A Mechanistic–Empirical Model for Top–Down Cracking of Asphalt Pavements Layers (2018)

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.

101 CHAPTER 5. INTERPRETATIONS, APPRAISAL, AND APPLICATIONS Introduction This project developed an asphalt pavement top-down cracking design program which is based on calibrating predicted top-down cracking lives to observed top-down cracking distress in asphalt pavements. This program is written to be a subprogram to the AASHTOWare Pavement ME Design software. It uses the same types of inputs as the Pavement ME Design requires and additional data that is needed for a ME top-down cracking model. These include the material properties and aging characteristics of the asphalt layers which are directly measured from field asphalt mixtures and pavements. As with the Pavement ME Design, the top-down cracking design subprogram utilizes the traffic to be input as an axle load spectrum. The weather data in 65 counties of 29 states with identified top-down cracking are assembled in this project and incorporated into the subprogram. All data may be input at one of the three levels (Level 1, 2, and 3) selected by the user. In addition to the design model, the project has also developed a similar model calibration program that is used to create a set of calibration coefficients. These calibration coefficients are used in the design program for specific local or regional conditions and practices. The design program is arranged as a single mechanistic crack growth model with different sets of calibration coefficients for top-down cracking in different climatic zones. The Model Development Process The application of mechanics to the prediction of top-down cracking through asphalt pavement layers involves a number of very detailed steps including the use of finite element analysis of the J-integral and the modeling of those results with an Artificial Neural Network (ANN) algorithm in order to speed up the computational time. The computational task of determining the J-integral and fracture properties under a variety of loading conditions and temperatures, including traffic and thermal stresses must be done rapidly in order to make the Pavement ME Top-Down Cracking Design Model a practical tool for design. A third part is to generate accurate weather characteristics that can be used to provide realistic material properties and stresses throughout each day and over the observed service life of an asphalt pavement. A fourth part of the assembly of this model is to develop a consistent means of describing the distress that was observed in the field. The fifth part was to devise a means to relate the predicted distress to the observed distress in a simple way, and produce predicted distress that matched well with what was actually observed in the field. A discussion of these five steps follows. Mechanistic Prediction of Top-Down Cracking The model was selected for the Pavement ME Top-Down Cracking Design Program based on several factors. One of the most important of the factors was the speed with which daily crack growth could be computed to facilitate consideration of several material, thickness and age options. This led to the decision to use the ANN algorithms to compute both the changing modulus of the asphalt layer mixture due to aging and the growth of the cracks, and to develop

102 mechanistic data with which to “train” the ANN algorithms for crack growth. This was done by calculating a large set of J-integral data for a variety of material and pavement structures using a two-dimensional finite element approach. When the calculated results were compared to the correct answers generated by a full three-dimensional set of computational results, the errors were acceptably small as discussed in Chapter 4. As shown in Chapter 4, the ANN models fit all of computed data bases very well. Neither the two- nor the three-dimensional finite element analysis was used within the design program because of the long computational time that each requires. The ANN predicted J-integral is then used in the modified Paris’ law to calculate the “pre-fatigue life”. It is called “pre-fatigue life” in this study as it is the time when a single longitudinal crack in low severity develops. After that, multiple longitudinal cracks are connected which are rated as fatigue cracking. Asphalt Pavement Layer Material Properties The ANN algorithms were also used for generating the material properties (i.e. the unaged modulus) of an asphalt pavement layer material as it responds to traffic and thermal stresses. It was also necessary to use a well constituted and widely available data base of asphalt material properties to represent these properties. The fracture properties of an asphalt mixture depend on simple performance-based properties. Inputting binder properties, aggregate gradation, volumetric composition of the mix and frequency of loading and temperature into these ANN algorithms, very satisfactory mixture properties over a wide range of loading times and temperatures can be generated quickly. Inputting binder properties, aggregate gradation, volumetric composition and relaxation modulus of the mixture into the fracture coefficient prediction models, the fracture properties can be obtained quickly. A catalog of the fracture properties for different asphalt mixtures obtained in the laboratory and field is provided in Appendix F. The aging properties of an asphalt mixture are generated based on the kinetics- based approach and a catalog of aging properties is presented in Appendix G. Weather Data and Temperature Prediction Accurate temperature prediction is a key to making accurate predictions of thermal crack growth. Comparisons between the temperature predictions and actual temperatures measured in the field demonstrated the need for a higher degree of accuracy in calculating the temperature within the asphalt pavement layer. The temperature model mentioned in Chapter 4 was incorporated into the Pavement ME Top-Down Cracking Design Program. The temperatures calculated by this program and those measured in the field rarely differed by more than 2° C. A complete set of weather data were assembled from databases that are available to the public for about 65 different locations within the United States for use in the design program. Consistent Description of Top-Down Cracking The S-shaped curve for the accumulating extent of top-down cracking that was adopted in this project matches the pattern that is observed in the progressive development of many kinds of distress. Based on field observation, longitudinal wheelpath cracking mainly appears at the low severity level. The S-shaped curve of accumulating crack length is plotted as the total length of longitudinal wheelpath cracks in the three severity levels versus the service time. This curve is

103 defined by two parameters: ρ, the scale parameter and β, the shape parameter. The scale parameter is the number of days required for the percentage of longitudinal cracks to reach 36.8 percent of the maximum length of the cracks in the pavement surface. The shape parameter determines how steep the growth of the curve is as it reaches the 36.8 percent mark. This method allows a simple, consistent and comprehensive description of the distress history of an asphalt pavement. It also made possible the task of calibrating the calculated top- down cracking lives due to traffic and thermal stresses to the field observations of top-down cracking. Calibrated Results Compared with Observed Field Data The detailed calibration process is provided in Appendices H and I. The scale and shape parameters ρ and β that are fit to the field observations and the computation of the number of days to reach the critical crack depth are discussed in Appendix H. The method to determine the calibration coefficients for ρ and β and the plots of predicted values versus observed values of accumulating distress are presented in Appendix I. Determination of Number of Days for Critical Crack Depth It is observed from the LTPP data that most of the longitudinal cracking in the wheelpath is in the low severity level. Based on the LTPP definition of low, medium and high longitudinal cracking severity levels and the empirical relationship between crack width and crack depth for the low severity crack that was developed previously, 15 mm is selected as the critical crack depth, which corresponds to the boundary crack depth between low and medium severity levels. The crack depth when the crack makes its first appearance on the surface at the initiation time is 7.5 mm. The modified Paris’ law and the ANN models for different traffic levels developed previously are utilized to predict the number of days to reach the critical crack depth for each specific pavement section with the load spectra submodel. Calibration Coefficients by Regression Analysis The values of the calibration coefficients are presented in Appendix I. Figure 5.1 shows the predicted versus observed values of ρ and β for the LTPP pavement sections in the wet and dry zones. In this figure, N is the number of days for the top-down crack to reach the critical depth of 15 mm.

104 Figure 5.1 (a). Wet Zone 5.1 (b) Dry Zone R² = 0.8409 N= 16 Standard error=1496 0 2000 4000 6000 8000 10000 12000 14000 16000 0 5 10 15 20 25 Ca lc ul at ed  ρ (d ay ) Pre fatigue life (month) R² = 0.896 N=13 Standard error= 1972 0 5000 10000 15000 20000 25000 0 2 4 6 8 10 12 14 16 Ca lc ul at ed  ρ (d ay ) Pre fatigue life (month)

105 Figure 5.1. Prediction of Calibration Coefficients Prediction of Top-Down Cracking Distress Curve With the estimated calibration coefficients, the top-down design program can be used to predict the growth of top-down cracking length given the material properties, pavement structure, traffic, and climate information. The detailed information for the typical pavement sections in the four climatic zones are indicated in Table 5.1. Figures 5.2 presents comparisons of predicted distress curves for different combinations of the above factors in Table 5.1. R² = 0.5527 N=22 Standard error=0.583 0 0.5 1 1.5 2 2.5 3 3.5 0 0.2 0.4 0.6 0.8 1 1.2 Be ta log (N)

106 Table 5.1 Typical Pavement Sections in Different Climatic Zones Climatic Zone/ State AAD TT *HT (day) **LT (day) E1 (GPa) m- value Hac (inch) Hbase (inch) AV (%) t0 (day) ρ (day) β Pre-Nf (month) DF 1000 21 91 1.5 0.18 5 8 7 550 3148 0.95 3 DF 800 21 91 1 0.20 8 10 4 2948 8000 0.55 6 DNF 1000 94 45 1.5 0.18 3 6 7 497 1530 1.57 2 DNF 800 94 45 1 0.20 5 8 4 2552 4765 0.73 4 WNF 1000 76 5 1.5 0.18 3 6 7 499 3532 1.57 2 WNF 800 76 5 1 0.20 5 8 4 2563 4793 0.73 4 WF 1000 4 148 1.5 0.18 3 6 7 503 3532 1.57 2 WF 800 4 148 1 0.20 5 8 4 2583 4794 0.73 4 *HT- the number of days each year in which the temperature is above 32°C. This illustrates the effect of high temperature on top-down cracking. **LT-the number of days each year in which the temperature falls below 0°C. This illustrates the effect of low temperature on top-down cracking.

107 (a) Pavement Sections in DF Zone (b) Pavement Sections in DNF Zone 0 50 100 150 200 0 1000 2000 3000 4000 5000 6000 7000 Lo ng itu di na l C ra ck  L en gt h  (m ) Service Time (day) DF1 DF2 0 50 100 150 200 250 300 0 1000 2000 3000 4000 5000 6000 7000 Lo ng itu di na l C ra ck  L en gt h  (m ) Service Life (day) DNF1 DNF2

108 (c) Pavement Sections in WNF Zone (d) Pavement Sections in WF Zone Figure 5.2 Pavement Distress Curves for Pavement Sections in Different Climatic Zones 0 50 100 150 200 0 1000 2000 3000 4000 5000 6000 7000 Lo ng itu di na l C ra ck  L en gt h  (m ) Service Life (day) WNF1 WNF2 0 50 100 150 200 0 1000 2000 3000 4000 5000 6000 7000 Lo ng itu di nc al  C ra ck  L en gt h  (m ) Service Life (day) WF1 WF2