Вы находитесь на странице: 1из 7

Evaluation of Temperature Reponses in Concrete Pavement

Kamyar C. Mahboub, P.E., M.ASCE1; Yinhui Liu2; and David L. Allen, P.E.3

Abstract: Field data from a section of heavily instrumented concrete pavement on the Gene Snyder Freeway 共I-265兲, Louisville,
Kentucky were closely examined. To predict the mechanical behavior of the rigid pavement, a three-dimensional 共3D兲 finite element
Downloaded from ascelibrary.org by Birla Institute of Technology - Pilani on 08/26/16. Copyright ASCE. For personal use only; all rights reserved.

pavement model was developed using the ANSYS program. The 3D finite element model was successfully implemented for predicting the
stresses and strains in the concrete slabs. An important conclusion of this study was that temperature-induced pavement responses were
more significant than traffic-induced responses. This is a very important finding, and can potentially have serious design implications.
Applications of this work may include the development of a more scientifically based rigid pavement design and analysis systems for
Kentucky conditions. Results of this study will serve as background information for the implementation of new pavement design
procedures in Kentucky.
DOI: 10.1061/共ASCE兲0733-947X共2004兲130:3共395兲
CE Database subject headings: Rigid pavements; Concrete pavements; Finite element method; Pavement design; Environment
impacts; Temperature effects; Kentucky.

Introduction dimensional 共3D兲 finite element 共FE兲 model was developed to


account for these complex field conditions.
Environmental factors in pavement engineering usually refer to
the air temperature, moisture, precipitation, and frost/heave,
which can induce various types of stresses/strains in the concrete Effect of Environmental Loads on Concrete
pavement. Although the effects of climate on concrete pavement Pavement
performance have long been recognized, theses effects are not
always included in the concrete pavement design. To develop the Environmental factors that will affect rigid pavement perfor-
basis for a more scientifically based rigid pavement design and mance include temperature, humidity, precipitation, and frost/
analysis system for Kentucky conditions, a project was initiated heave. Daily and seasonal variations in temperature and moisture
by the Kentucky Transportation Cabinet 共KYTC兲 to monitor and affect not only the curling stresses but also the slab–subgrade
evaluate the impact of environmental factors on pavement stress/ contact. A partial loss of subgrade contact will affect the stresses
in concrete. It has been reported by some researchers that the
strains.
temperature gradient in a rigid pavement slab is related to fatigue
As a part of this project, a section of the Gene Snyder Freeway
damage 共Mashad 1996; Ahmad 1998兲. They also reported that a
共I-265兲 in Louisville, Kentucky was instrumented for various field
gradient of 1°F/in. in the slab increases the fatigue damage due to
measurements. These measurements included temperature profile,
truck traffic by a factor of 10 compared to that of a zero-
slab stresses, strain, and deflection. The experiment was designed
temperature–gradient condition. No precipitation and/or frost oc-
to separate the confounding effects of environmental loads and curred during the course of this study, and moisture fluctuation
traffic loads. To accomplish this, pavement responses were mea- was not significant. Therefore, the effects of moisture fluctuations
sured under environmental loads alone 共Load Combination No. were not considered in this study.
1兲, environmental loads plus static vehicle loads 共Load Combina- Experimental and analytical data reported by various research-
tion No. 2兲, and environmental loads plus moving vehicle loads ers suggest that temperature fluctuations produce a significant
共Load Combination No. 3兲. Since the environmental conditions amount of stresses and strains in concrete pavements 共Armaghani
change with time, the environmentally induced loads in the three et al. 1987; Faraggi et al. 1987; Hudson et al. 1987; Barenberg
load combinations are all different. Therefore, a three- et al. 1990兲. In order to address Kentcuky’s conditions, this study
focused on the modeling of a heavily instrumented concrete pave-
1
Associate Professor, Dept. of Civil Engineering, Univ. of Kentucky, ment and its responses to vehicle and temperature loads.
Lexington, KY 40506. E-mail: kmahboub@engr.uky.edu
2
Research Assistant, Graduate Student, Dept. of Civil Engineering,
Univ. of Kentucky, Lexington, KY 40506.
3
Head of Pavements and Materials Section, Kentucky Transportation
Center, Univ. of Kentucky, Lexington, KY 40506.
Note. Discussion open until October 1, 2004. Separate discussions
must be submitted for individual papers. To extend the closing date by
one month, a written request must be filed with the ASCE Managing
Editor. The manuscript for this paper was submitted for review and pos-
sible publication on August 16, 2002; approved on May 5, 2003. This
paper is part of the Journal of Transportation Engineering, Vol. 130,
Fig. 1. Sensors placements for I-265 共Gene Snyder Freeway兲
No. 3, May 1, 2004. ©ASCE, ISSN 0733-947X/2004/3-395– 401/$18.00.

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MAY/JUNE 2004 / 395

J. Transp. Eng., 2004, 130(3): 395-401


Fig. 2. Strain gauge placements 共䊊 denotes lower part of slab兲
Downloaded from ascelibrary.org by Birla Institute of Technology - Pilani on 08/26/16. Copyright ASCE. For personal use only; all rights reserved.

Applications of Three-Dimensional Finite Element sod. The California bearing ratio of the subgrade soil was 5%.
Modeling of Portland Cement Concrete Pavement The base at the test site was 0.153 m 共6 in.兲 dense-graded aggre-
gate 共DGA兲. The concrete used in this project met the require-
3D FE analysis is a powerful tool for linear and nonlinear pave- ment of Kentucky Department of Transportation standards for
ment responses to traffic and environmental loads. Researchers paving class concrete 共Kentucky Class-P兲, which requires a com-
have been using 3D FE pavement models for some time 共White pressive strength of 24.5 MPa 共3,500 psi兲, and Young’s modulus
1996兲. Such studies have included topics such as: Joint load trans- of 28,000 MPa 共4,000,000 psi兲. The dowel and tie bars were
fer 共Channakeshava et al. 1993; Uddin et al. 1996; Davis and A-616 steel, Grade 50. The concrete slabs were 3 months old at
Mahoney 1999; Bhattacharya 2000; Davis 2000兲, backcalculation the time of field tests.
and dynamic analysis 共Vepa and George 1997兲, and responses to To record the mechanical behavior of the slabs, sensors were
nonlinear temperature gradients 共Masad et al. 1997; Pane et al. embedded in the slabs. These sensors included: Pressure cells,
1998兲. Therefore, the selection of 3D FE modeling as a tool was strain gauges, thermistors, and linear variable differential trans-
based upon the success of previous researchers in this area. formers 共LVDT兲. Pressure cells were located under the slabs to
measure the pressure at the center and corner of the slabs. Upper
strain gauges were located 0.203 m 共8 in.兲 above the DGA layer,
Field Test Data and the lower strain gauges were located about 0.076 m 共3 in.兲
above the DGA layer. The strain gauges were used to measure the
Three slabs on Gene Snyder Freeway 共outside of Louisville, Ky.兲 longitudinal, transverse, or perpendicular strain inside the slabs.
were selected for field testing. These nonreinforced slabs were All thermistors were colocated with pressure cells and strain
part of a four-lane highway 共two lanes per approach兲 with con- gauges and at the center of each slab to record the temperature
crete shoulders. All slabs had dowel bars, and adjacent slabs and inside the slabs. LVDTs were embedded in the slabs to record the
shoulders were connected with tie bars. The instrumented slabs deflection of the slab. The locations of sensors are illustrated in
were located on the outside lane. The slab dimensions were: 3.66 Figs. 1 and 2. The sensor data included: Deflection 共LVDT data兲,
m 共12 ft兲 wide, 0.228 m 共9 in.兲 thick, and lengths were: 3.54 m strain, pressure on surface of the DGA layer, and temperature at
共11 ft–7.25 in.兲 for Slab No. 1, 3.96 m 共12 ft–11.75 in.兲 for Slab various sensor locations and inside the slabs. The weather data
No. 2, and 4.15 m 共13 ft–7.5 in.兲 for Slab No. 3. were also monitored, which included air temperature, solar en-
Soils at the test site near Louisville in Northwestern Kentucky ergy, humidity, and precipitation. A sample tracking of pavement
were mainly Pembroke silt loam, overlain with clipped fescue temperature is presented in Fig. 3.

Fig. 3. Field measurement of air temperature and pavement temperature

396 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MAY/JUNE 2004

J. Transp. Eng., 2004, 130(3): 395-401


Table 1. Dowel Bars and Tie Bars: Properties Table 2. Stiffness of Nonlinear Elastic Spring 共Base and Subgrade兲
Property Value Displacement,
mm 共in.兲 Force, kN 共lbs兲
Young’s modulus, MPa 共ksi兲 28,000 共4,000,000兲
Poisson ratio 0.2 ⫺1.318 共⫺0.0519兲 ⫺53.4 共⫺12,000兲
Expansion coefficient, cm/cm/°C 0.000005 ⫺0.6982 共⫺0.02749兲 ⫺35.6 共⫺8,000兲
Density, g/cm3 共lbs/in.3兲 2.5 共0.09兲 ⫺0.4547 共⫺0.0179兲 ⫺26.7 共⫺6,000兲
Dowel bar stiffness 35,000 共2,000,000兲 ⫺0.1052 共⫺0.00414兲 ⫺8.9 共⫺2,000兲
共longitudinal linear elastic spring兲 kN/m 共lbs/in.兲 ⫺0.03531 共⫺0.00139兲 ⫺3.56 共⫺800兲
Tie bar 共transverse linear elastic spring兲 17,500 共1,000,000兲 ⫺0.015748共 ⫺0.00062兲 ⫺1.78 共⫺400兲
kN/m 共lbs/in.兲 25.4 共1兲 8.9 共2,000兲
Downloaded from ascelibrary.org by Birla Institute of Technology - Pilani on 08/26/16. Copyright ASCE. For personal use only; all rights reserved.

To compare the effects of environmental loads 共mainly the dowel bars and tie bars were assumed to be elastic springs,
temperature-induced loads兲 with that of traffic loads and filter out and the base and subgrade were assumed to be nonlinear springs.
the confounding factors, the field data were collected in stages. Therefore, a unidirectional nonlinear spring element was used for
The data were recorded at 5-min intervals for two consecutive the dowel bar and tie bar, and a spring-damper element with 3D
days. On the first day, the pavement was open to traffic and all capability was used for the subgrade. The materials properties of
measurements were conducted under the combination of traffic these elements are given in Tables 1 and 2.
loads and environmental loads. On the second day, the road was
closed to traffic in order to record the pavement responses to
Assumptions and Boundary Conditions
environmental factors alone. Additionally, a portion of the second
day experiment was dedicated to the application of known static When the 3D FE model was generated, certain requirements had
loads at designed slab locations. to be met in order to make the model more representative of field
conditions. The requirements included:
1. The ability to simulate the mechanical responses of the ac-
Modeling and Calibration Process tual pavement, and
2. The placement of nodes at each sensor and assurance of
good agreement between the model solutions and the field
Three-Dimensional Finite Element Model Description
data.
Initially, a 3D FE model was developed and calibrated using the Assuming a solid foundation, the resilient modulus data were
measured field data under environmental loading only. The cali- used as an input for the compressive loading curve of the nonlin-
brated model was then used to predict the response of the con- ear spring 共Huang 1993兲. In the calibration of the FE model, the
crete under environmental load plus vehicle loads. The 3D FE magnitude of the foundation spring stiffness was adjusted to fit
model was developed on an ANSYS platform. Three element types the field data. Initially, the material properties of the concrete
were chosen for the modeling of the concrete slabs, dowel bars, slabs were assumed to be isotropic. During the model calibration
tie bars, the DGA, and subgrade soil. An eight-node brick solid process, the material properties at different directions were ad-
element was selected to represent concrete slabs. The slab model justed slightly in order to find a better fit with the field pavement
was developed to accommodate sensor locations. In the model, response data. It was also assumed that the restriction offered by

Table 3. Comparisons of Field Measurement and Model Calculation Strain Data Prior to Model Calibration 共cm/cm兲
Strain gauge No. 6 Strain gauge No. 7 Strain gauge No. 8 Strain gauge No. 9
Time of day⶿Data Field Model Field Model Field Model Field Model
6:00 am ⫺21.6 ⫺0.649 ⫺13.44 ⫺23.95 27.34 26.04 ⫺37.32 ⫺31.37
6:50 am ⫺19.65 1.4 ⫺11.22 ⫺23.23 30.17 28.97 ⫺36.36 ⫺35.17
7:40 am ⫺18.48 1.9 ⫺9.84 ⫺22.47 31.81 29.67 ⫺35.84 ⫺36.21
7:45 am ⫺18.39 2.255 ⫺9.39 ⫺22.45 31.83 29.57 ⫺35.86 ⫺36.02
8:50 am ⫺16.94 3.63 ⫺7.56 ⫺22.51 33.29 30.16 ⫺34.95 ⫺36.49
9:40 am ⫺14.21 5.52 ⫺9.75 ⫺25.89 35.16 31.52 ⫺34.26 ⫺38.32
11:00 am ⫺11.92 11.17 ⫺17.82 ⫺39.53 40.18 39.8 ⫺41.74 ⫺53.17
11:50 am ⫺16.97 15.8 ⫺23.69 ⫺60.19 37.98 41.88 ⫺49.13 ⫺58.5
1:00 pm ⫺14.08 10.52 ⫺37.33 ⫺70.53 41.15 42.1 ⫺66.1 ⫺90.57
1:50 pm ⫺21.37 5.54 ⫺43.79 ⫺82.36 35.91 37.68 ⫺76.6 ⫺103.19
2:40 pm ⫺26.77 2.43 ⫺47.04 ⫺85.83 30.88 35.49 ⫺83.8 ⫺108.38
3:30 pm ⫺30.36 0.64 ⫺51.2 ⫺89.34 27.33 34.13 ⫺90.2 ⫺114.22
4:20 pm ⫺33.57 ⫺1.07 ⫺56.62 ⫺92.45 22.52 32.34 ⫺97.3 ⫺118.5

Sum ⫺264.31 59.086 ⫺338.69 ⫺660.73 425.55 439.35 ⫺719.46 ⫺860.11

Average difference 24.87 ⫺24.77 1.061 ⫺10.82


between field and
model data

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MAY/JUNE 2004 / 397

J. Transp. Eng., 2004, 130(3): 395-401


Table 4. Comparisons of Field Measurement and Adjusted Model Data After Model Calibration 共cm/cm兲
Strain gauge No. 6 Strain gauge No. 7 Strain gauge No. 8 Strain gauge No. 9
Time of
Day/Data Field Model Difference Field Model Difference Field Model Difference Field Model Difference
12:00 am ⫺20.86 ⫺27.9004 ⫺7.0404 ⫺4.352 10.242 14.594 28.96 21.558 ⫺7.402 ⫺29.11 ⫺15.087 14.023
1:00 am ⫺18.15 ⫺26.154 ⫺8.004 ⫺0.614 14.047 14.661 31.45 22.43 ⫺9.02 ⫺24.97 ⫺11.804 13.166
2:00 am ⫺14.94 ⫺24.6265 ⫺9.6865 2.9 16.975 14.075 34.4 22.72 ⫺11.68 ⫺21.42 ⫺9.74 11.68
3:00 am ⫺12.53 ⫺23.403 ⫺10.873 6.111 19.827 13.716 37.07 22.775 ⫺14.295 ⫺17.85 ⫺7.908 9.942
4:00 am ⫺9.89 ⫺22.225 ⫺12.335 9.26 22.3517 13.0917 39.77 22.938 ⫺16.832 ⫺14.6 ⫺6.384 8.216
5:00 am ⫺7.27 ⫺21.024 ⫺13.754 12.19 24.82 12.63 42.54 23.195 ⫺19.345 ⫺11.53 ⫺5.08 6.45
6:00 am ⫺4.609 ⫺19.346 ⫺14.737 15.17 26.453 11.283 45.44 24.19 ⫺21.25 ⫺9.28 ⫺5.47 3.81
Downloaded from ascelibrary.org by Birla Institute of Technology - Pilani on 08/26/16. Copyright ASCE. For personal use only; all rights reserved.

7:00 am ⫺2.216 ⫺17.93 ⫺15.714 17.51 28.144 10.634 47.97 24.53 ⫺23.44 ⫺7.36 ⫺5.33 2.03
8:00 am 0.298 ⫺16.586 ⫺16.884 20.31 29.636 9.326 50.35 24.88 ⫺25.47 ⫺5.399 ⫺5.29 0.109
9:00 am 3.24 ⫺14.662 ⫺17.902 23.22 29.836 6.616 53.33 26.137 ⫺27.193 ⫺3.473 ⫺6.958 ⫺3.485
10:00 am 6.603 ⫺9.562 ⫺16.165 20.7 22.595 1.895 57.48 31.913 ⫺25.567 ⫺6.785 ⫺16.469 ⫺9.684
11:00 am 8.91 ⫺3.714 ⫺12.624 10.32 8.485 ⫺1.835 60.95 40.26 ⫺20.69 ⫺17.54 ⫺32.965 ⫺15.425
12:00 pm 9.17 0.69 ⫺8.48 ⫺4.021 ⫺8.67 ⫺4.649 61.31 48.21 ⫺13.1 ⫺32.01 ⫺51.683 ⫺19.673
1:00 pm 6.638 ⫺1.367 ⫺8.005 ⫺20.45 ⫺33.83 ⫺13.38 59.89 56.27 ⫺3.62 ⫺49.66 ⫺69.09 ⫺19.43
2:00 pm 1.188 ⫺2.329 ⫺3.517 ⫺36.4 ⫺51.44 ⫺15.04 55.11 60.734 5.624 ⫺66.79 ⫺84.23 ⫺17.44
3:00 pm ⫺6.12 ⫺5.712 0.408 ⫺49.72 ⫺64.93 ⫺15.21 49.1 62.236 13.136 ⫺81.1 ⫺94.15 ⫺13.05
4:00 pm ⫺14.42 ⫺11.285 3.135 ⫺59.75 ⫺73.992 ⫺14.242 40.57 60.048 19.478 ⫺91.3 ⫺99.24 ⫺7.94
5:00 pm ⫺22.22 ⫺17.6297 4.5903 ⫺65.34 ⫺76.71 ⫺11.37 32.38 56.65 24.27 ⫺96.5 ⫺97.48 ⫺0.98
6:00 pm ⫺29.3 ⫺24.5299 4.77014 ⫺66.37 ⫺74.3 ⫺7.93 24.59 50.739 26.149 ⫺96.7 ⫺90.56 6.14
7:00 pm ⫺34.9 ⫺31.31 3.59 ⫺62.31 ⫺66.486 ⫺4.176 17.36 43.524 26.164 ⫺91.3 ⫺79.329 11.971
8:00 pm ⫺37.43 ⫺36.287 1.143 ⫺53.46 ⫺55.704 ⫺2.244 13.56 36.972 23.412 ⫺82.4 ⫺65.42 16.98
9:00 pm ⫺37.18 ⫺38.174 ⫺0.994 ⫺45.26 ⫺46.814 ⫺1.554 12.55 33.625 21.075 ⫺74.5 ⫺55.038 19.462
10:00 pm ⫺35.94 ⫺38.97 ⫺3.03 ⫺39.84 ⫺40.014 ⫺0.174 14.05 30.928 16.878 ⫺68.04 ⫺47.777 20.263
11:00 pm ⫺34.32 ⫺28.7516 5.5684 ⫺34.81 ⫺19.865 14.945 14.4 28.329 13.929 ⫺62.64 ⫺43.704 18.936
12:00 am ⫺32.6 ⫺28.8454 3.7546 ⫺31.02 ⫺15.912 15.108 15.82 27.084 11.264 ⫺57.99 ⫺39.843 18.147

Average ⫺6.11142 2.430828 ⫺1.501 2.96872


Maximum 0.69 29.836 60.734 ⫺5.08
Minimum ⫺38.97 ⫺66.486 21.558 ⫺99.24
Range 39.66 96.322 39.176 94.16
Average range % 15.4% 2.5% 3.8% 5.3%

the concrete slabs in the adjacent lane was the same as that ap- ronmental conditions, the slabs never reached a fully flat position.
plied by the concrete shoulder. The dowel bars were assumed to Therefore, it was assumed that the slabs in this study were manu-
have full contact and were treated as elastic springs in the model. factured with some level of residual strain 共caused by concrete
Both the DGA base and the subgrade soil were treated as one curing, loss of moisture, and shrinkage兲. According to the general
composite layer. FE theory, the equivalent nodal force is equal to the combination
Field observations revealed that the slabs responded to tem- of the forces due to distributed body forces, initial strain, and
perature cycles by curling. However, upon the reversal of envi- residual stresses 共Zienkiewicz and Taylor 2000兲. In order to simu-

Fig. 4. Model calibration by longitudinal strain, gauge No. 6 共bottom of the center slab兲

398 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MAY/JUNE 2004

J. Transp. Eng., 2004, 130(3): 395-401


Downloaded from ascelibrary.org by Birla Institute of Technology - Pilani on 08/26/16. Copyright ASCE. For personal use only; all rights reserved.

Fig. 5. Model calibration by transverse strain, gauge No. 9 共top of the center slab兲

late the initial strain and residual stresses in the slab, distributed The data obtained under environmental load only or under
body forces were applied to each slab. The magnitude of these static vehicle loads were used to calibrate the pavement model to
body forces were adjusted during the calibration of the model. find the best fit for material properties and nonlinear spring stiff-
ness. To apply static edge loads, a parked single-unit dump truck
Solution Process with a rear tandem axle group was put along the right-hand side
edge of the pavement. When the tandem axle was located at the
A two-step solution was employed in order to enable the model corner of Slab No. 2, the load condition was referred to as the
better predict the field conditions. The first step was a thermal corner load, and when tandem axle was located at the center of
analysis, in which temperature readings at various thermistor lo- Slab No. 2; the load condition was referred to as the center load.
cations were entered in the model. These temperature profiles in
During the calibration process, the data obtained under envi-
the slabs were used to calculate thermally induced strains and
ronmental loads alone were used first to find the best fit for pave-
stresses in the slabs. The second step was a structural analysis. In
ment material properties and nonlinear spring stiffness. After that,
this step, gravity and static loads were entered into the model
the data obtained under the static load were used to see if the
along with the temperature-induced stresses as pavement loading.
Finally, tentative solutions were generated for pavement deflec- model would produce acceptable results. With the exception of a
tions, stress, and strains. Obviously, the model had to be cali- few anomalous points, the calibration results show that the cali-
brated to provide a closer match with actual field measurements. brated model did a good job of predicting various field conditions.
The calibration process is described in the next section. The comparison of model solutions and field data under the
environmental loads alone, as well as environmental plus static
vehicle loads are shown in Tables 3 and 4 and Figs. 4 and 5.
Three-Dimensional Finite Element Model Calibration Strain gauge No. 6, located at the bottom of slab, a was used to
The objective of model calibration was to find a set of material measure the strain in the longitudinal direction. Strain gauge No.
properties and element constants, which can closely match the 7, located at the top of slab, was also used to measure the strain in
field data. To calibrate the model for the Gene Snyder Freeway the longitudinal direction. Strain gauge No. 8, located at the bot-
pavement, a parametric study was conducted. The stiffness of tom of slab, was used to measure the strain in the transverse
linear and nonlinear springs and Young’s modulus of the concrete direction. Strain gauge No. 9, located at the top of slab, was used
slab were adjusted to allow the model to better correspond with to measure the strain in the transverse direction. The average
field conditions.

Fig. 6. Effect of load condition on longitudinal strain, gauge No. 6 共bottom of the center slab兲

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MAY/JUNE 2004 / 399

J. Transp. Eng., 2004, 130(3): 395-401


Downloaded from ascelibrary.org by Birla Institute of Technology - Pilani on 08/26/16. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Effect of load condition on transverse strain, gauge No. 9 共top of the center slab兲

difference of the field strain predictions and model strain data at sively based on traffic loads, should be modified to include the
these locations appear to be reasonable. effects of temperature cycles. More weight must be given to the
role of temperature fluctuation.

Traffic versus Temperature Loads


Disclaimer
Using the field data and modeling, an attempt was made to quan-
tify the effects of traffic and temperature loads on concrete slabs. The contents of this report reflect the views of the writers who are
Figs. 6 and 7 show the comparison between strains induced by responsible for the facts and accuracy of the data presented
different vehicle load types at the location of strain gauge Nos. 6 herein. The contents do not necessarily reflect the official views
and 9, respectively. In Figs. 6 and 7, the field measurements under or policies of the University of Kentucky, the KYTC, or the Fed-
traffic load plus environmental load were compared with the ad- eral Highway Administration. This report does not constitute a
justed model solutions. The information in Figs. 6 and 7 demon- standard, specification, or regulation. The inclusion of manufac-
strates that the contribution of the ordinary highway traffic- turer names and trade names is for identification purposes only
induced pavement response is rather small as compared to the and is not to be considered an endorsement.
temperature-induced response. Therefore, the trends of the change
in strains in concrete pavement are dominated by the change in air
temperature, while the static vehicle load and traffic load induce a
References
relatively small fluctuation in the strain change tendency. This is
an important finding with potentially serious ramifications in con-
Ahmad, I., Rahman, M. H., Seraj, S. M., and Hoque, A. M. 共1998兲.
crete pavement design and analysis.
‘‘Performance of plain concrete runway pavement.’’ J. Perform. Con-
str. Facil., 12共3兲, 145–157.
Armaghani, J. M., Dempsey, B. J., Hill, H., and Vogel, J. 共1987兲. ‘‘Tem-
Summary and Recommendations perature Response of Concrete Pavements.’’ Transportation Research
Record 1121, Transportation Research Board, National Research
This paper focused on a section of concrete pavement on Gene Council, Washington D.C., 23–33.
Snyder Freeway 共I-265兲, Louisville, Ky. These concrete slabs Barenberg, E. J., and Zollinger, D. G. 共1990兲. ‘‘Validation of concrete
were heavily instrumented in this study in order to measure pave- pavement responses using instrumented pavements.’’ Transportation
ment responses to temperature and vehicle loads. A 3D FE pave- Research Record 1286, Transportation Research Board, National Re-
ment model was developed using the ANSYS program. The 3D FE search Council, Washington D.C., 67–77.
Bhattacharya, K. 共2000兲. ‘‘Nonlinear response of transverse joints of air-
model was successfully implemented for predicting the stresses
field pavements.’’ J. Transp. Eng., 126共2兲, 168 –177.
and strains in the concrete slabs. Based upon various comparisons
Channakeshava, C., Barzegar, F., and Voyiadjis, G. 共1993兲. ‘‘Nonlinear
of field data and model solutions, it was demonstrated that the 3D FE analysis of plain concrete pavement with doweled joints.’’ J.
FE model was successfully correlated to the stresses and strains in Transp. Eng., 119共5兲, 763–781.
the concrete slabs. Davis, W., and Mahoney, J. 共1999兲. ‘‘Experimental verification of rigid
The study demonstrated that environmentally induced pave- pavement joint load transfer modeling with EverFE.’’ Transportation
ment responses overshadowed the traffic-induced responses. The Research Record 1648, Transportation Research Board, National Re-
change in strain induced by traffic load is approximately less than search Council, Washington D.C., 81– 89.
20% of the change due to temperature flux for a specific day and Davis, W. 共2000兲. ‘‘Effect of dowel looseness on response of jointed
time. The trends for strains as a function of temperature at each concrete pavements.’’ J. Transp. Eng., 126共1兲, 50–57.
point in the slabs were not significantly affected as a result of Faraggi, V., Jofre, C., and Kraemer, C. 共1987兲. ‘‘Combined effect of traf-
fic loads and thermal gradient on concrete pavement design.’’ Trans-
static vehicle or moving traffic load applications. This means that
portation Research Record 1136, Transportation Research Board, Na-
environmental conditions produce a far more significant response tional Research Council, Washington D.C., 108 –118.
than traffic in concrete pavements. Huang, Y. H. 共1993兲. Pavement analysis and design, Prentice-Hall, New
It is recommended that the findings of this paper be incorpo- York.
rated in various pavement design and analysis routines. The cur- Hudson, W. R., and Flanagan, P. R. 共1987兲. ‘‘An examination of environ-
rent concrete pavement fatigue models, which are almost exclu- mental versus load effects on pavements.’’ Transportation Research

400 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MAY/JUNE 2004

J. Transp. Eng., 2004, 130(3): 395-401


Record 1121, Transportation Research Board, National Research Uddin, W., Hackett, R. M., Joseph, A., Pan, Z., and Crawley, A. B.
Council, Washington D.C., 34 –39. 共1996兲. ‘‘Three-dimensional finite-element analysis of jointed con-
Masad, E., Taha, R., and Muhunthan, B. 共1996兲. ‘‘Finite element analysis crete pavement having discontinuities.’’ Transportation Research
of temperature effects on plain-jointed concrete pavements.’’ J. Record 1482, Transportation Research Board, National Research
Transp. Eng., 122共5兲, 388 –398. Council, Washington D.C., 26 –32.
Pane, I., Hansen, W., and Mohamed, A. R. 共1998兲. ‘‘Three-dimensional Vepa, T. S., and George, K. P. 共1997兲. ‘‘Deflection response model for
finite element study on effects of nonlinear temperature gradients in cracked rigid pavements.’’ J. Transp. Eng., 123共5兲, 377–384.
concrete pavements.’’ Transportation Research Record 1692, Trans- White. 共1996兲.
portation Research Board, National Research Council, Washington Zienkiewicz, O. C., and Taylor, R. L. 共2000兲. The finite element method,
D.C., 58 – 66. 5th Ed., Butterworth–Heinemann, Oxford, U.K.
Downloaded from ascelibrary.org by Birla Institute of Technology - Pilani on 08/26/16. Copyright ASCE. For personal use only; all rights reserved.

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MAY/JUNE 2004 / 401

J. Transp. Eng., 2004, 130(3): 395-401

Вам также может понравиться