Pavement Performance Measures and Forecasting and The Effects of Maintenance and Rehabilitation Strategy on Treatment Effectiveness (Revised)

CHAPTER 9. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS

This chapter presents a summary of the studies described in chapters 2 through 8, followed by significant conclusions reached organized by topic, and finally, significant recommendations also organized by topic.

SUMMARY

For the benefit of the reader, this section is divided into several paragraphs. Each paragraph includes the summary of a particular chapter of the final report.

At the outset, the research team conducted a comprehensive review of the state of the practice of various State transportation departments with regard to several aspects of pavement condition measures, pavement condition and distress data analyses, and treatment selection. The review also included previous related studies that were conducted using the LTPP database. The detailed literature review can be found in chapter 2. The topics covered include the following:

• Pavement distress severity levels.
  
  
• Pavement condition and distress descriptions.
  
  
• Pavement performance modeling and treatment benefit calculations.
  
  
• Treatment type and time selection.
  
  
• Preservation costs and LCCA.
  
  
• Effectiveness of pavement treatments at the project and network levels.
  
  
• LTPP Program, its objectives, and the SPS and GPS test sections.
  
  
• Previous findings regarding the impacts of pavement treatments and various design factors on pavement performance.

At the same time, the research team downloaded from the six data volumes housed in the LTPP database Standard Data Release 28.0 the data elements of the more than 2,500 test sections included in the LTPP Program. The data were organized in a special format and readied for analyses. In addition, the pavement management databases from three State transportation departments (CDOT, WSDOT, and LADOTD) were requested and received. From each database, several pavement projects that were subjected to certain treatments in the past were identified, and their data were downloaded from the respective databases and formatted for analyses. The details of the LTPP and State data mining are in chapter 4.

Based on the literature review, and general review of the LTPP data, dual pavement condition rating systems were developed based on the pavement function and its structural integrity. One system was based on three CSs, and the other was based on five CSs. For each pavement section, the functional CS was based on ride quality in terms of IR) and safety in terms of rut depth. The functional CS was expressed in term of RFP in years for the pavement to reach the prespecified threshold value for IRI or rut depth. The structural CS was based on RSP in years for the pavement section to reach the threshold values in terms of alligator, transverse, or longitudinal cracking or rut depth. The rating system for each CS consisted of numerical classification, color coding, range of RFP and RSP, and the average cost per 0.1 mi (0.16 km) of preserving the pavement. Further, based on the literature review and common engineering practice, individual threshold values for IRI, rut depth, alligator, transverse, and longitudinal cracking were recommended and used in the analyses of the LTPP and State data. Details of the two pavement rating systems and the threshold values used in the analyses are in chapter 3 of this report.

The performance of each of the LTPP flexible pavement test sections included in the SPS-1, -3, and -5, and GPS-6 experiments was analyzed. In the analyses, the available before and after treatments time-series pavement condition (IRI and rut depth) and distress (rut depth, alligator, longitudinal, and transverse cracking) data were used. The data were modeled as a function of time using the proper mathematical function form (power function for rut depth, exponential for IRI, and logistic for cracking). Results of the analyses were expressed in terms of RFP for IRI, RFP/RSP for rut depth, and RSP for each cracking type. Thus, for each test section, two RFPs and four RSPs were calculated. These values were used to assess the impacts of regional climatic and design factors on pavement performance. In addition, the LTPP data on the flexible pavement test sections were used to develop a new and novel method to estimate the pavement performance based on a single data point. The ORCSE method was developed to be applied to pavement sections that had experienced less frequent data collection and/or pavement sections that might not yet have sufficient data records for modeling owing to age. The analyses and the results of the analyses of the SPS-1, -3, and -5, and GPS-6 test sections and the development and validation of the ORCSE method are detailed in chapter 5.

The performance of each of the LTPP rigid pavement test sections included in the SPS-2, -4, -6, and -7, and GPS-7 experiments was also analyzed. In the analyses, the available before and after treatments time-series pavement condition (IRI and rut depth) and distress (rut depth, alligator, longitudinal, and transverse cracking) data were used. The intent was to study the impact of each design variable on pavement performance. When the data were divided into various groups based on separation of variables, the number of test sections under each design variable was statistically insignificant (i.e., for some variables, there was only one or no test section). Therefore, the impacts of the design variables on pavement performance were not analyzed or discussed any further. Rather, the data were used to study the impacts of the conditions in climatic regions on pavement performance. Results of those analyses are detailed in chapter 6.

The measured flexible and rigid pavement deflection data were organized and analyzed. The analyses were conducted to accomplish the following two objectives:

• Determine whether the measured flexible pavement deflection data could be used to estimate the time to initiate bottom-up cracking before those cracks appeared on the pavement surface and whether those data could be used in the RFP and RSP algorithms.
  
  
• Study the LTE using the measured rigid pavement deflection data and its impact on ride quality in terms of IRI.

At the onset, it was envisioned that measured pavement deflections and/or LTE data would be correlated with pavement condition or distress, and hence, they could be used as an early warning of pending deterioration before it became visible on the pavement surface. For flexible pavements, the measured pavement deflections are a function of the pavement temperatures. Therefore, the accuracy of the existing temperature adjustment methods were reviewed and scrutinized using the measured LTPP deflection data. Based on the results, a new global temperature correction methodology was developed that could be applied to all deflection sensors and all climatic regions. Results of the analyses of the measured deflection data and the methodology and algorithms of the newly developed temperature correction method are detailed in chapter 7.

The pavement condition and distress databases of three pavement networks were requested and received from CDOT, LADOTD, and WSDOT. Each database was searched and pavement projects that received one of the following five treatment types were identified:

• Thin overlay (≤ 2.5 inches (63.5 mm)).
  
  
• Thick overlay (> 2.5 inches (63.5 mm)).
  
  
• Thin mill and fill (≤ 2.5 inches (63.5 mm)).
  
  
• Thick mill and fill (> 2.5 inches (63.5 mm)).
  
  
• Single chip seal.

The research team analyzed the pavement condition and distress data measured before and after treatment of each 0.1-mi (0.16-km)-long pavement segment along each selected pavement project that was treated using one of these five treatments. The main objective of the analyses was to calculate the treatment benefits in terms of the following:

• RFPs and RSPs before and after treatment.
  
  
• CFP and CSP resulting from the treatment.
  
  
• FCROP and SCROP.

For each treatment type, the weighted average treatment benefits, in terms of each pavement condition and distress type, were then calculated. The results were submitted to FHWA and are available from the LTPP Customer Support Services.⁽⁷⁹⁾ The weighted average treatment benefits were then compared with the weighted average treatment benefits of the LTPP test sections. Results of the comparison are detailed in chapter 8.

CONCLUSIONS

Based on the review of various national and international papers published in various journals and mainly on the results of the analyses, various conclusions (detailed in each chapter) were drawn that cover various study-related topics. For convenience, only significant conclusions were selected and listed by topic in the follow subsections.

Pavement Performance Measures

The following significant conclusions were drawn regarding pavement performance measures:

• The pavement cracking data were typically collected and stored based on three severity levels (low, medium, and high). For most cases, the problem was that the data could not be analyzed per severity level because of their excessive variability from one year to the next. Analyses of the cracking data based on the sum of all severity levels were proven to overcome the problem.
  
  
• For pavement projects received the same treatment type, T²Ms were developed to display the distribution of the pavement conditions along the project before and after treatment. The data in the T²Ms were used to estimate the benefits of the various treatments.
  
  
• Pavement condition rating should be based on current conditions and distresses as well as the pavement’s rates of deterioration.
  
  
• The three- and five-level dual pavement condition rating systems developed in this study were useful and were equally applied to both State and LTPP data. The systems were flexible and could be easily tailored to fit the needs and constraints of any road agency.
  
  
• The estimated average cost of pavement preservation for each level of the dual pavement rating system could be used in the lifecycle cost analyses and in strategy optimization.
  
  
• Threshold values were provided for calculation of RFP and RSP. The values were based on minimum level of service to the user (functional), and loss of structural integrity (structural).

Flexible Pavements

The following significant conclusions were drawn regarding flexible pavements:

• The conditions in WF regions had significant adverse impacts on pavement performance in terms of IRI, rut depth, and cracking.
  
  
• Drainable bases decreased the impacts of the conditions in WF regions on pavement performance. The pavement performance owing to drainable bases was slightly better than that owing to increasing the AC thickness from 4 to 7 inches (102 to 178 mm). Therefore, the use of drainable bases in the WF region was cost effective.
  
  
• The inclusion of drainable bases in the DF and DNF regions did not affect pavement performance in terms of RFP or RSP. This was expected because the volume and frequency of available water was low. Furthermore, most rainfall occurred over short periods of time with most water running off the surface and not penetrating the pavement layers.
  
  
• Increasing the thickness of the AC layer from 4 to 7 inches (102 to 178 mm) increased the frost protection of the lower layers, and hence decreased the impacts of the climatic conditions in the WF region on pavement performance. However, this option was not a cost-effective one.
  
  
• The climatic conditions in the WNF, DF, and (DNF) regions did not affect the pavement performance in terms of rutting potential and IRI.
  
  
• The climatic conditions in the DF region had more adverse effects on cracking potential than those in the DNF region. This was mainly attributed to the higher oxidation (aging) potential of the AC layer in the DF region.
  
  
• The thin overlay treatment improved the pavement performance of the SPS-3 test sections in terms of IRI and rut depth in all climatic regions except the DNF region. This could be related to construction issues and, perhaps, the relatively high solar radiation (accelerated oxidation/aging) in the DNF region.
  
  
• In general, the thin overlay treatment did not improve the pavement performance of the SPS-3 test sections in terms of alligator, longitudinal, and transverse cracking. This is mainly because of the high rate of reflective cracking. Immediately after treatment, all cracks were hidden by the thin overlay. However, one or few years later, most cracks were reflected through the overlay, which implied relatively high rate of deterioration and hence short RSP. The exception was in the DNF region where the two test sections showed an increase of 12years in the average RSP compared with the one control section. This oddity was mainly attributable to the limited number of sections. That is, the conclusion was not reliable because of the limited number of test sections and control sections.
  
  
• The slurry seal treatment improved the pavement performance of the SPS-3 test sections in terms of IRI and rut depth but did not have much impact on alligator, longitudinal, and transverse cracking.
  
  
• Crack sealing appeared to improve pavement performance of the SPS-3 test sections in terms of rutting. However, it did not improve the pavement performance in terms of cracking.
  
  
• Aggregate seal coats appeared to improve the pavement performance of the SPS-3 test sections in all climatic regions in terms of IRI, rut depth, and cracking.
  
  
• In general, the worse the pavement conditions were before treatment, the shorter the benefits of treatments were in terms of RFP and/or RSP.
  
  
• On average, the impact of 2- and 4-inch (51- and 102-mm)-thick virgin or recycled AC overlays on pavement performance of the SPS-5 test sections was almost the same.
  
  
• The 2-inch (51-mm)-thick AC overlay (virgin or recycled mix) did not provide long-term remediation of transverse cracking. The cracks in the lower pavement structure typically reflect through the overlay in few years.
  
  
• On average, SLE in terms of alligator crack for a flexible pavement structure as a result of application of a 4-inch (102-mm) AC overlay was slightly better than that which resulted from application of a 2-inch (51-mm) overlay.
  
  
• In each climatic region, the impacts of the thin and thick overlay or thin and thick mill-and-fill treatments on IRI and rut depths were almost the same. This was expected because good quality construction can decrease the pavement surface roughness substantially regardless of the overlay thickness and because most pavement rutting occurs early in the pavement life and can be removed during the treatment.

ORCSE Method

The following significant conclusions were drawn regarding the ORCSE method:

• The ORCSE method was successful in predicting the CSs of the test sections based on RFP and RSP. For older test sections, the average estimation error for the most critical CSs was less than 1 percent.
  
  
• The ORCSE method would have significant potential benefit for local roadway owners as well as State transportation department managers when only two or fewer data points were available.

Rigid and Composite Pavements

Significant conclusions regarding rigid and composite pavements were the following:

• The conditions in the WF region had more damaging impacts on the performance of the SPS-4 test sections in terms of transverse cracking than on test sections located in the WNF, DF, and DNF regions. This was expected owing to the combined effects of subfreezing temperatures and moisture.
  
  
• On average, the majority of the SPS-2 test sections located in the WNF region performed worse relative to longitudinal cracking than those in the DNF region. This was mainly due to the impact of excessive moisture on pavement performance.
  
  
• On average, in terms of IRI, joint and crack sealing treatment had a positive impact on the performance of the SPS-4 test sections located in the WNF region, no impact in the WF region, and negative impact in the DF and DNF regions.
  
  
• Joint and crack sealing was effective in the WF region and not effective in the other three climatic regions, and joint undersealing was not effective in any region.
  
  
• The performance of treated SPS-6 test sections in terms of IRI was independent of the climatic region and pavement type. It was also independent of treatment type in terms of rut depth.
  
  
• The alligator cracking data in the SPS-6 database were highly likely an advanced form of top-down fatigue cracking. (Top-down cracks are fatigue cracks that initiate at the pavement surface and, over time, propagate downward.) The short transverse and longitudinal cracks resembled the traditional bottom-up alligator cracking pattern.
  
  
• The performance of the test sections in terms of longitudinal cracking was worse after subjecting the section to any of the seven analyzed treatment types.
  
  
• Minimum and maximum pavement restoration with no AC overlay treatments did not improve the performance of the JRCP test sections.
  
  
• The IRI-based performance of the treated CRCPs (SPS-7) test sections was independent of the eight treatment types and the two climatic regions (WF and WNF).
  
  
• The performance of the JPCP test sections subjected to a 3-inch (76-mm) concrete overlay with milling (703) or with shot blasting (704) treatments was lower than the performance of the other JPCP test sections subjected to the other six treatments.
  
  
• None of the eight applied treatments were effective in treating transverse cracking problems of the CRCP test sections.

Deflection

The following significant conclusions were drawn regarding deflection:

• The LTPP deflection data collected for the SMP test sections did not support the use of the AI temperature adjustment procedure for the measured deflection data.
  
  
• The newly developed global flexible pavement deflection temperature adjustment algorithm was applicable to all deflection sensors and in all climatic regions.
  
  
• The LTPP deflection data collected for the SMP test sections did not correlate with the measured pavement condition and distress data. Therefore, the data were not included in the RFP and RSP algorithms.
  
  
• The LTPP deflection and LTE data collected for the SPS-2 test sections in Arizona, Colorado, Delaware, and Michigan did not support modeling of the data as a function of time.

State Data

The following significant conclusions were drawn regarding State data:

• The weighted average benefits of each of five treatment types (thin overlay (≤2.5inches (63.5 mm)), thick overlay (> 2.5 inches (63.5 mm)), thin mill and fill (≤ 2.5 inches (63.5mm)), thick mill and fill (> 2.5 inches (63.5 mm)), and single chip seal) with regard to each pavement condition and distress type, obtained from the analyses of the LTPP data, were similar to the benefits obtained from the three sets of State data.
  
  
• The methodologies described in this report for the analyses of the LTPP pavement condition and distress data applied to the State data.
  
  
• The three-level (poor, fair, and good) and five-level (very poor, poor, fair, good, and very good) pavement rating systems developed in this study, and presented in chapter 4 based on the time-series pavement condition and distress, were equally applicable to the LTPP and the State data.
  
  
• The average variability in the pavement condition and distress data over time for the LTPP test sections was almost the same as the average variability of the State-measured data.
  
  
• The percent of the LTPP test sections that were excluded from the analyses because of an inadequate number of data points or because of improving pavement condition and/or distress over time without the application of treatments was similar to the percent of the 0.1-mi (0.16-km)-long pavement segments of a given pavement project that was excluded from the analyses for the same reasons.

RECOMMENDATIONS

Based on the results of the LTPP and State data analyses and the conclusions listed in the previous section, various recommendations were developed. For convenience, the significant recommendations are listed by topic in the following subsections.

Performance Measures

The following recommendations were developed regarding performance measures:

• The sum of crack lengths or crack areas of all severity levels should be used to model the data as a function of time.
  
  
• Accurate pavement planning and management decisions should be based on the pavement conditions and rates of deterioration.
  
  
• The three and/or the five-level rating systems should be adopted by FHWA and submitted to AASHTO for approval.
  
  
• The threshold values used in this study should be adopted or similar ones should be developed by highway owners to estimate RFP and RSP of the various pavement sections.
  
  
• Each highway agency should develop the average cost of pavement preservation for each RFP and RSP level of the dual pavement rating systems using their own cost records.
  
  
• LCCA should be performed at the project level and strategy optimization at the network level to improve the overall cost effectiveness of the pavement management application.
  
  
• The T²M procedure should be adopted and used by road owners to assess treatment effectiveness and to select the optimum treatment time.
  
  
• The dual pavement rating system described in chapter 3 should be used in future analyses and assessments of the benefits of pavement rehabilitation and/or maintenance treatments.

Flexible Pavements

The following recommendations were developed regarding flexible pavements:

• Drainable bases should be constructed to enhance the performance of pavement sections located in the WF region.
  
  
• For future study, the control or linked test sections should be selected to border the regular test sections in question, and their history should be included in the database. This would eliminate unnecessary variability.
  
  
• The pavement condition and distress data should be measured before and after treatments. The quality control data for project acceptance should be included in the PMS database.
  
  
• The frequency of pavement condition and distress data collection should be a function of treatment type. Treatments having short TL should be surveyed more often than long-life treatments.

ORCSE Method

The following recommendation was developed regarding the ORCSE method:

• The ORCSE method should be expanded to address other conditions and distresses as well as be applied to a wider range of LTPP and State transportation department data to further verify its successful prediction of the pavement CSs.

Deflection

The following recommendations were developed regarding deflection:

• The new global flexible pavement deflection temperature adjustment algorithm should be adopted and used to adjust the measured pavement deflection to a standard temperature of 70 ºF (21 ºC).
  
  
• The estimation of the AC modulus at the standard temperature of 70 ºF (21 ºC) could be accomplished using one of the following two procedures:
  
  
  • Adjust the measured deflection data to 70 ºF (21 ºC) using the newly developed global model and then backcalculate the modulus.
    
    
  • Conduct a minimum of three FWD tests at temperatures above and below the standard temperature of 70 ºF (21 ºC) and measure the pavement deflections. Backcalculate the AC modulus value for each of the three FWD tests. Plot the modulus as a function of the measured pavement temperature and estimate the AC modulus at 70 ºF (21 ºC) from the graph.
    
    
• For new LTPP experiments or road tests, FWD tests should be conducted before and after treatment.

State Data

The following recommendations were developed regarding State data:

• The dual pavement condition rating systems should be adopted by FHWA, AASHTO, and the State transportation departments. This would unify the analyses of pavement performance nationwide.
  
  
• The benchmark benefit values of the five treatment types included in this study should be expanded to include additional pavement treatments.
  
  
• The treatment benefits described in this report that were found using the LTPP data, could be used as benchmark values for the national practice. State transportation departments could use such data to do the following:
  
  
  • Gauge the effectiveness of their current practices using similar analyses.
    
    
  • Conduct lifecycle cost analyses of various treatment alternatives to optimize the pavement rehabilitation and treatment strategy at the network level.

Future Studies

The following actions are recommended regarding future studies:

• Initiate studies to produce a national catalog regarding the service life of various treatment types as a function of pretreatment pavement conditions and distress. The studies should be based mainly on the LTPP data and applied to some State data.
  
  
• Initiate studies to establish automated data collection processes, quality control procedures, and data storage to minimize the impact of subjective factors and human errors on pavement conditions and distress.
  
  
• Initiate studies to scrutinize the newly developed self-powered wireless Pico sensors (10¹² mm) that can be embedded in pavement and transportation infrastructures to measure their performance in terms of cracking and induced stresses and strains.
  
  
• Explore the accuracy of the newly developed chemical sensors that can be included in concrete and asphalt mixes to measure future pavement conditions and distresses. This would eliminate the need for traditional data collection.
  
  
• Develop short course materials with examples to train engineers and staff of State transportation departments to use the MATLAB® computer program for the analyses of pavement condition and distress and to emphasize the benefits of the LTPP Program.