DEVELOPMENT OF CRCP-9 COMPUTER PROGRAM FOR ANALYSIS OF CRC PAVEMENTS

(Paper No. 01-3365)

Seong-Min Kim, Ph.D. Research Associate Center for Transportation Research The University of Texas at Austin 3208 Red River, Suite 200 Austin, Texas 78705 Telephone: (512) 232-3134 Fax: (512) 232-3151 E-mail: seong-min@mail.utexas.edu Moon C. Won, Ph.D., P.E. Assistant Branch Director Concrete/Cement, Materials Section Construction Division Texas Department of Transportation 125 E. 11th St. Austin, Texas 78701-2483 Telephone: (512) 465-7502 Fax: (512) 465-3093 E-mail: mwon@dot.state.tx.us and B. Frank McCullough, Ph.D., P.E. Professor Emeritus Center for Transportation Research The University of Texas at Austin 3208 Red River, Suite 200 Austin, Texas 78705 Telephone: (512) 232-3141 Fax: (512) 232-3151 E-mail: bfmccullough@mail.utexas.edu

A Paper Prepared for the 2001 Annual Meeting of the Transportation Research Board Washington, D.C.

Kim, Won, and McCullough

DEVELOPMENT OF CRCP-9 COMPUTER PROGRAM FOR ANALYSIS OF CRC PAVEMENTS

by Seong-Min Kim, Moon C. Won, and B. Frank McCullough

ABSTRACT

A new computer program, CRCP-9, has been developed using finite element formulations and probabilities theories to analyze the behavior of continuously reinforced concrete (CRC) pavements. This paper presents details of the numerical modeling and characteristics of CRCP-9. This mechanistic model predicts crack spacing, crack width, and longitudinal steel stress at given days. The distributions of crack spacings are also provided. Using the predicted crack spacing distributions, this program predicts the number of punchouts per unit length related to the number of wheel load applications. The characteristics of CRCP-9 also include consideration of nonlinear variations of temperature and drying shrinkage through the depth of concrete slab, curling and warping effects, creep effect of concrete, nonlinear bond-slip between concrete and steel, and consideration of changes in material properties with time. The sample analysis results are presented and compared with the CRCP-8 computer program.

Kim, Won, and McCullough INTRODUCTION

The use of continuously reinforced concrete pavement (CRCP) has increased over the years primarily in urban areas in Texas because CRCP provides a durable pavement requiring little maintenance. The effectiveness of CRCP is ensured by tight crack widths and adequate transverse crack spacings, which are determined by design, material, construction, and environmental variables. To ensure tight crack widths and adequate crack spacings, the effect of each variable and the interaction among these variables need to be investigated, so that optimum combinations of design, materials, and construction techniques can be determined. Mechanistic modeling provides a useful tool toward this end. The first mechanistic model, CRCP-1, was made in the mid-1970s under a study sponsored by the National Cooperative Highway Research Program (NCHRP) (1). In 1991, Won et al. developed an improvement to the CRCP program, CRCP-5, that simulates material variance to concrete tensile strength and includes fatigue failure models (2). The normalized curing curves were determined for different coarse

aggregates commonly used in Texas pavements (3), and these curves and the calibrated failure prediction model were included in CRCP-7 (4). In 1995, previous versions of the CRCP programs were integrated into one program, CRCP-8, with simplification of the user input process (5). Although CRCP-8 has permitted pavement engineers to develop designs of the CRC pavements, there are some limitations due to the simplified assumptions of the one-dimensional analysis. In 1996, a research project was conducted to expand the ability of the mechanistic model by incorporating the variations in

Kim, Won, and McCullough

temperature and moisture changes through the depth of concrete slab. As a result of the project, a two-dimensional finite element model was developed (6, 7). In 1998, Texas Department of Transportation (TxDOT) decided to extend the project to complete the development of a new mechanistic model, CRCP-9. CRCP-9 uses two-dimensional finite element theories to reduce the cost of computation, but in order to increase the accuracy of the 2-D model, three-dimensional analyses were performed first and the differences between 2-D and 3-D analysis results were investigated (8). The CRCP-9 computer program has been developed recently, and details of the program are described in this paper.

NUMERICAL MODELING

Theoretical Background The finite element method has been used to model the CRC pavement systems. The concrete slab has been discretized using two-dimensional plane elements with the thickness of the element as the distance between longitudinal steel bars. The longitudinal steel bar has been discretized using frame elements. The element stiffness matrix is obtained by
K = V B T DB dV

(1)

where V = volume of element, D = material property matrix, and B = strain-displacement matrix used to define strain vector.

Kim, Won, and McCullough

When there are initial strains without any other loads, the equilibrium equation can be written as
V B T DB dV U = V B T D
0

dV

(2)

where U = displacement vector, and

0

= initial concrete strain vector due to temperature and drying shrinkage changes.

The stresses can be obtained by

= D(BU
0

(3)

The stresses are calculated at the integration points and the average values are used for each element. The accuracy of the two-dimensional finite element model has been investigated by comparison with the three-dimensional model (8). It is noted that the stresses are little overestimated if the plane strain elements are used and are little underestimated with the plane stress elements compared with the 3-D analysis results. The differences between the stresses from 2-D and 3-D models are very small with the plane stress elements, and can be negligible. The bond-slip between concrete and longitudinal steel bar has been modeled using spring elements between the nodes at contact positions with the active degree of freedom in the longitudinal direction. In this program, the bond-slip relationship can be defined as several different approaches such as linear, bilinear, and bilinear with ultimate slip. The frictional bond-slip that occurs at the interface between concrete and base layer has also been modeled using spring elements.

Kim, Won, and McCullough

The underlying layers have been modeled using vertical spring elements. If any side of the concrete slab curls up, there can be a gap between the slab and the base layer. To properly model this curling effect, the tensionless vertical springs that can sustain only the compressive forces have been developed. It is noted that the curling effect in the CRC pavement is significant when the transverse crack spacing is long and the temperature variation between the top and bottom of the concrete slab is large (8). The boundary conditions of the finite element model should be correctly defined to obtain viable results. At cracks, there are no restraints for concrete and no longitudinal and rotational displacements for longitudinal steel bars. At the center of the two

transverse cracks, vertical degrees of freedom exist, and the longitudinal and rotational displacements are restrained. Figure 1 shows the finite element model developed in this research. To find the optimal size of the finite element, the convergence test has been performed and the element size has been selected as 38.1 mm (1.5 in.) for each side. If the external load is applied to any finite element node, this load will act as a line load because the finite element model developed in this project is two-dimensional. Therefore, the stresses due to the wheel load applications have been obtained using Westergaard equations (9), and added to the stresses from the finite element model. It is noted that further studies are being conducted to obtain more realistic wheel load stresses including effects of discontinuities at cracks and multiple wheel loads. To verify the developed computer code, the same model has been made using the commercial finite element program ABAQUS (10), and the same results could be obtained.

Kim, Won, and McCullough Material Properties

The creep of strain and the relaxation of stress in concrete have been modeled using the effective modulus method (11). The reduced elastic modulus of concrete (effective modulus) is defined as
Eeff (t ) = E (t0 ) 1 + (t , t0 )

(4)

where t = time of consideration, t0 = time at first application of load, E(t0) = modulus of elasticity at time t0, and

(t,t0) = creep coefficient at time t for concrete loaded at time t0, and defined in this
research by

(t , t0 ) = max [1 (1 x ) ]
where

t tx

(5)

max = maximum creep ratio to instantaneous elastic strain, x = ratio to max to define a point on creep curve, and
tx = time corresponding to x. Concrete material properties such as elastic modulus, tensile strength, compressive strength, and drying shrinkage change with time. The equations for

variations of these material properties with time were developed for concrete with various coarse aggregate types in a previous research (3) and are written by
F (t ) = A(2 e Bt e Ct )

(6)

where

Kim, Won, and McCullough t = time of curing (days), e = base of natural log,

F(t) = concrete properties such as elastic modulus, tensile and compressive strengths, and drying shrinkage at time t, and A, B, and C = coefficients of curvature specific to a given aggregate. In this program, eight different coarse aggregate types of limestone, siliceous river gravel (SRG), granite, dolomite, Vega, Bridgeport, western tascosa, and Ferris are used. The coefficient of thermal expansion of concrete depends on coarse aggregate types, and the typical value is used for each aggregate type. The frictional bond-slip between concrete and subbase depends on subbase types. Five different subbase types of flexible, asphalt-stabilized, cement-stabilized, lime-treated clay, and untreated clay subbases are used, and the typical values of bond-slip stiffness have been defined.

Crack Spacing Prediction The crack will occur when and where the concrete stress exceeds the tensile strength of concrete. If the concrete slab is assumed to be homogeneous, the new crack will occur at the center of the two previously formed transverse cracks because the maximum concrete stress occurs at the center. However, since the tensile strength of concrete is governed by the weakest element in it, there exists variation in concrete tensile strength from location to location. The concrete tensile strength depends largely on the bond characteristics between cement paste and the aggregate surface. The bond strength between cement paste and the aggregate surface, for a given water/cement ratio and aggregate type,

Kim, Won, and McCullough

depends on local conditions, such as aggregate surface shape, texture, and the existence of voids between cement paste and the aggregate surface. To include the effect of the variation of the tensile strength along the pavement length in the model, the concrete tensile strength at each finite element is selected randomly using a normal distribution because the concrete tensile strength distribution along the pavement length is reported to be sufficiently close to the normal distribution (2, 12). Once the tensile strength at each finite element is determined and the stresses are calculated from the model, the difference between the tensile strength and the concrete stress is obtained at each finite element. Where the concrete stress exceeds the tensile strength and the difference between them is the largest, the new crack will occur. This methodology is known as the Monte Carlo method (2). As for initial lengths of the slab for the analysis, several different spacings were evaluated for their effects on the analysis results. Using more than two different slab lengths resulted in little variations in the analysis results. Therefore, in this finite element model, five different initial crack spacings of 15.24, 13.72, 12.19, 10.67, and 9.14 m are used as primary crack spacings.

Pavement Life Prediction There are structural failures and functional failures in the pavement systems (13). The structural failures lead to functional failures in CRC pavements, and the major failure manifestation is the punchout (14). The punchout is a structural failure in which a small segment of pavement is loosened from the main body and displaced downward under traffic. The punchout usually is bounded by two closely spaced transverse cracks, a longitudinal crack, and the pavement edge. Even though the punchout development

Kim, Won, and McCullough

mechanism is complicated, it is assumed that the longitudinal crack is the most significant contributing factor in the punchout development. Therefore, it is assumed that longitudinal cracks result in punchouts. Once the crack spacings along the pavement length are obtained from the analysis, the transverse stresses are calculated for each crack spacing by (2)
s = e9.8474 D 1.8143 X 0.4477

(7)

where s = stress in transverse direction for 9000 lb. single wheel load (psi), e = base of natural log, D = slab thickness (in.), and X = crack spacing (ft). The most widely used form of the fatigue failure equation is
f N = C1 ( )C 2 s

(8)

where N = number of load applications, f = flexural strength, s = flexural stress from Eq. (7), and C1 and C2 = coefficients Using the above equation, the number of load applications corresponding to various probabilities of fatigue failure is calculated. After the relationship between the fatigue failure (or punchouts) and load applications has been obtained for each crack spacing, the final fatigue failure versus load application curve can be obtained by adding each curve from each crack spacing. The appropriate coefficients of C1 and C2 in Eq. (8) for CRC

Kim, Won, and McCullough

10

pavements have been found by Suh et al. (4). These values change with the percentage of reliability and the swelling conditions of subgrade soil.

SAMPLE ANALYSIS RESULTS

The analysis has been performed using CRCP-9, and the results are shown in this section. The material properties and the temperature and drying shrinkage variations through the depth of the concrete slab are listed in Tables 1 and 2, respectively. The temperature variations used in this example are selected based on the previous field studies (4, 15). The results from CRCP-9 include mean and standard deviation of crack spacing, crack width, steel stress at crack for each day, crack spacing distribution at a given day, and the relationship between fatigue failure and the number of wheel load applications. Figure 2 shows the variations in mean crack spacings for the first 28 days. The large decrease in the crack spacing is observed within the first several days. For the various amounts of the longitudinal steel, the smaller crack spacings are obtained as the amount of steel increases as shown in Figure 2(a). When the different coarse aggregates are used, the pavement with limestone has larger crack spacings than that with siliceous river gravel (SRG) as shown in Figure 2(b). Figure 3 shows the distribution of the crack spacings after 28 days for pavements with limestone and SRG. The crack spacings for the limestone aggregate are larger than those for SRG. Figure 4(a) shows the cumulative crack spacing distributions from the results shown in Figure 3. These distributions agree well with field observations (16). Figure 4(b) shows the cumulative crack spacing distributions for different values of the coefficient of variation (CV) in concrete tensile

Kim, Won, and McCullough

11

strength. The smaller the CV values are, the smaller variability in the tensile strength is achieved. The crack spacings tend to decrease when CV values increase. This means that the material variability should be controlled well to prevent very narrow crack spacings. Figure 5 shows the frequency of punchouts for various load applications. The SRG concrete pavement has more punchouts than the limestone pavement as shown in Figure 5(a). This is because of the shorter crack spacings in the SRG concrete pavement as already investigated in Figures 2(b), 3, and 4(a). The variability of the concrete tensile strength also affects the frequency of punchouts as shown in Figure 5(b). As the

coefficient of variation (CV) increases, the frequency of punchouts increases. Figures 6 and 7 show sample CRCP-9 output screens.

COMPARISON WITH CRCP-8 COMPUTER PROGRAM

The results from CRCP-8 and CRCP-9 have been compared as shown in Figure 8 to investigate how much difference exists in the results between the two programs. Because the input parameters in those two computer programs are not the same, efforts have been made to have close input values for the two programs. The crack spacing, crack width, and steel stress are compared for different values of the coefficient of variation (CV) in concrete tensile strength. The crack spacings from CRCP-8 are larger than those from CRCP-9 when CV values are small, and are very close with CV values of 20 and larger as shown in Figure 8(a). The crack widths are larger and the steel stresses are smaller in the results from CRCP-9 compared with CRCP-8 (Figures 8(b) and 8(c)). It should be noted that the analysis results are very sensitive to the bond-slip relationship between

Kim, Won, and McCullough

12

concrete and steel bars. Because CRCP-8 and CRCP-9 have totally different approaches to obtain the bond-slip relationship, further experimental studies should be conducted to have more reasonable input values for the bond-slip relationship. CRCP-9 includes all the input parameters required in CRCP-8; and there are additional input parameters in CRCP-9 as listed below. - The variations of temperature and drying shrinkage through the depth of concrete slab can be considered. - The location of the longitudinal steel bar can be changed. - The specific weights of concrete and steel are included for the calculation of the curling effect. - The stiffness of the bond-slip between concrete and steel is included. - The parameters for the creep effects are included. - The type of the plane element can be selected (plane stress or plane strain).

SUMMARY

The new mechanistic model of the CRC pavement, CRCP-9, has been developed using two-dimensional finite element theories. By using CRCP-9, the effects of nonlinear variations of the temperature and drying shrinkage through the depth of the concrete pavement can be evaluated. The curling effect, the creep effect of concrete, and the nonlinear bond-slip between concrete and steel are also considered. The analysis results from CRCP-9 include mean and standard deviation of crack spacing, crack width, and steel stress at given days. The distribution of the crack spacings at a given day is also

Kim, Won, and McCullough

13

provided. The number of punchouts per unit length (km or mile) due to the wheel load applications can be obtained. The CRCP-9 computer program can be used for various purposes. It can be used, for instance, to determine the optimum longitudinal steel design, such as amount of steel, bar diameter, and location of the steel bar, for given pavement structure, material properties, and environmental conditions. To improve the accuracy of the program, it is needed that the calibration of the program with field data to obtain more reasonable ranges of input values and results should be further performed.

ACKNOWLEDGMENTS

The research described in this paper was sponsored by TxDOT in cooperation with FHWA under the project number 0-1831 and was part of a study concerning the improvement of concrete pavement performance model. TxDOT is sincerely appreciated. The financial support of

REFERENCES

1. McCullough, B. F., A. A. Ayyash, W. R. Hudson, and J. P. Randall. Design of Continuously Reinforced Concrete Pavements for Highways. NCHRP 1-15. Center for Transportation Research, The University of Texas at Austin, 1975. 2. Won, M. C., K. Hankins, and B. F. McCullough. Mechanistic Analysis of

Continuously Reinforced Concrete Pavements Considering Material Characteristics,

Kim, Won, and McCullough

14

Variability, and Fatigue. Report 1169-2. Center for Transportation Research, The University of Texas at Austin, 1991. 3. Dossey, T., and B. F. McCullough. Characterization of Concrete Properties with Age. Report 1244-2. Center for Transportation Research, The University of Texas at Austin, 1991. 4. Suh, Y. C., K. Hankins, and B. F. McCullough. Early-Age Behavior of Continuously Reinforced Concrete Pavement and Calibration of the Failure Prediction Model in the CRCP-7 Program. Report 1244-3. Center for Transportation Research, The University of Texas at Austin, 1992. 5. Won, M. C., T. Dossey, S. Easley, and J. Speer. CRCP-8 Program Users Guide. Center for Transportation Research, The University of Texas at Austin, 1995. 6. Kim, S. M., M. C. Won, and B. F. McCullough. Development of a Finite Element Program for Continuously Reinforced Concrete Pavements. Report 1758-S. Center for Transportation Research, The University of Texas at Austin, 1997. 7. Kim, S. M., M. C. Won, and B. F. McCullough. Numerical Modeling of

Continuously Reinforced Concrete Pavement Subjected to Environmental Loads. In Transportation Research Record 1629, TRB, National Research Council, Washington, D.C., 1998, pp. 76-89. 8. Kim, S. M., M. C. Won, and B. F. McCullough. Three-Dimensional Analysis of Continuously Reinforced Concrete Pavements. Proceedings of the 79th Annual

Meeting of the Transportation Research Board, TRB, National Research Council, Washington, D.C., 2000. (Accepted for publication in Transportation Research Record).

Kim, Won, and McCullough

15

9. Westergaard, H. M. Stresses in Concrete Pavements Computed by Theoretical Analysis. Public Roads, Vol. 7, 1925, pp. 25-35. 10. ABAQUS. Users Manual Version 5.8, Hibbit, Karlsson & Sorensen, Inc., Pawtucket, R.I., 1998. 11. Faber, O. Plastic Yield, Shrinkage and Other Problems of Concrete and Their Effect on Design. Minutes of Proc. ICE, Vol. 225, Part 1, London, 1927, pp. 27-73. 12. Hankins, K., and M. C. Won. Condition Survey on US 290. Technical Memorandum 422-25. Center for Transportation Research, The University of Texas at Austin, 1988. 13. Yoder, E. J., and M. W. Witczak. Principles of Pavement Design. 2nd Edition, John Wiley & Sons, Inc., New York, 1975. 14. McCullough, B. F., J. C. M. Ma, and C. S. Noble. Limiting Criteria for the Design of CRCP. Report 177-17, Center for Highway Research, The University of Texas at Austin, 1979. 15. Van Metzinger, W. A., J. R. Lundy, B. F. McCullough, and D. W. Fowler. Design and Construction of Bonded Concrete Overlays. Report 1205-4F. Center for Transportation Research, The University of Texas at Austin, 1991. 16. McCullough, B.F., Zollinger, D., and Dossey, T. Evaluation of the Performance of Texas Pavements Made with Different Coarse Aggregates. Report 3925-1, Center for Transportation Research, The University of Texas at Austin, 1998.

Kim, Won, and McCullough

16

LIST OF TABLES AND FIGURES

TABLE 1 TABLE 2

Input values for sample analysis Variation of temperature and drying shrinkage through depth

FIGURE 1 FIGURE 2

Finite element model Variations in mean crack spacing with age: (a) for various longitudinal steel percentages, (b) for different coarse aggregates

FIGURE 3 FIGURE 4

Crack spacing distribution with different coarse aggregates Cumulative crack spacing distribution: (a) for concrete with different coarse aggregates, (b) for different variations in concrete tensile strength

FIGURE 5

Relationship between frequency of distress and wheel load applications: (a) for concrete with different coarse aggregates, (b) for different variations in concrete tensile strength

FIGURE 6

Sample output screens: (a) time history of mean crack spacing, (b) time history of mean crack width.

FIGURE 7

Sample output screens: (a) time history of mean steel stress at crack, (b) predicted punchouts per mile.

FIGURE 8

Comparison of the results between CRCP-8 and CRCP-9: (a) mean crack spacing, (b) crack width, (c) longitudinal steel stress

Kim, Won, and McCullough

17

TABLE 1

Input values for sample analysis

Concrete slab thickness Percent longitudinal steel Diameter of longitudinal steel Expansion coefficient of steel Steel location from surface Elastic modulus of steel Wheel load

30.48 cm (12 in) 0.6

Poissons ratio of concrete Curing temperature

0.15

40.6 C (105 F)

19.05 mm (0.75 in) 0.000009 /C (0.000005 /F) 15.24 cm (6 in) 200100 MPa (29000000 psi) 40 kN (9000 lbs)

Vertical stiffness of underlying layers Bond-slip stiffness between conc. & steel Specific weight of concrete Specific weight of steel Maximum creep ratio

0.1085 MPa/mm (400 psi/in) 8.14 MPa/mm (30000 psi/in) 22.8 kN/m3 (145 pcf) 76.9 kN/m3 (490 pcf) 2.0

Radius of loaded area

15.24 cm (6 in.)

Load duration for creep

12 hr

Days before wheel load applied Reliability for punchout prediction

14 days

x
tx

0.99

95 %

30 days

Kim, Won, and McCullough

18

TABLE 2

Variation of temperature and drying shrinkage through depth

Day 1
Depth (cm) Morning (C) Afternoon (C) Drying shrinkage ratio Morning (C)

Day 2
Afternoon (C) Drying shrinkage ratio Morning (C)

Day 3-5
Afternoon (C) Drying shrinkage ratio

0.00 3.81 7.62 11.43 15.24 19.05 22.86 26.67 30.48

35.00 35.83 36.67 37.50 37.78 37.78 37.78 37.78 37.78

48.89 46.11 43.33 40.56 37.78 37.78 37.78 37.78 37.78 Day 6-10

1.00 0.80 0.65 0.55 0.50 0.50 0.50 0.50 0.50

28.89 29.44 30.00 30.56 31.11 31.39 31.67 31.95 32.22

43.33 41.67 40.00 38.33 37.22 36.67 36.11 35.56 35.00 Day 11-28

1.00 0.80 0.65 0.55 0.50 0.50 0.50 0.50 0.50

26.67 27.22 27.78 28.33 28.89 29.17 29.44 29.72 30.00

40.56 38.33 36.11 33.89 31.67 30.83 30.56 30.28 30.00 Day 180

1.00 0.80 0.65 0.55 0.50 0.50 0.50 0.50 0.50

Depth (cm)

Morning (C)

Afternoon (C)

Drying shrinkage ratio

Morning (C)

Afternoon (C)

Drying shrinkage ratio

Morning (C)

Afternoon (C)

Drying shrinkage ratio

0.00 3.81 7.62 11.43 15.24 19.05 22.86 26.67 30.48

23.89 25.00 26.11 27.22 28.33 28.61 28.89 29.17 29.44

40.56 38.33 36.11 33.89 31.67 31.11 30.56 30.00 29.44

1.00 0.80 0.65 0.55 0.50 0.50 0.50 0.50 0.50

21.11 22.22 23.33 24.44 25.56 25.83 26.11 26.39 26.67

37.78 35.56 33.33 31.11 28.89 28.33 27.78 27.22 26.67

1.00 0.80 0.65 0.55 0.50 0.50 0.50 0.50 0.50

0.00 1.11 2.22 3.33 4.44 5.00 5.56 6.11 6.67

13.33 12.22 11.11 10.00 8.89 8.33 7.78 7.22 6.67

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Kim, Won, and McCullough

19

Center of cracks

Plane element for concrete Horizontal spring for bond-slip effect

Crack

Roller

No rotation No horizontal disp.

Vertical spring for underlying layers

Frame element for steel

Horizontal spring for frictional bond-slip effect

FIGURE 1

Finite element model.

Kim, Won, and McCullough

20

(a)
5 Mean Crack Spacing (m) 0.40% 4 3 2 1 0 0 4 8 12 16 20 24 28 Pavement Age (Days) 0.50% 0.60%

(b)
5 Mean Crack Spacing (m) 4 3 2 1 0 0 4 8 12 16 20 24 28 Pavement Age (Days) Limestone SRG

FIGURE 2

Variations in mean crack spacing with age: (a) for various longitudinal steel percentages, (b) for different coarse aggregates.

Kim, Won, and McCullough

21

20 Limestone 15 Frequency (%) SRG

10

0 0.1524

0.762

1.3716

1.9812

2.5908

3.2004

Crack Spacing (m)

FIGURE 3

Crack spacing distribution with different coarse aggregates.

Kim, Won, and McCullough

22

(a)
100 Cumulative Value (%) 80 60 40 20 0 0 1 2 Crack Spacing (m) 3 4 Limestone SRG

(b)
100 Cumulative Value (%) 80 60 40 20 0 0 1 2 Crack Spacing (m) 3 4

CV = 10% CV = 15% CV = 20%

FIGURE 4

Cumulative crack spacing distribution: (a) for concrete with different coarse aggregates, (b) for different variations in concrete tensile strength.

Kim, Won, and McCullough

23

(a)
40 Number of Punchouts / km Limestone SRG 20

30

10

0 0 25 50 ESALS (Millions) 75 100

(b)
15 Number of Punchouts / km CV = 10% 10 CV = 15% CV = 20% 5

0 0 25 50 ESALS (Millions) 75 100

FIGURE 5

Relationship between frequency of distress and wheel load applications: (a) for concrete with different coarse aggregates, (b) for different variations in concrete tensile strength.

Kim, Won, and McCullough

24

(a)

(b)

FIGURE 6

Sample output screens: (a) time history of mean crack spacing, (b) time history of mean crack width.

Kim, Won, and McCullough

25

(a)

(b)

FIGURE 7

Sample output screens: (a) time history of mean steel stress at crack, (b) predicted punchouts per mile.

Kim, Won, and McCullough

26

(a)
2 Crack Spacing (m) 1.5 1 0.5 0 5 10 15 20 25 Coefficient of Variation in Concrete Tensile Strength (%) 30 CRCP-8 CRCP-9

(b)
Crack Width (mm) 2 1.5 1 0.5 0 5 10 15 20 25 Coefficient of Variation in Concrete Tensile Strength (%) 30

(c)
Steel Stress (MPa) 500 400 300 200 100 0 5 10 15 20 25 Coefficient of Variation in Concrete Tensile Strength (%) 30

FIGURE 8

Comparison of the results between CRCP-8 and CRCP-9: (a) mean crack spacing, (b) crack width, (c) longitudinal steel stress.