Rehabilitation of Asphalt Concrete Pavements: Initial Evaluation of the SPS-5 Experiment—Final Report

6. SUMMARY AND CONCLUSIONS

The SPS-5 experiment entitled Rehabilitation of Asphalt Concrete Pavement is a key experiment of the LTPP program. The main objective of SPS-5 is to determine the long-term effectiveness of different rehabilitation techniques on pavement performance and service life. There are some concerns about whether SPS-5 can meet its expectations given that several projects were constructed on coarse-grained, subgrade soils. Although this is considered a significant deviation from the experimental plan, it is not believed to be detrimental to the overall expectations for this experiment as long as it is fully considered in the data analyses.

This report presents results from the first comprehensive review and evaluation of SPS-5. Issues of experiment design, construction quality, data availability and completeness, and early performance trends have been addressed. That unavailable data have been identified here does not necessarily mean that the data were not collected or submitted by the RCO or owner agency that built the individual projects. There can be several reasons that good data can be delayed before reaching Level E status. Following are some reasons that some data elements could be shown as unavailable when the data actually had been collected.

• Data were under regional review.
• Data failed one of the quality checks and were being reviewed.
• Data failed one of the quality checks and were identified as anomalies.
• Data needed to be quality checked.

Some initially unavailable data were located and forwarded to the IMS during the course of this study. The key findings or observations from this detailed review are summarized in this chapter.

SPS-5 EXPERIMENTAL SITE STATUS

As of January 2000, 18 SPS-5 projects had been identified throughout North America. The full factorial of the original experiment design had been completely filled except for 2 cells. One project was required in a dry-freeze environment on a pavement in poor condition. This project would serve as a replicate to the Manitoba project (refer to table 8). Two projects were required for the wet-no-freeze climate with a pavement in fair condition. These missing projects were not believed to be critical because the factorial covered the range of environments and pavement conditions previously included in the experiment design.

Each SPS-5 project had 9 core test sections, and some had supplemental test sections built by the individual agency; 162 core test sections and 48 supplemental test sections were available, a total of 210 test sections. This number of test sections should provide excellent data for future studies for calibrating and validating design procedures.

The primary value of the supplemental sections is to serve as a direct comparison to the core test sections within that specific SPS-5 project. However, these supplemental sections also can be used in regional or national studies through the use of mechanistic analysis principles. Thus, efforts should be made to ensure that their construction and monitoring data are collected and entered in the IMS for future use.

An important issue in the experimental factorial is the different soil classifications. Half the projects were built over coarse-grained and half over fine-grained soils. This deviation from the original experiment design is not believed to be critical, but should be considered when analyzing the data to determine the main factor effects on performance.

DESIGN VERSUS ACTUAL CONSTRUCTION

Experimental design factors were compared to the actual values measured during construction. These include both the site condition factors and rehabilitation design features in the IMS database. Most SPS-5 sections followed the experiment design for the large majority of design factors. Overall, very few construction deviations were reported for the SPS-5 projects, with the exception of overlay thickness.

Most overlay thickness measurements deviated from the project’s experiment design requirements. However, none of the thickness data for the thin and thick overlays overlapped. Two projects (Maine and Manitoba) had significant thickness deviations from the planned overlay thickness. In both cases, the SHA was attempting to correct problems in the cross-slope of the pavement.

The other construction deviations were primarily related to the HMA mixtures. The percentage passing the number 4 sieve for the HMA mixture and air voids of the compacted mat exceeded the specified values for many test sections. These deviations were considered minor and should not be critical to the overall experiment. Other minor deviations were noted in the construction of these projects. For example, breakdowns of the hot mix facility and paving equipment caused delays in construction. Most of these types of deviations were considered minor and should not be critical to the overall experiment.

Three projects incorporated a control section that was overlaid during project construction of the other test sections: Colorado, Montana, and New Mexico. In each case, the condition of the existing pavement was believed to be a risk to the traveling public.

DATA AVAILABILITY AND COMPLETENESS

The data availability and completeness for the SPS-5 experiment were good overall, with the exception of two data elements. These two elements were materials testing and traffic data. Furthermore, some monitoring data still needed to be collected and/or checked to fill in the gaps of the time-history performance data for selected projects. Three projects (Minnesota, New Mexico, and Oklahoma) did not have sufficient time-history data for transverse profile to establish performance trends. The transverse profile should be measured at each of these sites. The reasons that data had not achieved Level E status need to be ascertained and the situation rectified before detailed analyses of the experiment can be completed. Most other data elements that had been collected at each site were at level E.

The SPS-5 data deficiencies are summarized below.

• Level E traffic data were not available for 50 percent of the SPS-5 sites. More important, traffic monitoring equipment had not been installed at 7 of the 18 sites.
• Materials test data were very deficient for most of the pavement and overlay materials and subgrade soils. The resilient modulus and other fundamental properties of these materials need to be measured and entered in the database if these sites are to be used for mechanistic studies. The testing program for many projects was still underway, and data were continually being forwarded to the RCOs for processing.
• The Missouri project had been constructed, but data were not yet in the database.

It is recommended that a significant effort be put forth to obtain these missing data. The following sections summarize the availability of each data element and its effect on following studies, such as for the 2002 Design Guide (NCHRP 1-37A). ⁽¹⁰⁾

Construction Reports/Data

The construction and deviation reports were extremely valuable in reviewing and explaining performance anomalies of individual test sections. Construction and deviation reports were available for all of projects except Missouri. The Missouri project was recently constructed, but the construction report was unavailable at the time of the detailed review.

Materials Data

The materials data were partially complete for all of the projects, with the exception of the new projects that had no test data at Level E. The laboratory material testing was divided into preconstruction and postconstruction tests. Preconstruction tests were to be performed on each existing pavement layer and the subgrade, while postconstruction tests were confined to the HMA overlay mixture.

Tables 10 through 12 summarized the availability of selected test data by material type for each project. Extensive test data were unavailable at the time of the data extraction–especially for the HMA materials. In general, the younger the project, the less testing had been completed. None of the resilient modulus, indirect tensile strength, and creep compliance tests had been completed on the HMA mixtures for any project.

Unavailable materials test data to determine the physical properties of the pavement and soils will be a significant limitation in the SPS-5 project’s use in mechanistic studies (such as NCHRP 1-37A). Completion of the materials testing program should be a high priority to ensure achieving the full benefit of the SPS-5 experiment. The RCOs recognized the importance of this data element and were pursuing these data. The materials testing program was still underway, and materials test data were being submitted to the RCOs periodically.

Climatic Data

No climatic data were missing. The SPS-5 experiment design called for a project to be located in one of four different climates (refer to tables 1 and 3). The climatic data are estimated from historical data from five nearby weather stations (virtual weather stations). The IMS contains monthly and average annual summary statistics for all 18 projects.

Traffic Data

The SPS-5 experiment design calls for continuous AVC monitoring, with WIM data collected at least 2 days of the year, as permitted by WIM scale operating divisions. Continuous AVC monitoring was defined as more than 300 AVC monitoring days in a given year. Table TRF_MONITOR_BASIC_INFO was examined to identify the SPS-5 records with WIM, AVC, and annual ESAL estimates.

Table 15 summarizes the amount of data for the SPS-5 sites and identifies those projects that had no traffic data at Level E. In summary, 14 (about 75 percent) of the SPS-5 sites had no traffic monitoring equipment at the site. Most of the older projects did have some traffic data, while most of the newer projects were missing the traffic data. All projects had an annual estimate of the number of ESALs; however, the reliability of these data was unknown for 15 of the 18 projects.

Performance Indicator Data

Several types of monitoring data are included in the LTPP IMS, including distresses (from both manual and photographic surveys), longitudinal profiles, transverse profiles, and deflection. Performance data are collected for both preconstruction and postconstruction time frames. All projects have preconstruction information on the surface condition of the pavement. Table 16 summarizes the number of postconstruction distress and other performance indicator measurements made at each project site; table 17 summarizes the average number of years between the surveys for each performance indicator.

Performance indicator monitoring data were available for all projects. However, the time interval between data collections was beginning to exceed the recommended frequency for a few of the projects, as noted below.

• Longitudinal profiles: New Mexico.
• Transverse profiles: Alabama and New Mexico.
• Distress surveys: None.
• Deflection surveys: None.

The RCOs had taken steps to collect some missing data or submit data that had been collected but not forwarded to the IMS. In summary, the amount of performance indicator data was good. The time-series data for each measure of performance will be a significant benefit for future studies on the design and performance of rehabilitated flexible pavements.

Friction Data

With few exceptions, friction surveys had not been performed on the SPS-5 projects. This testing, however, was not required at the time nor is it essential to the SPS-5 experiment. That friction data were missing should have no impact on future studies on structural behavior and performance of HMA overlays.

Summary

Table 28 summarizes the unavailable and limited data for the SPS-5 experiment, as of January 2000, in terms of the revised experimental factorial. Table 29 summarizes the limitations and action items to correct these deficiencies in each SPS-5 project. Every effort should be made to obtain these data elements.

EARLY PERFORMANCE TRENDS

Most SPS-5 projects were still relatively young. As of January 2000, less than 45 percent of the test sections had distress magnitudes that exceeded values believed necessary to complete meaningful comparisons. Based on preliminary statistical analyses and comparisons, age, surface preparation, and pre-existing surface condition were found to have an effect on performance indicators. The long-term performance trends could be significantly different from these early observations as more and more data are collected for these test sections.

The specific experimental expectations of the SPS-5 experiment were to determine the main effects and interactions of the following key design features:

• Amount of surface preparation before overlay.
• Material type used for the overlay.
• Overlay thickness.

These main effects and interactions were to be determined for each of the following subgrade and climatic conditions:

• Condition of pavement before overlay.
• Wet climates and dry climates.
• Freeze or cold climates and no-freeze or warm climates.
• Fine- and coarse-grained subgrade soils.

The following conclusions were drawn from the preliminary performance analyses conducted for this report.

• More fatigue cracking occurred on test sections placed in a climate with less precipitation but higher freeze indices.
• Less transverse cracking occurred on sections with intensive surface preparation than on sections with minimal surface preparation before overlay.
• The IRI values were found to be lower for: (1) overlays placed over pavements in the fair category and (2) when the existing surface was milled before overlay.

This evaluation has shown that several problems will limit results that can be obtained by the SPS-5 experiment, two of minor and two of major importance. Of minor importance are (2) the misinterpretation of the different distress types with time by the distress surveyors and (2) the measurement error of low levels of distress, which results in difficulties in interpreting performance trends and in determining the effects between the experiment factors.

The two problems that could result in major limitations to the value of SPS-5 are the materials test data and traffic data that are unavailable in the IMS. Without these data, the experimental objectives can be accomplished only in an empirical sense in terms of the general performance of different sections, but the development and calibration of mechanistic procedures will not be possible.

EXPECTATIONS FROM THE OWNER AGENCIES

At one national workshop, input was received from the States and Provinces on the SPS-5 experiment (April 27, 2000, in Newport, Rhode Island).⁽⁵⁾ Several agencies made presentations on the status of their individual SPS-5 project and on their expectations for the experiment. Panel discussions on the future direction and analysis of the SPS-5 data are summarized in this section.

In general, the owner agencies seem to be satisfied with the experiment and believed that it would produce valuable information on different design factors and features. Many agencies had been conducting or were planning their own analyses on their individual SPS-5 projects. Some of these analyses were yielding useful results; however, the agencies wanted a focus on implementation.

First and foremost, agencies wanted a research-quality database from SPS-5. Second, the agencies wanted to be able to determine the impacts of the rehabilitation design features on overlay performance and the effectiveness of the SPS-5 experiment design factors, such as:

• Condition of Existing Pavement–What effect does the condition of the existing pavement or types of distress (both in severity and extent) have on the performance of the rehabilitation strategy?
• HMA Overlay Thickness–How thick should the overlay be and what properties of the HMA have a significant effect on performance?
• Surface Preparation Intensity–How effective are certain surface preparation strategies in extending the service life of the overlay?

In addition to the structural design features, the agencies also wanted to know what major site condition factors influence the performance of HMA overlays over flexible pavements, including:

• Climate.
• Traffic volume and weights.
• Existing pavement condition.

Other expectations from the agencies included:

• Evaluation of existing performance prediction equations.
• Better rehabilitation design procedures.
• Better understanding of the distress mechanisms of HMA overlays.
• Validation and confirmation of pavement analysis methods.
• Calibration of mechanistic-empirical distress prediction models.
• Comparison of laboratory measured and field derived (back-calculated) material properties.

As to the future analysis plan for SPS-5, the agencies believed that it was worthwhile first to fill in the missing data–specifically, obtain traffic and materials test data. Some presenters at the SPS conference requested that fundamental studies be conducted to determine how the SPS-5 sections were responding to load and environmental stresses and loads. It also was suggested that an integrated analysis plan be developed for future research studies.

CAN THE SPS-5 EXPERIMENT MEET EXPECTATIONS?

This data review and evaluation of early performance trends showed that several significant data issues will limit the results that can be obtained from SPS-5. The missing traffic data and key material test data must be obtained before meaningful global analysis can be performed. A few SPS-5 sites had significant construction deviations. However, these construction deviations will not have a detrimental effect on the value of the experiment if the materials test data become available.

This does not mean that many important and useful findings cannot be obtained from the SPS-5 experiment even if no more traffic and materials data become available. Some important early trends were already identified that will be useful for the design and construction of HMA overlays of flexible pavements, even though all projects were less than 10 years old. Continual monitoring of the projects and test sections will provide valuable performance data that can be used in future studies to answer the questions asked by the owner agencies. Thus, it is concluded from this comprehensive study of the SPS-5 experiment that the expectations from the owner agencies and HMA industry can be met.

RECOMMENDATIONS FOR FUTURE ACTIVITIES

As stated in chapter 1, the key objective of the SPS-5 experiment is to determine the relative influence of different rehabilitation techniques on flexible pavement performance. It is believed that the experiment will be able to achieve this key objective with time. At the time of this report, the oldest SPS-5 project was just over 10 years (most were 5 to 7 years old), so many test sections had a moderate amount of surface distress, but only a few had been taken out service. The real benefit from this experiment will occur in the years ahead, as a greater percentage of test sections exhibit higher levels of distress–magnifying the effect of the experimental and other rehabilitation factors on performance.

The SPS-5 assessment report focused on the quality and completeness of SPS-5 construction and monitoring data, and on the adequacy of the experiment to achieve the original expectations and objectives. Some data were unavailable, but this will not significantly limit the value of the results. Detailed analysis of the effect of different design factors on performance was outside the scope of work for this study. Thus, future studies using SPS-5 experimental data should be planned and prioritized so they can be initiated as the SPS-5 projects exhibit higher levels of distress.

These future studies should be planned in two stages that focus on local and national expectations from SPS-5. The first stage would be a detailed assessment of each structural cell in the experiment to support local interests. The second would examine selected data elements to evaluate the effect of different structural features across the whole experiment. Both are discussed below.

Initial Stage–Analysis of Local Expectations or Individual Factorial Cells

A detailed evaluation of the projects within each cell should be completed as soon as some test sections begin to exhibit higher levels of at least one distress type. The purpose of these case studies is to:

• Resolve construction and monitoring anomalies and experimental cell differences for projects that changed cell locations from the original experiment design.
• Conduct comparative analyses of the individual test sections at each site including the supplemental test sections to identify differences in pavement performance and response. These comparative studies should include performance measures, material properties, and as-built conditions.
• Determine the effect of any construction difficulties and problems and material noncompliance issues with the SPS-5 project specifications, if any, on pavement performance and response.
• Determine the effect of any construction difficulties and problems and material noncompliance issues with the SPS-5 project specifications, if any, on pavement performance and response.
• Develop findings on comparisons made between the projects and test sections and prepare a case study report that can be used for the national studies.

Second Stage–Analysis of National Expectations or Experimental Findings

The second-stage analyses should not be pursued until the first stage is complete. It is expected that the analyses performed as a part of the second stage can be coordinated with the Strategic Plan for LTPP Data Analysis. The SPS-5 experiment can contribute to the following specific analyses outlined in the strategic plan.

• Develop relationships to enable interchangeable use of laboratory- and field-derived material properties (Strategic Plan No. 2B).
• Establish procedures for determining as-built material properties (Strategic Plan No. 2C).
• Identify quantitative information on the performance impact of different levels of material variability and quality (Strategic Plan No. 2D).
• Estimate material design parameters from other materials data (Strategic Plan No. 2E).
• Quantify information as to the relationship between as-designed and as-built material characteristics (Strategic Plan No. 2F).
• Develop recommendations for collection of climatic data collection to adequately predict pavement performance (Strategic Plan No. 3D).
• Develop models relating functional and structural performance (Strategic Plan No. 4C).
• Calibrate relationships or transfer functions between pavement response and individual distress types (Strategic Plan No. 5C).
• Identify quantitative information on the impact of design features on measured pavement responses (deflection, load-transfer, strains, etc.) (Strategic Plan No. 7B).
• Develop guidelines for the selection of pavement design features (Strategic Plan No. 7C).

A description of some future studies that could be pursued at the national level using all of the SPS-5 experimental data are summarized in tables 30 through 41. The future research studies were prepared based on the discussions with and presentations from SHA personnel at the various SPS conferences that were held in 1999 and 2000. These future analysis objectives are believed to be achievable from data collected within the SPS-5 experiment and have been subdivided into two categories. The first category includes the analysis objectives that are related to the main factors of the SPS-5 experiment; objectives in the second are related to other experimental factors.

The following second-stage analysis objectives are recommended for the SPS-5 experiment; they are presented in more detail in tables 30 through 41.

Future Analysis Objectives Related to Main Experimental Factors (by table number)

Future Analysis Objectives Related to Other Experimental Factors (by table number)

Data-Collection Efforts

It is recommended that the following data-collection efforts be emphasized in the future in support of the second-stage analyses:

• Collect routine data:
  • WIM and AVC traffic monitoring should receive close attention. Steps should be taken to ensure the routine data-collection guidelines are followed for both AVC and WIM.
  • Resolve irregular distress measurements over time for each SPS-5 section.
• Collect new data:
  • Dynamic modulus of asphalt concrete to predict fatigue and other load-related distresses.
  • Indirect tensile creep tests to predict low temperature cracking.
  • Performance-grade of the asphalt binder in the HMA overlay mixtures.
  • Coring along the cracks in HMA to determine the initiation of the crack and direction of its propagation (top-down or bottom-up cracking).

Note: The future research topics or objectives that follow for the individual main or primary factors of the experiment are included as individual project objective statements.

* The initial IRI values (longitudinal profile measured within 6 months of construction, assuming reasonable performance of the test sections) are needed to obtain the full benefit of the research study. The initial IRI values will need to be predicted from the time series data for some of the test sections or SPS-5 projects.

* GPS–General Pavement Studies

Note: One component needed to improve the accuracy of the results is comparable time measurements of deflection data and distress surveys. In addition, the resilient modulus of the HMA mixtures will be needed to improve the universal application of the results.

* The probability of success will increase greatly if cores are performed as a part of special interim studies and all forensic studies to confirm the location of where the fatigue cracks initiated.

* The probability of success will increase greatly if trenches are performed as a part of all forensic studies to confirm any subgrade distortion.