Professional Documents
Culture Documents
011 Kohler Accelarated Load Tests
011 Kohler Accelarated Load Tests
011 Kohler Accelarated Load Tests
Erwin Kohler
ABSTRACT
1. INTRODUCTION
Continuously Reinforced Concrete Pavements (CRCP) are known to perform better than
other types of concrete pavements (Gharaibeh at al. 1999, Smith et al. 1998) and are
considered the preferred option by some highway agencies for extended life pavements. This
paper describes an experiment that investigated the repeated load behavior of CRCP
sections subjected to accelerated traffic, and presents examples of in service CRCP in Illinois
(Dan Ryan Expressway), California (Intersate I-80 at Fairfield), and the general experience in
the state of Texas.
Ten experimental sections were built at the Advanced Transportation Research and
Engineering Laboratory (ATREL) in December 2001 with funding from the Illinois Department
of Transportation. Sections 1 through 5 were loaded with accelerated traffic and the results
are presented in this paper. Sections 6 to 10 were built to compare the effect of induced
cracks and have not yet been subjected to traffic. A summary of the field test section results
and overall benefits of active crack induction for these sections can be obtained from Kohler
and Roesler (2004). All sections are approximately 26 m long with transition zones between
sections to accommodate steel content changes. End restraints were constructed at the end
of each lane to prevent excessive slab movement and to anchor the longitudinal steel. The
end restraints consisted of a double lug system with a depth of 1.22 m and the width of the
concrete pavement, which is 3.66 m. The layout and main design characteristics of the
sections are presented in Figure 1.
Proceedings of the International Conference on Concrete Roads (ICCR2007) 16 – 17 August 2007
ISBN Number: 978-1-920017-31-6 Midrand, South Africa
Produced by: Document Transformation Technologies cc Conference organised by: Cement and Concrete Institute
125
150 m
Lane 2 26 m
6 7 8 9 10
p=0.55%, #5 p=0.80%, #6 p=1.09%, #7 p=0.80%, #6 p=0.80%, #6
h=254, d=89 h=254, d=89 h=254, d=89 h=254, d=89 h=254, d=178
Lane 1
1 2 3 4 5
p=0.55%, #5 p=0.80%, #6 p=1.09%, #7 p=0.78%, #7 p=0.78%, #7
h=254, d=89 h=254, d=89 h=254, d=89 h=356, d=178 h=356, d=89 & 178
p : percent of steel
# : bar size (US system)
h : concrete thickness (mm)
d : depth of the steel layer (mm)
All the transverse cracks developed naturally in the pavement sections of this study, i.e., there
were no induced cracks in Lane 1 sections. The thickness of the concrete was 254 mm
except in sections 4 and 5 where it was 356 mm. The slab reinforcement consisted of 26
longitudinal bars, spaced 140 mm apart. The bar diameter in sections 1 and 2 was 15.9 mm
(#5) and 19.1 mm (#6), respectively. Sections 3 to 5 had 22.2mm bars (#7). The depth of the
longitudinal reinforcement was 89 mm in sections 1, 2 and 3, and it was 114 mm in section 4.
The reinforcement in section 5 was split in two layers, at 89 and 178 mm from the surface.
Transverse steel reinforcement (#4 bars) spaced every 1.2 m supported the longitudinal bars.
LVDT holders
A variety of sensors were installed in the test sections in order to monitor the environmental
and repeated load performance of the CRC pavement so that differences in pavement
responses between test sections could be identified. Horizontal movements were measured,
via LVDTs, relative to each side of the crack, at different depths in the slab. Vertical
movements were measured with two LVDTs suspended from a reference beam supported
away from the loaded pavement edge. Vertical and horizontal sensors were placed
concurrently at each monitored crack, as shown in Figure 2. During the CRC pavement
construction, four strain gages were embedded in each section. The strain gages were placed
25 mm from the CRCP surface, oriented transversely, and spaced at a distance from the
loaded edge where the maximum stress was expected to occur under a rolling wheel (strain
126
gage results are omitted in this article due to space restriction). Thermocouples were placed
at 25 mm from the top and bottom of the slab to determine the pavement temperature
(differential and average through the thickness). Additional thermocouples in the slab depth
were used as part of a static data collection system that functioned permanently in all sections.
A picture of the “Advanced Transportation Loading Assembly” (ATLAS), with a full view of the
device is shown in Figure 3.
Station(m)
0 5 10 15 20 25
0.0
0.1
Vertical Deflection (mm)
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9 Load=200 kN
1.0
Traffic Sensors
127
Similar influence lines were collected for transverse strains and for horizontal movements at
the instrumented cracks. The effect of the moving wheel load is represented by the rebound
responses, defined as the peak value measured as the wheel passes over the sensor
compared to the reading after the wheel has moved away from the sensor. These responses
depend on the load level being applied and also on the thermal state of the pavement, and
may vary along a uniform section due to variability in material and support conditions. Figure
5 shows rebound vertical deflections measured continuously over a two-week period for a
fixed load level of 133 kN. These measurements were obtained from section 3 during the
summertime when the daily thermal cycles were more pronounced. Figure 6 also shows the
temperature difference through the thickness of the slab, which fluctuated each day between
-2 and 4°C approximately (note the temperature scale is inverted in the plot). Despite the
fixed load level, the rebound deflections were not constant but varied with thermal conditions
in the slab. Deflections in the morning could be 20 percent higher than in the late afternoon.
1.6
1.5
Vertical deflection (mm)
1.4
1.3
1.2
1.1 Rebound deflection under
repeated 133 kN wheel load
1.0
0.9
0.8
Temp.difference (C)
-4
Temperature difference top to bottom
-2
0
2
4
6
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Days
Similar behavior was observed in all pavement sections were the edge uplift caused by
transverse curling increased the rebound deflection and the transverse strain near the top of
the slab. The effect slab curling has on the pavement responses make it difficult to compare
strain and deflection measurements between sections, which were tested in different seasons.
As a reference, Figure 6 shows typical pavement temperature profiles measured in winter and
summer and in the morning and the afternoon.
128
Temperature (C)
-20 -15 -10 -5 0 5 10 15 20 25 30 35 40 45 50
0
1 Winter
50 50
Top
40 40
Temperature (C)
Temperature (`C)
Bottom
30 30
20 20
10 10
Top
0 0 Bottom
-10 -10
Day Night Day Night Day
Day Night Day Night Day -20
-20
Hour Hour
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Section 1 2 3 4 5
T-avg (°C) 7.8 +/-2.8 32.8 +/-2.9 23.3 +/-3.1 -2.2+/-1.2 7.2 +/-0.2
T-diff (°C) -1.1 +/-0.2 -1.1 +/-2.8 5.6 +/-5.5 1.9 +/-0.8 3.2 +/-0.2
Season Fall Summer Summer Winter Spring
129
Sections 1, 2, and 3 consist of 254 mm thick slabs, and the deflections were approximately
twice the deflections observed in the 356 mm thick slabs of sections 4 and 5. The additional
reinforcement in section 3 (1.09 percent) with respect to sections 1 and 2 (0.55 and 0.80
percent respectively) did not significantly alter the elastic deflection measurements. At time of
testing in section 2, the subgrade was saturated due to heavy rainfall, which could help to
explain the higher deflections that this section experienced compared to sections 1 and 3. The
356 mm thick sections experienced similar deflections values to each other.
The highest number of cracks occurred in section 3, which was in the middle of Lane 1 and
had the largest steel content. The average crack spacing was 0.78 m. In section 2, the
average crack spacing was 0.90 m with the only design difference between them being the
steel content. The additional steel in section 3 with respect to section 2 resulted in a reduction
in the average crack spacing. Section 1 had even less steel than section 2, and the average
crack spacing increased to 1.40 m. Some movement likely occurred at the lug end of section
1, which had the effect of increasing the average spacing. It is not definitive whether the 1.40
m spacing corresponded to the true crack spacing of the section 1 due the unknown end
effects. It is reasonable to expect if section 1 had been at an interior location in Lane 1, the
spacing would be between 0.90 and 1.40 m (larger than section 2 and shorter than observed).
Crack spacing between 1.0 and 1.5 m are considered suitable for good performing CRCP,
and therefore sections 1 to 3 can represent desirable field sections. Section 4 was 112 mm
thicker and had approximately the same steel content as section 2, although placed deeper
from the surface. Section 4 has half the number of cracks as section 2, and therefore it can be
concluded that the two effects, thicker slab and greater steel depth, combined to generate a
spacing of 1.44 m in section 4 that is considerably higher than the 0.90 m in section 2.
Existing field data (Burke and Dhamrait 1968, Dhamrait et al. 1973) and mathematical models
show an increase in thickness and deeper reinforcement result in larger crack spacing. The
cracking pattern in Section 5 was difficult to assess since only four cracks developed over the
entire pavement length. The large crack spacing can be attributed to the lug movement at the
end of Lane 1 and the effect of double layer steel placement. Crack spacing in all sections
(except section 5) can be considered within a typical range of field CRCP.
130
ratio of 20 to account for channelized trafficking of the ATLAS wheel at the pavement edge.
The estimated ESALs at time of the first longitudinal crack was between 230 and 548 millions
for the 254mm sections, and no damage occurred on section 4 (356mm) after 764 millions
ESALs.
Longitudinal cracks initiated about 1.5 m from the loaded edge at transverse cracks and
eventually propagated toward the pavement edge with more load applications. At the end of
loading, they resembled half-moon cracks. The places where the longitudinal crack extended
to the edge coincided with longer panels. The presence of other closely spaced transverse
cracks allowed the crack propagation to continue advancing longitudinally through multiple
transverse cracks. The half-moon cracks had the potential to develop faulting and the
enclosed area was considered a punchout. A cascade effect created failures associated with
the original punchouts, affecting a considerable length of the test section. The length of each
punchout was between 3 and 14 m. Secondary longitudinal cracks formed closer to the edge
in sections 1 and 2. Figure 8 presents the crack maps for all sections after the loading was
finished.
131
132
Figure 8. Crack maps at the end of loading.
The measurements of vertical deflection on both sides of each crack revealed that the load
transfer across the cracks did not decrease as described in the traditional punchout failure
mechanism for CRCP. Based on the observation of the vertical deflections, the longitudinal
cracking was the result of permanent deformation under the slab produced by the repetitive
action of heavy wheel loads.
The longitudinal cracks started and developed with no decrease in load transfer efficiency
across the transverse cracks. Traditionally, the loss of LTE is cited as the cause of the
increased stresses that generate the longitudinal cracks. However, it was found that under
small crack widths, the longitudinal cracks start and propagate without loss in transverse
crack shear capacity (see Figure 9). The sequence of events that lead to faulted punchouts is
described in Figure 10.
110
100
90
LTE (%)
80
70
45 kN 133 kN 156 kN
60
50
0 25 50 75 100 125 150 175
Passes (Thousands)
Punchouts
(rectangular) Punchouts
133
3. CASES OF CRCP IN THE UNITED STATES
After six years in service, traffic levels had already surpassed original estimates for the entire
life of the project, and there were some pavement performance issues associated with
smooth bars that were used and with corrosion. A reconstruction effort took place in 1972 in
the local lanes for nearly 8 of the 18 Kms. The local lanes were reconstructed with 250mm
CRCP (replacing the 200-mm CRCP). The granular subbase was removed under the traffic
lanes (not under the shoulders) and replaced with 100 mm of stabilized subbase. The
reinforcing steel, now deformed #25 (25mm) bars, raised the percentage of reinforcement to
0.7. The express lanes were not reconstructed. Repairs, when needed, consisted of full-depth
concrete patches without reinforcement, and sometimes asphalt overlays: The 200-mm
CRCP express lanes were overlaid with 76mm of asphalt; the 254-mm CRCP local lanes
were overlaid with a total of 160mm of asphalt. In 1997 the asphalt in the express lanes was
resurfaced, removing the top 25mm and then adding 140 mm, resulting in a total asphalt layer
of 191mm.
The Illinois Department of Transportation (IDOT) and the City of Chicago began
reconstructing the Dan Ryan Expressway in 2003. The pavement reconstruction began in
2006 and is scheduled to be finished in 2007, and it is being built once again as CRCP.
Based on the results of accelerated load testing at ATREL, the pavement thickness is 356
mm with 0.7 percent reinforcing steel (#22 [22mm] longitudinal reinforcing bars), and lug
system to control pavement expansion. The concrete is placed over a 150mm recycled
asphalt stabilized base layer over a 610mm granular layer on the existing subgrade. To
prevent corrosion of the reinforcing bars, IDOT specifies epoxy coating on the longitudinal
and transverse bars, and also requires the bar supports and ties to be either epoxy-coated or
otherwise no corrosive (e.g., plastic ties).
134
Figure 12. Reconstruction of Dan Ryan Expressway continuously reinforced concrete
pavement in 2006. (Talley 2006)
The Fairfield pavement consisted of two experimental sections, totaling 1.6 Km (1 mile). The
road was then highway US40, and later became part of the Interstate System, highway I-80, a
major road that crosses the country virtually from New York to San Francisco.
The project was originally built as a four-lane divided highway with two 3.7-m lanes in each
direction. The CRCP was constructed in the west-bound lanes. The concrete slab was built
200 mm thick, with a tongue and groove joint at the junction of the two lanes. The
reinforcement consisted of 13mm-diameter deformed steel bars placed at two different
spacings: 102-mm (4-in.) and 127-mm (5-in.) centers, along each half of the project which
constituted section 1 and 2 respectively. The steel content amounted to 0.62 percent and 0.50
percent. The reinforcement was placed at mid depth of the slab. The California Bearing Ratio
(CBR) of the subgrade ranged from 2 to 20 percent. A 250mm layer of a material with CBR of
150 to 200 percent was placed over the existing subgrade. The top 100 mm of this material
were treated with 3.5 percent of cement to result in a cement treated base.
135
Crack spacing in the traffic and passing lanes of the two sections ranged between 1.2 m and
1.7 m, with an average of 1.4 m, and stabilized after the first year. Only one punchout
developed over the entire length of the section in the first 20 years, and a second one is
currently observable, more than 50 years after construction.
The cracks have remained tight and in very good condition, as it can be observed in Figure 8.
Figure 12. The two left lanes are the CRCP built in 1949 (cracks in the second lane were
unnecessarily sealed during maintenance). The outer two lanes were constructed later
with 5m concrete slabs
136
3.3 Texas Experience
The Texas Department of Transportation (TxDOT) built its first CRCP in 1951 in the Fort
Worth District, which became IH-30. It consisted on a concrete slab 200 mm thick over 20 mm
of crushed stone with a two-course surface treatment. The amount of longitudinal steel was
0.69 percent. Performance has been excellent. Since then, many kilometers of CRCP have
been built in Texas. As of 2005, Texas had about 16,000 lane-km of CRCP. The overall
performance of CRCP has been satisfactory, even though some distress types developed
that required repairs. Satisfied with the performance of CRCP, in 2000 TxDOT made CRCP
use an official policy when rigid pavement type is selected for a project.
A research team from the Center for Transportation Research of the University of Texas at
Austin selected about 60 test sections and has been conducting detailed field testing
including visual condition surveys, deflection testing, and LTE evaluations. No punchouts
have been observed in the CRCP since new designs (thicker slabs, stabilized base, and tied
concrete shoulder) were implemented (Won et al 2006). Punchouts are observed in CRCP
with old designs and construction practices. Also, no wide transverse cracks have been
observed. Spalling continues to be a problem in pavement sections built with a coarse
aggregate type. The TxDOT 2005 Pavement Management Information System Annual Report
states that “CRCP condition, distress, and ride quality improved to the highest level in four
years in FY 2005. CRCP was the only pavement type to show improved ride quality in FY
2005—in fact, it now has the best overall ride quality of the major pavement types in Texas.”
This statement highlights the durable nature of CRCP.
137
4. SUMMARY
The performance of the CRCP from full-scale experimental sections and also from real
sections in roadways in operations, indicate that this type of pavement provides exceptionally
long life. The example in Illinois showed that after 44 years of service, a project is now being
reconstructed again on CRCP, with design modifications derived from the accelerated load
test. The example in California showed that after 57 years in service an originally
experimental CRCP project is still serving well at a national corridor. The experience in Texas
with CRCP has been very positive, to the point that is now the official pavement type (when
concrete is selected) and recent evaluation revealed that it provides the best overall ride
quality among the major pavement types.
5. ACKNOWLEDGMENTS
The accelerated load test part of this material is based upon work supported by the Illinois
Department of Transportation (IDOT) conducted by the author of this article in collaboration
with Dr. Jeff Roesler of the University of Illinois at Urbana-Champaign. The material about the
Texas experience with CRCP is based on work by Dr. Moon Won of the University of Texas
Austin. The cases presented from Illinois and California are based on compilations by the
author of this article and papers from Andrea Talley formerly with CRSI) and Chetana Rao, of
ARA Inc, respectively.
6. REFERENCES
[1] Burke, J.E., Dhamrait, J.S. (1968). A Twenty-Year Report on the Illinois Continuously
Reinforced Pavement; Highway Research Record No. 239, Highway Research Board,
pp197-211
[2] Dhamrait, J.S., Jacobsen, F.K., and Schwartz, D.R. (1973). Condition of Longitudinal
Steel in Illinois Continuously Reinforced Concrete Pavements, Physical Research Report
No. 89 (IHR-36), Illinois DOT
138
[3] Gharaibeh, N. G., Darter, M.I., Heckel, L.B. (1999). Field Performance of Continuously
Reinforced Concrete Pavement in Illinois, Transportation Research Record - Journal of
the Transportation Research Board 1684, pp 44-50.
[4] Kohler, E.R. (2005). Experimental Mechanics of Crack Width in Full-Scale Sections of
Continuously Reinforced Concrete Pavements. PhD Dissertation, University of Illinois at
Urbana-Champaign.
[5] Kohler, E.R. and Roesler, J.R. (2005). Crack Width Measurements in Continuously
Reinforced Concrete Pavements, ASCE Journal of Transportation Engineering.
[6] Rao, C., Darter, M.I., and Pyle, T. (2006). Extended Service Life of Continuously
Reinforced Concrete Pavement in California. Proceedings of the International
Conference on Long-Life Concrete Pavements, Illinois October 25–27, 2006
[7] Smith, K. D., Wade M. J., Peshkin D. G., Khazanovich L., Yu H. T., Darter M.I., (1998).
Performance of Concrete Pavements, Vol. II: Evaluation of in-service Concrete
Pavements, FHWA-RD-95-110
[8] Talley, A. (2006). The Dan Ryan Expressway: A Look Back (and Forward) at the
Continuously Reinforced Concrete Pavement That Works. Proceedings of the
International Conference on Long-Life Concrete Pavements, Rosemont, Illinois October
25–27, 2006
[9] Won, M. (2005). Improvements of Testing Procedures For Concrete Coefficient of
Thermal Expansion, Proceedings Transportation Research Board 84th meeting
(CD-ROM), Washinton D.C. paper nr. 05-2521
[10] Won, M., Kim, D-H., Cho, Y-H., and Medina-Chavez, C. (2006). Long-Term Performance
of Continuously Reinforced Concrete Pavement in Texas. Proceedings of the
International Conference on Long-Life Concrete Pavements, Illinois October 25–27, 2006
139