the reliability of the design recommendations, the characterization of local and new materials, the characterization of existing pavement layers, the characterization of traffic, the confidence in distress pre- diction, and the knowledge of in-house staff in pavement design and pavement performance. Challenges Indiana DOT indicated that one of the most challenging efforts in the implementation effort was incorporating traf- fic data. Although traffic data for Indiana DOT was readily available, significant data retrieval and manipulation was required prior to use in the MEPDG. Indiana DOT also indicated that having buy-in from the executive staff early in the implementation process was criti- cal to its success. In addition, it was also important to provide the executive staff with information on how the agency would benefit with the implementation of the MEPDG. Lessons Learned When asked about the lessons learned from the MEPDG implementation process, Indiana DOT proposed the following: • Review data to identify potential errors. • Setting up the traffic input data (Level 1 or Level 2) requires a significant length of time; therefore, traffic input data needs to be resolved first. • Know the agency construction practice. Be practical in setting up the layers in the MEPDG. Some layers can be designed, but cannot be constructed. Functional Class Asphalt Pavements Concrete Pavements Reliability (percent) IRI (in./mi) Alligator cracking (percent area) Asphalt rutting (in.) Transverse cracking (ft/mi) Transverse cracking (percent slabs) Joint faulting (in.) Freeway 160 10 0.40 500 10 0.15 90 Arterial, Urban 190 20 0.40 700 10 0.20 90 Arterial, Rural 200 25 0.40 600 10 0.22 85 Collector, Urban 190 30 0.40 700 10 0.25 80 Collector, Rural 200 35 0.40 600 10 0.25 75 Local 200 35 0.40 700 10 0.25 70 Modified from INDOT (2013). TABLE 30 PERFORMANCE THRESHOLD LIMITS—INDIANA DOT

34 In 2005, Missouri DOT released the ME Design Manual, which provides details related to pavement design life, dis- tress threshold values, reliability levels, and input values for traffic, pavement structure, and materials (MODOT 2005). In 2009, ME Design Manual—Volume II was released pro- viding the pavement designer with input values for direct use with the AASHTOWare Pavement ME Design™ software (MODOT 2009). In 2009, a research project, Implementing the AASHTO Mechanistic-Empirical Pavement Design Guide in Missouri, was completed; this provides details related to the steps and activities needed to locally calibrate the distress prediction models to Missouri conditions (Mallela et al. 2009). Missouri DOT determined that the implementation efforts should focus on the analysis of state-collected WIM data; materials characterization of typical asphalt, concrete, aggregate base, and subgrade soils; and field testing of in situ pavements for use in distress prediction model calibra- tion. Each of these areas is further described in the following sections. Traffic The traffic data evaluation was conducted by a consultant and included a comparison of the Missouri-specific traffic data with the MEPDG default values and the development of input values based on historical traffic data. In this study, the Missouri DOT’s continuous and portable WIM sites were evaluated and the results of the traffic data analysis indicated the following (Mallela et al. 2009): • The continuous WIM data were of sufficient quality for use in the MEPDG. MISSOURI DEPARTMENT OF TRANSPORTATION Missouri DOT manages and maintains a highway network of 75,999 lane-miles, and includes 5,621 lane-miles of inter- state routes, 10,607 lane-miles of non-interstate NHS routes, and 59,771 lane-miles of non-NHS routes (BTS 2011). The pavement types, new, preservation, and rehabilitation, cur- rently constructed by Missouri DOT are listed in Table 31. Missouri DOT operates under a centralized organizational structure—pavement designs are conducted, reviewed, and approved by the Central Office (Construction and Materials Division, Pavement Section). MEPDG Implementation Process In 2005, Missouri DOT determined that the benefits obtained from implementation of the MEPDG would outweigh the risks associated with adopting an ME-based pavement design pro- cedure that was not yet fully evaluated, calibrated, or validated. Specifically, the reasons for moving forward with MEPDG implementation assumed improved reliability in prediction of pavement condition, potential cost savings, consideration of local and new materials, consideration of local traffic con- ditions, ability to model the effects of climate and materials aging, and improved characterization of existing pavement layer parameters. Missouri DOT became the first highway agency to initiate and implement the use of the MEPDG for the design of new asphalt and concrete pavements by imple- menting the MEPDG for design of thick asphalt on rubblized concrete and concrete overlays in 2008. Missouri DOT also adopted the use of input values primarily based on Level 3. As better inputs became available, those values were adopted and substituted into the MEPDG analysis (MODOT 2005). Table 32 lists pavement types currently evaluated using the MEPDG. New Construction Preservation and Rehabilitation Thick asphalt (>6 in.) over aggregate base JPCP Asphalt overlays of existing asphalt, concrete, and composite pavement Bonded and unbonded JPCP overlay HIR with an asphalt overlay CIR with an asphalt overlay Full-depth reclamation with an asphalt overlay Rubblization with an asphalt overlay Dowel bar retrofit Diamond grinding TABLE 31 PAVEMENT TYPES—MISSOURI DOT Asphalt Concrete (JPCP) New construction New construction Asphalt overlay of existing asphalt Bonded concrete overlay of existing JPCP Asphalt overlay of existing JPCP JPCP overlay of existing asphalt pavement Asphalt overlay of fractured JPCP Unbonded JPCP overlay of existing JPCP TABLE 32 MEPDG IMPLEMENTED PAVEMENT TYPES—MISSOURI DOT

35 is shown in Table 34. Each test section was 500 ft long. Pave- ment testing included coring to quantify asphalt layer proper- ties (thickness, air voids, asphalt content, bulk and maximum specific gravity, and aggregate gradation), concrete properties (thickness, compressive strength, elastic modulus, and coef- ficient of thermal expansion), FWD testing to quantify in situ layer stiffness, manual condition surveys, and analysis of test section historical IRI data (Schroer 2012). The results of the test sections were used during the local calibration of the MEPDG performance prediction models. Input Levels Based on the local calibration effort, Missouri DOT specifies the input levels for each input feature, as shown in Table 35. For the most part, MEPDG default values are used, with regional values for specific inputs (e.g., vehicle class dis- tribution, asphalt mixture volumetrics, concrete mix prop- erties, resilient modulus, and sieve analysis for unbound layers). • The MEPDG default truck traffic classification groups adequately describe the highway traffic distribution on Missouri principal arterial highways. • The MEPDG defaults for vehicle class distributions, axle-load spectra, axles per truck, hourly truck usage, and default monthly adjustment factors are appropriate for routine design. Materials Characterization The laboratory testing for materials characterization was conducted by Missouri DOT on asphalt and concrete mix- tures, dense-graded aggregate base, and subgrade materials. Table 33 provides a summary of the material testing needs and the findings of the analysis. Test Sections In situ pavement testing was conducted on 36 agency-specified sites and 41 LTPP sites. A list of test sites by pavement type Material Type Study Results Asphalt Mixture Asphalt material property inputs were determined and included in the MEPDG materials library for typical Missouri DOT mixtures. The MEPDG dynamic modulus regression equation adequately reflects Missouri DOT mixtures. The MEPDG prediction equation has a tendency to under predict E*, especially at high frequencies Concrete Mixture Concrete material property inputs were determined and included in the MEPDG materials library for typical Missouri DOT mixtures. Laboratory-determined values for compressive strength, flexural strength, and elastic modulus are not statistically different from MEPDG Level 3 default values; therefore, use the MEPDG Level 3 default values. Until long-term data are available use the strength and modulus gain model contained in MEPDG. When Missouri DOT compressive-to-flexural strength correlation is very close to MEPDG, use the MEPDG default values. MEPDG underestimates compressive strength-to-elastic modulus; therefore, use the Missouri DOT-developed relationship. Unbound Material Unbound material property inputs were determined and included in the MEPDG materials library for standard (Type 5) base material and subgrade soils. Source: Mallela et al. (2009). TABLE 33 MATERIAL CHARACTERIZATION—MISSOURI DOT Pavement Type Agency LTPP Total New JPCP 25 7 32 New asphalt 6 14 20 Asphalt overlay of existing asphalt 0 11 11 Asphalt overlay of existing JPCP 0 5 5 Asphalt overlay of rubblized JPCP 0 4 4 Unbonded concrete overlay 5 0 5 Total 36 41 77 Source: Mallela et al. (2009). TABLE 34 FIELD TESTING SITES—MISSOURI DOT

36 and reflective cracking reflect Missouri conditions; local cali- bration was required for transverse cracking, rut depth (asphalt and total), and IRI. For concrete pavements, the MEPDG per- formance prediction models were determined to be acceptable for transverse cracking and joint faulting, and local calibration of the IRI performance prediction model was required. Table 37 summarizes the performance criteria used by the Missouri DOT for asphalt (new and overlays) pavements. Even though several performance prediction models have been locally calibrated, the Missouri DOT currently requires only performance criteria for alligator cracking and rut depth in the asphalt layers. The performance criterion for alligator cracking is based on the expected level of alligator cracking for a “perpetual” pavement (i.e., no deep-seated structural distress). However, the level of expected alligator cracking has not yet been verified from in-service asphalt pavements. The asphalt rutting criterion is based on the approximate depth to reduce the potential for hydroplaning. At this time, Missouri DOT has not implemented the IRI criteria in the pavement design process because it is difficult Data Availability Missouri DOT reported that the existing pavement structure (layer type and thickness), traffic, and pavement condition data are all readily available; however, material properties of the existing pavement structure are difficult to obtain. Both the traffic and pavement condition data are available agency- wide, whereas the existing pavement structure (layer type and thickness) data and material properties are only available at the district level. Agency definitions of pavement distress are similar to those recommended in the Distress Identification Manual. Local Calibration Missouri DOT has conducted an evaluation of the MEPDG performance prediction models to determine applicability to Missouri conditions. A summary of the performance predic- tion model selection is shown in Table 36. For asphalt pave- ments, Missouri DOT determined that MEPDG performance prediction models for longitudinal cracking, alligator cracking, Feature Input Level All traffic inputs (except as indicated below) MEPDG default values AADTT Site-specific Vehicle class distribution Regional values Axles per truck Regional values All asphalt layer inputs (except as indicated below) MEPDG default values Mixture volumetrics Regional value Mechanical properties Regional value All concrete layer inputs MEPDG default values Poisson’s ratio Regional value Unit weight Regional value Mix properties Regional value Strength properties Regional value All chemically stabilized layer inputs Do not use All sandwiched granular layers Do not use All non-stabilized base layers (except as indicated below) MEPDG default values Modulus Regional value Sieve analysis Regional value All subgrade layers Modulus Regional value Sieve analysis Regional value All bedrock layers MEPDG default value TABLE 35 MEPDG INPUT LEVELS—MISSOURI DOT Pavement Type Performance Indicator Selected Model All IRI Locally calibrated model Asphalt Longitudinal cracking MEPDG model Alligator cracking Nationally calibrated model Transverse cracking Locally calibrated model Rut depth—asphalt layers Locally calibrated model Rut depth—total Locally calibrated model Reflective cracking MEPDG model JPCP Transverse cracking MEPDG model Joint faulting MEPDG model TABLE 36 PERFORMANCE PREDICTION MODEL SELECTION—MISSOURI DOT

37 ings resulting from more economical designs, improved characterization of local materials, existing pavement layers and traffic, and improved confidence in distress prediction. OREGON DEPARTMENT OF TRANSPORTATION Oregon DOT manages and maintains a highway network of 18,606 lane-miles, and includes 3,126 lane-miles of interstate routes, 7,267 lane-miles of non-interstate NHS routes, and 8,213 lane-miles of non-NHS routes (BTS 2011). Table 38 lists all new construction, and preservation and rehabilitation pavement types currently constructed by Oregon DOT. Pavement Design Process Pavement designs for Oregon DOT are conducted by agency staff as well as private consultants. The state pavement design engineer is responsible for evaluating, conducting, review- ing, or overseeing all pavement designs for the state highway network. Currently acceptable pavement design procedures include: AASHTO 1993 Guide, MEPDG, Asphalt Institute, Portland Cement Association, Asphalt Pavement Association of Oregon (based on AASHTO 1993 Guide), and American Concrete Pavement Association (ODOT 2011). The standard pavement design procedure used by Oregon DOT for asphalt pavements is the AASHTO 1993 Guide, while the MEPDG analysis is conducted concurrently for comparison purposes. The MEPDG has been fully adopted for new concrete pave- ment design. Oregon DOT operates under a centralized organizational structure—pavement designs are conducted, reviewed, and approved by the central office (Pavement Services Unit). The Pavement Services Unit reports to the state construc- tion and materials engineer and is responsible for pavement design, pavement management, and pavement materials and construction. MEPDG Implementation Process Oregon DOT began evaluation of the MEPDG in 2006, with implementation for new construction (or reconstruc- tion) high-volume routes in 2009. Oregon DOT has devel- oped calibration coefficients for each of the pavement types to determine the initial as-constructed IRI. This is particu- larly problematic for the agency because new construction and reconstruction pavement projects are let as alternate bid contracts. Missouri DOT includes the pavement thickness requirements for both pavement types in the project proposal documents. However, because the determined layer thick- ness may be affected by the initial IRI value, an unfair advan- tage may arise because of differences in as-construction IRI compared with the initial IRI used in the design process. The initial calibration effort for total rut depth was con- ducted in 2006. As pavement designs were being conducted and reviewed, Missouri DOT questioned the validity of the rut depth predictions for unbound base and subgrade layers. The more recent local calibration effort has yet to be accepted; therefore, for now, only the asphalt layer rut depth criteria has been implemented. Training There is currently no formal training for MEPDG and the AASHTOWare Pavement ME Design™ software by Missouri DOT staff; it is self-taught. Additional Efforts Before adopting the MEPDG (and AASHTOWare Pavement ME Design™ software) Missouri DOT was required to obtain buy-in from and address any concerns from the industry, as well as address any concerns or issues with the information technology department. Benefits As discussed previously, Missouri DOT moved forward with the MEPDG implementation process because of the assumed benefits that it would bring. These benefits include cost sav- Performance Indicator Alligator cracking (percent lane area) Rut depth—asphalt only (in.) Threshold Limit 2.00 0.25 Reliability (percent) 50 50 TABLE 37 PERFORMANCE THRESHOLD LIMITS—MISSOURI DOT New Construction Preservation and Rehabilitation Thin (<6 in.) asphalt over unbound aggregate Thick (>6 in.) asphalt over unbound aggregate Asphalt over cementitious stabilized layers Composite (new asphalt over new concrete) JPCP CRCP Asphalt overlays of existing asphalt, concrete, and composite pavements Unbonded CRCP overlay CIR with an asphalt overlay Full-depth reclamation with an asphalt overlay Rubblization with an asphalt overlay Rubblization with a concrete overlay Diamond grinding TABLE 38 PAVEMENT TYPES—OREGON DOT

38 axles per truck, and average individual axle spacing. However, the axle group categories are not combined because each virtual truck classification has a distinct distribution of tandem, tridem, and quad axles. • For more critical roadways, the virtual truck classifica- tion associated with low, moderate, or high truck vol- umes is used. Input Levels The Oregon DOT input levels for each of the MEPDG design inputs are listed in Table 40. For the majority of data inputs, the agency has chosen to use the MEPDG default values. Only a few data inputs are based on site-specific values and these include vehicle class distribution, asphalt mixture volu- metrics, concrete strength properties, and modulus and sieve analysis of unbound and subgrade layers. Identifying Existing Conditions The existing pavement structure (layer type and thickness), associated material properties, traffic, and condition data are readily available for all state highways. Data availability and acquisition is not viewed as an insurmountable issue because the Pavement Design Unit provides a centralized connection between traffic, pavement management, materials testing, and pavement design. Pavement condition data are collected in accordance with the Pavement Distress Identification Manual (Miller and Bellinger 2003). However, modifications to distress defini- tions and measurements have been conducted to better reflect Oregon conditions (ODOT 2011). The distress survey is conducted every other year using semi-automated pavement condition survey procedures. Local Calibration Oregon DOT has conducted a local calibration of the pave- ment performance prediction models for asphalt pavements, JPCP, and CRCP (Table 41). Local calibration was based on the evaluation of 108 pavement test sections representing typical Oregon DOT pavement designs, regional locations (coastal, valley, and eastern), and traffic levels (low, moder- ate, and high) (Williams and Shaidur 2013). shown in Table 39. Currently, Oregon DOT is re-evaluating the calibration coefficients for asphalt pavements (including overlays) and determining the implementation plan for appli- cation to asphalt and concrete overlay designs. Oregon DOT indicated that reasons for implementing the MEPDG included the potential cost savings owing to more economical pavement structure recommendations, consid- eration of local traffic conditions, the effects of climate and materials aging on pavement performance, and consider- ation of the characterization of existing pavement layers. Oregon DOT also indicated that it has improved commu- nications between the pavement design and pavement man- agement offices. As part of the implementation effort, Oregon DOT, through university research projects, focused on material characterization, traffic, and local calibration. Each of these efforts is summarized in the following sections. Materials Characterization Lundy et al. (2005) determined the dynamic modulus for Oregon DOT standard asphalt mixtures. Asphalt mixtures were varied according to air void level, binder grade, and binder content. During testing, the same aggregate source and gradation were used for all mixtures. One of the pri- mary findings from this study was that the MEPDG regres- sion equation resulted in good agreement with the laboratory results. From this analysis, the Oregon DOT adopted Level 3 inputs for asphalt material characterization. Traffic Oregon DOT collects WIM data on 22 sites across the state. The raw data from four of the WIM sites was used to generate “virtual” truck classifications representing three typical daily truck traffic volumes: 500 (low), 1,500 (moderate), and 5,000 (high) trucks per day. The virtual truck classifications are electronically available for import into the AASHTOWare Pavement ME Design™ software. Oregon DOT uses the following WIM data options (Elkins and Higgins 2008): • On less critical roadways, the virtual truck classifica- tion is combined for all seasons and sites to determine average hourly volume distribution, average number of Asphalt JPCP CRCP New construction Overlay of existing JPCP Overlay of existing CRCP New construction New construction Overlay of existing asphalt TABLE 39 MEPDG IMPLEMENTED PAVEMENT TYPES—OREGON DOT

39 Additional Work to Justify Implementation The Pavement Design Unit is currently evaluating the ben- efits of expanding the use of the MEPDG for the design of all new construction and rehabilitated asphalt pavements. SUMMARY This chapter described the successful MEPDG implementa- tion efforts of three state highway transportation agencies. Implementation efforts were presented as case examples that are based on the agency survey responses and follow- up questions, and supplemented with agency documents and research reports. The case examples provided a summary of the MEPDG implementation processes for the Indiana, Missouri, and Oregon DOTs; specifically, information related to the agency pavement design process, MEPDG implementation process, local calibration efforts, staff train- ing efforts, and development of MEPDG materials and traffic databases. Training There is currently no formal training for MEPDG and the AASHTOWare Pavement ME Design™ software by Oregon DOT staff; it is self-taught. Benefits Oregon DOT indicated that implementation of the MEPDG will improve the confidence in the performance prediction models; however, benefits have yet to be quantified. Challenges Oregon DOT indicated that a number of issues have impeded the MEPDG implementation, including availability of materials characterization data, funding restrictions, limited time available, and justification of benefits for implementing more advanced procedures. Feature Input Level All traffic inputs (except as indicated below) Regional value Vehicle class distribution Site-specific value AADTT Site-specific value Axles per truck MEPDG default value Axle configuration MEPDG default value Wheelbase MEPDG default value All asphalt layer inputs (except as indicated below) MEPDG default value Mixture volumetrics Site-specific value All concrete layer inputs MEPDG default value Strength properties Site-specific value All chemically stabilized layer inputs Do not use All sandwiched granular layers Do not use All non-stabilized base layers (except as indicated below) MEPDG default value Modulus Site-specific value Sieve analysis Site-specific value All subgrade layers MEPDG default value Modulus Site-specific value Sieve analysis Site-specific value All bedrock layers MEPDG default value TABLE 40 MEPDG INPUT LEVELS—OREGON DOT Pavement Type Performance Indicator Selected Model All IRI MEPDG model Asphalt Longitudinal cracking Locally calibrated model Alligator cracking Locally calibrated model Transverse cracking Locally calibrated model Rut depth—asphalt layers Locally calibrated model Rut depth—total Locally calibrated model Reflective cracking MEPDG model JPCP Transverse cracking MEPDG model Joint faulting MEPDG model CRCP Punchouts MEPDG model TABLE 41 PERFORMANCE PREDICTION MODEL SELECTION—OREGON DOT

40 prediction model to local conditions. When the MEPDG performance prediction model did not adequately represent measured conditions, the agency recalibrated the perfor- mance prediction model. In Table 44, “National” indicates that the MEPDG performance prediction model was selected for use, while “Local” indicates that the performance predic- tion model was calibrated to local conditions. For asphalt pavements, Indiana DOT reviews but does not consider the longitudinal cracking, total rut depth, thermal cracking, or reflective cracking criteria for determining the final pavement layer thicknesses, but has locally calibrated the alligator cracking and IRI performance prediction mod- els, and adopted the MEPDG asphalt rut depth performance prediction model. For JPCP pavements, Indiana DOT has adopted the MEPDG performance prediction models for Table 42 summarizes the pavement types evaluated by these three agencies using the MEPDG for each agency. As noted, all agencies utilize the MEPDG analysis for quan- tifying the pavement structure for asphalt and JPCP new construction, and Oregon DOT also includes CRCP new construction. Agencies utilize the MEPDG for analyzing a variety of asphalt and concrete overlay options. Table 43 lists the input levels selected by each of the three agencies. For the majority of inputs, Indiana DOT has selected regional values; Missouri DOT uses a combination of regional and MEPDG default values; and Oregon DOT uses predominately site-specific and MEPDG default values. Table 44 lists the performance prediction models selected by each agency for asphalt and concrete pavements. All agen- cies evaluated the applicability of the MEPDG performance Pavement Type Indiana DOT Missouri DOT Oregon DOT New construction—Asphalt New construction—CRCP New construction—JPCP Asphalt overlay of existing asphalt Asphalt overlay of existing CRCP Asphalt overlay of existing JPCP Asphalt overlay of fractured JPCP CRCP overlay of existing asphalt JPCP bonded concrete overlay JPCP overlay of existing asphalt Unbonded overlay TABLE 42 SUMMARY OF MEPDG IMPLEMENTED PAVEMENT TYPES Feature Indiana DOT Missouri DOT Oregon DOT All traffic inputs (except as noted) Regional MEPDG Regional AADTT Site-specific Site-specific Site-specific Vehicle class distribution Regional Site-specific Axles per truck Regional MEPDG Axle configuration MEPDG Wheelbase MEPDG All asphalt layer inputs (except as noted) Regional MEPDG MEPDG Mixture volumetrics Regional Site-specific Mechanical properties Regional All concrete layer inputs (except as noted) Regional MEPDG MEPDG Poisson’s ratio Regional Unit weight Regional Mix properties Regional Strength properties Regional Site-specific All chemically stabilized layer inputs Regional Do not use Do not use All sandwiched granular layers Regional Do not use Do not use All non-stabilized base layers (except as noted) MEPDG MEPDG Poisson’s ratio Regional Modulus Regional Regional Site-specific Sieve analysis Regional Regional Site-specific All subgrade layers (except as noted) MEPDG Poisson’s ratio Regional Modulus Regional Regional Site-specific Sieve analysis Regional Regional Site-specific All bedrock layers MEPDG MEPDG MEPDG TABLE 43 SUMMARY OF MEPDG INPUT LEVELS

41 be required for the ME design process, MEPDG, and soft- ware. In addition, it will be necessary for agencies to deter- mine MEPDG-specific details, such as threshold criteria and reliability levels, input levels, materials and traffic inputs, and applicability of predicted performance to field conditions. To address these issues and concerns, agencies identified a num- ber of aids (e.g., user guides, data libraries, and training) that can be used to assist an agency in the implementation process. A number of these implementation aids are listed in Table 45. All agencies, to some extent, have conducted materials charac- terization. All agencies have also characterized traffic accord- ing to the data requirements contained within the MEPDG. Both Indiana and Missouri DOTs have identified calibration sections, developed materials and traffic libraries, developed implementation plans, and created an agency-specific ME user guide. Indiana DOT has also conducted concurrent designs and modified the agency design manual. Oregon DOT is in the process of developing an implementation plan for asphalt and concrete overlay design. Only Indiana DOT has developed an in-house training program for agency staff. both slab cracking and joint faulting, and has locally cali- brated the IRI performance prediction model. Missouri DOT has adopted the MEPDG performance prediction model for longitudinal cracking, thermal cracking, and reflective cracking, and for asphalt pavements has locally calibrated the IRI, alligator cracking, and asphalt and total rut depth performance prediction models. For JPCP, Missouri DOT has adopted the MEPDG performance prediction models for slab cracking and joint faulting, and has locally calibrated the IRI prediction models. Oregon DOT has adopted the MEPDG IRI prediction model, and has locally calibrated all other asphalt pavement performance prediction models. For concrete pavements, Oregon DOT adopted the MEPDG performance prediction models for both JPCP and CRCP. Implementation of the MEPDG (or any new process) requires more effort than just evaluating the applicability of the process to agency conditions. For example, training may Pavement Type Performance Indicator Indiana DOT Missouri DOT Oregon DOT Asphalt IRI Local Do not use National Longitudinal cracking Do not use Do not use Local Alligator cracking Local Local Local Transverse cracking Do not use Do not use Local Rut depth asphalt layers National Local Local Rut depth total Do not use Do not use Local Reflective cracking Do not use Do not use National JPCP Transverse cracking National National National Joint faulting National National National IRI Local Local National CRCP Punchouts Not applicable Not applicable National IRI National Oregon DOT has adopted the MEPDG IRI prediction model, and has locally calibrated all other asphalt pavement performance prediction models. For concrete pavements, Oregon DOT adopted the MEPDG performance prediction models for both JPCP and CRCP. TABLE 44 SUMMARY OF AGENCY-SELECTED PERFORMANCE PREDICTION MODELS Feature Indiana DOT Missouri DOT Oregon DOT Materials characterization Asphalt Concrete Unbound aggregate Subgrade soils Traffic characterization Test sections Training Utilization of pavement management data Material library Traffic library Implementation plan In progress MEPDG user guide Concurrent designs Design manual revisions Indicates agency-developed implementation aid. TABLE 45 SUMMARY OF AGENCY IMPLEMENTATION AIDS