U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-06-121
Date: November 2006

Long-Term Pavement Performance (LTPP) Data Analysis Support: National Pooled Fund Study TPF-5(013)

Executive Summary

Pavements subjected to frost effects have different service lives than do similar pavements with no exposure to frost; however, limited national research is available quantifying the effect frost has on pavement performance let alone the costs resulting from reduced service life. This study provides some insight into pavement performance and service life, considering conditions of both deep-frost and moderate-frost depth with multiple freeze-thaw cycles (FTC).

The study included a review of all available relevant literature to provide guidance and to support the project work. Literature directly related to this investigation was quite limited, with most of the literature regarding frost effects dealing with quantifying the change in material properties and performance characterization on particular projects. In addition, relatively limited information was found on modeling pavement performance in frost areas using Long-Term Pavement Performance (LTPP) data. State Highway Agency (SHA) Web sites were also reviewed to accumulate reports documenting studies of frost mitigation techniques.

To study pavement performance in the various frost settings, models were developed using multivariate regression analysis. LTPP data from General Pavement Study (GPS) projects 1, 2, 3, and 6 as well as Specific Pavement Study (SPS) experiments 1, 2, and 8 were used to generate the models. Data from more than 520 test sections were used in developing the prediction models for flexible pavements, while approximately 270 test sections were used for rigid pavement modeling. More than 20 models were developed to represent the rate of pavement deterioration with age unique to environmental regions. Of these, nine models were selected that were determined to best predict basic pavement trends with time. A summary of these nine models can be found in Table 1.

Using the models presented above, statistical comparisons of pavement performance were made for the following five climatic region scenarios:

All of the models (with the exception of flexible pavement roughness) predicted significantly different performance, at 95 percent confidence, between two or more of the climatic scenarios.

Table 1. List of models and basic logistic and regression statistics.

Model

Pavement

Model
Type

Logistic
Cutoff
Probability

Percent
Correct

R- Squared

Number of
Observations

 

Roughness

 

Flexible

Regression
(shifted)

 

NA

 

NA

 

0.78

 

4544

Roughness

Rigid

Regression

NA

NA

0.78

2652

Rut Depth

Flexible

Regression

NA

NA

0.45

1966

Faulting

Rigid

Regression

NA

NA

0.47

1384

Fatigue and
Wheelpath
Cracking

Flexible- deduct

Logistic

0.7

72.6

NA

1977

Regression

NA

NA

0.63

1486

Fatigue and
Wheelpath
Cracking

Flexible- percent

Logistic

0.7

72.6

NA

1977

Regression

NA

NA

0.49

1481

Transverse
Cracking

Flexible

Logistic

0.7

78.4

NA

1920

Regression

NA

NA

0.71

1077

Longitudinal
Cracking

Rigid

Logistic

0.55

63.5

NA

475

Regression

NA

NA

0.38

240

Transverse
Cracking

Rigid

Logistic

0.6

63.5

NA

489

Regression

NA

NA

0.54

228

To gain an understanding of state design practices, a questionnaire was developed and sent to the pooled fund State participants. Basic information on standard roadway sections including structural design for given scenarios, standard specifications, and test procedures were requested.

Responses to the survey revealed that there is a large variation in the roadway section for similar design situations. However, most of the States in the study experiencing deep frost did include a construction specification requiring additional surfacing or the
replacement of frost-susceptible soils with frost-free surfacing for a depth of 1 to 2 meters
(m) (3 to 6 feet (ft)).

Agency responses also revealed that the use of Superpave mix design procedures has, to a large extent, eliminated local adaptations in mix designs and specifications that might have provided improved performance in areas of deep frost penetration or numerous FTCs, or both. The Superpave® mix design procedure does not differentiate between mix designs where pavements will be exposed to numerous FTCs and those that will experience little or no FTCs. Because many SHAs are in the process of adopting the Superpave binder specifications as well as the mix design procedures, local adaptations
as far as mix designs and specifications were not found that would indicate improved pavement performance in areas with either deep frost penetration or numerous FTCs.

An additional objective of the study involved evaluating the costs associated with performance differences in the various environments. Life cycle cost analysis (LCCA) was used to evaluate pavement costs in the various climatic settings because it produces comparable results (i.e., equivalent uniform annual costs). Comparisons were made using both deterministic and probabilistic methods of LCCA.

Standard flexible pavement sections were developed based on the 1993 AASHTO Guide for Design of Pavement Structures design procedures(1) and using input variables from the questionnaire. A cost comparison was performed using this standard section for all environmental zones; therefore, the initial and rehabilitation costs were constant for all regions. Cost differences were the result of changes in treatment timing because of performance variations between the regions. Predictions from the models were used to determine treatment timing for each climatic scenario.

To account for local adaptations used to mitigate damage associated with freezing and thawing climates, an additional cost evaluation was performed in which the initial costs of the deep- and moderate-freeze regions included extra frost-free material (i.e., unbound base) to obtain a pavement structure with a total depth of 1 m (3 ft). Based on responses from the participating Agencies, this is a typical frost-free depth for many SHAs experiencing 1 to 1.5 m (3 to 4 ft) of frost penetration. The initial construction costs for the no-freeze region did not include additional base material.

Using the standard section for all regions resulted in costs that were not significantly different for the five climatic scenarios. When the cost of additional surfacing was considered, the life cycle costs in the no-freeze region were significantly lower than in the deep- and moderate-freeze regions.

Consideration was given to the use of the developed performance models in the implementation of mechanistic-empirical (M-E) design procedures. The National Cooperative Highway Research Program (NCHRP) Guide for Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures,(2) Project 1-37A final report, was developed using damage models that represent average pavement damage trends for the entire United States. The models developed in this project can be used to predict average rutting or fatigue cracking trends for a specific regional or statewide environment. In turn, these estimates can be used in the iteration and verification process described in the NCHRP 1-37A Guide design procedure to determine if modified calibration factors are required in the design program for the specific environment. An example of how to use the models from this study to provide regional calibration is described in the report.

The models developed for this project will also be useful in pavement management applications. The pavement distress trend models developed in this study can be used to provide general pavement deterioration trends for a specific environment. These trends can be used to develop a family of curves for use in a pavement management system (PMS) where an SHA or local agency does not have sufficient data to develop those curves.

Previous | Contents | Next

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101