Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home
Public Roads
Featuring developments in Federal highway policies, programs, and research and technology.
This magazine is an archived publication and may contain dated technical, contact, and link information.
Federal Highway Administration > Publications > Public Roads > Vol. 72 · No. 2 > Steel Versus GFRP Rebars?

Sept/Oct 2008
Vol. 72 · No. 2

Publication Number: FHWA-HRT-08-006

Steel Versus GFRP Rebars?

by Roger H. L. Chen, Jeong-Hoon Choi, Hota V. GangaRao, and Peter A. Kopac

Field studies show that glass fiber-reinforced polymer offers a low life-cycle cost option for reinforcement in concrete pavements.

This truck is traveling on the Nation's first GFRP-CRCP test section, constructed on a segment of Route 9 in West Virginia. The research team took this photograph while conducting a field observation on January 31, 2008.
This truck is traveling on the Nation's first GFRP-CRCP test section, constructed on a segment of Route 9 in West Virginia. The research team took this photograph while conducting a field observation on January 31, 2008.

Glass fiber-reinforced polymer rebar is one of the new products on the market that could offer a number of benefits to the transportation industry. Because it is lightweight and free of corrosion, construction costs should be lower and pavements should last longer. However, laboratory studies offer limited help in determining the real-world performance of glass fiber-reinforced polymer (GFRP) reinforcing bars in continuously reinforced concrete pavements (CRCPs). The reasons: the difficulties in modeling field boundary conditions, such as friction from the subbase and restraints from the shoulders or adjoining pavements; environmental changes; traffic loads; and possible variations in construction work. To overcome these limitations and gain a better understanding of GFRP-CRCP behavior, researchers turned to field investigations.

With support and cooperation from the Federal Highway Administration (FHWA), the West Virginia Department of Transportation (WVDOT), and contractors, West Virginia University (WVU) researchers recently completed the Nation's first GFRP-CRCP test section, along with a steel-CRCP test segment, to study the performance of the two rebar materials. The GFRP and steel test segments are located on Route 9 in Martinsburg, in the northeastern corner of West Virginia.

"The use of GFRP reinforcing bars in lieu of conventional steel reinforcement in CRCP, as demonstrated last fall [2007] in West Virginia, offers some interesting performance considerations," says Sam Tyson, concrete pavement engineer, FHWA. "First, the corrosion resistance of GFRP bars makes them attractive for obvious reasons, particularly in a State where winter conditions require frequent applications of deicer chemicals. In addition, the high tensile strength and low unit weight of GFRP, its matching thermal and matching stiffness characteristics, provide for a unique approach to the design and construction of CRCP. Finally, because GFRP is not magnetic, its use in concrete pavements where various traffic- and toll-monitoring devices are to be installed could be advantageous."

These qualities are clear advantages of GFRP, but the WVU study has not reached a conclusion regarding performance, including corrosion resistance, because not enough time has passed to obtain sufficient results. However, the study did show that GFRP-reinforced CRCPs can be constructed at low cost and without added construction time.

Overview of the Study

WVDOT allocated a 610-meter (2,000-foot)-long, two-lane section on Route 9 as the testing ground for the study. The experimental design incorporated two CRCP sections for comparison. The GFRP- and the steel-reinforced segments are both 305 meters (1,000 feet) long and 25 centimeters (10 inches) thick. WVU specified that both segments were to be constructed of concrete containing limestone coarse aggregate placed on a cement-stabilized subbase.

The contractor constructed the two experimental CRCP sections on September 25, 2007, and WVU monitored them continuously during the first 3 days to investigate the early-age cracking behavior. As the concrete cured during this period, WVU researchers recorded changes in concrete strain, reinforcement strain, and temperature. WVU researchers located, counted, and measured early-age cracks to estimate the spacing and width. The research team then analyzed and compared the data, along with additional crack data obtained about 1 month and 4 months after construction.

Members of the contractor's team are placing the subgrade layer during construction of the test segments.
Members of the contractor's team are placing the subgrade layer during construction of the test segments.

Design Details

Each CRCP section consists of two travel lanes with asphalt shoulders. A layer of subgrade, consisting of cement-treated aggregate, provides uniform support to the CRCP sections. On top of the subgrade, an open-graded, free-draining base course with # 57 aggregate serves as a subbase. The contractor stabilized the subbase with Type 1 portland cement to obtain erosion-resistant stabilized support below both sections.

For the GFRP-reinforced section, the design called for # 7 GFRP longitudinal rebars. For the steel-reinforced section, the design specified # 6 steel longitudinal rebars. In both test segments, the contractor placed the longitudinal rebars at the middepth of the slab.

For the transverse reinforcement that supports the longitudinal reinforcement, the contractor placed # 6 GFRP and # 5 transverse black steel rebars at 1.2-meter (4-foot) spacing. The contractor placed the transverse reinforcement on plastic chairs for the GFRP rebars and steel chairs for the steel rebars. Chairs are supports to keep rebars in their proper position during the placement of concrete.

Ensuring that adequate bond strength develops in the lapped splices of the longitudinal reinforcing rebars is important to prevent crack widening and subsequent structural failures. Therefore, a minimum splice length of 40 times the rebar diameter for GFRP and 25 to 30 times for steel is required, with at least three secure ties for each lap splice. Regular steel tie wires were used for the steel rebars and plastic zip ties for the GFRP. The contractor also staggered the lapped splices across the pavement to prevent localized strains in the slab.

The contractor used three wide-flange beam terminal joints between the two test sections and the abutting conventional jointed plain concrete pavement (JPCP) lanes on Route 9. A wide-flange beam joint is designed to accommodate rather than restrain movement of the free end of a CRCP slab. In a wide-flange beam joint system, the bottom of the beam is partially embedded in a reinforced concrete sleeper slab, the large horizontal slab that supports the ends of abutting pavements. The sleeper slab beneath the joint provides a large bearing area and additional support for the free ends. The steel flange helps protect the corners against spalling and aids in load transfer across the joint.

Concrete Mixes and Reinforcement Properties

For both test sections, the contractor used the same concrete mix design in accordance with Section 601 of the West Virginia Division of Highways Standard Specifications and Materials Procedure MP 711.03.23 for portland cement concrete. The contractor used Type I portland cement in the concrete mix along with Class F flyash. The coarse aggregate was # 57 limestone, and the fine aggregate was natural sand. The contractor also included an air-entraining admixture and a water-reducing admixture. The water-to-cement ratio was 0.42. The WVU designers specified that the concrete mix have a relatively high concrete strength to avoid excessively narrow crack spacings.

This photo shows the continuous reinforcement assemblies with GFRP rebars.
This photo shows the continuous reinforcement assemblies with GFRP rebars.

 

Shown here are the continuous reinforcement assemblies with steel rebars.
Shown here are the continuous reinforcement assemblies with steel rebars.

The GFRP rebar properties, provided by the FRP manufacturer, include a longitudinal elastic modulus (a measure of how rebar deforms) of 40.8 gigapascals, GPa (5.92 by 106 pounds per square inch, psi), and tensile strengths of 620.6 megapascals, MPa (90 kips per square inch, ksi) for # 6 rebar and 586.1 MPa (85 ksi) for # 7 rebar. The GFRP rebars are composed of calcium aluminosilicate glass fibers and a urethane-modified vinylester resin matrix with 70 percent minimum fiber content by weight. The contractor used typical grade 60 steel deformed rebars for the steel-CRCP section.

Construction

Concrete placement for the steel-CRCP section began at about 9:00 a.m. in an ambient temperature of about 20 degrees Celsius, oC (68 degrees Fahrenheit, oF). The contractor completed the steel-CRCP section at about 12:30 p.m. and then began concrete placement for the GFRP-CRCP section.

Properties of the Test Sections

Item

Steel-Reinforced CRCP

GFRP-Reinforced CRCP

Slab Width

7.32 meters (24 feet): two 3.66-meter (12-foot) travel lanes

7.32 meters (24 feet): two 3.66-meter (12-foot) travel lanes

Asphalt Shoulder Width

3.05 meters (10 feet) wide next to the right lane;

1.22 meters (4 feet) wide next to the passing lane

3.05 meters (10 feet) wide next to the right lane;

1.22 meters (4 feet) wide next to the passing lane

Longitudinal Reinforcement

# 6 rebar at 15.24-centimeter (6-inch) spacing

# 7 rebar at 15.24-centimeter (6-inch) spacing

Diameter of Longitudinal Rebars

1.91 centimeters (0.75 inch)

2.22 centimeters (0.875 inch)

Subbase

10.16-centimeter (4-inch)- thick open graded free draining base course stabilized with cement content of 9062 kilograms per cubic meter, kg/m3 (15065 pounds per cubic yard, lb/yd3)

10.16-centimeter (4-inch)-thick open graded free draining base course stabilized with cement content of 9062 kg/m3 (15065 lb/yd3)

Subgrade

27.31-centimeter (10.75-inch)-thick cement-treated aggregate subgrade

27.31-centimeter (10.75-inch)-thick cement-treated aggregate subgrade

Required Concrete Compressive Strength

40.7 megapascals, MPa (5,900 pounds per square inch, psi)

40.7 MPa (5,900 psi)

Longitudinal Coefficient of Thermal Expansion

11.88 microstrain per degree Celsius, µe/°C (6.6 microstrain per degree Fahrenheit, µe/°F)

6.58 µe/°C (3.66 µe/°F)

As the placement continued, the temperature of the subbase surface increased due to continuous sun exposure. The contractor measured the subbase surface temperature as about 39 oC (103 oF) at 1:30 p.m. To avoid temperature-related impairment of workability due to dry subbase aggregates absorbing water from the concrete mix and undesirable cracking from accelerated rates of moisture loss, the contractor sprayed water on the subbase from a sprinkler truck before placing the concrete. Workers completed both CRCP sections at about 6:30 p.m., when the ambient temperature was about 29 oC (85 oF).

The construction crews placed the CRCP sections using a slip-form paving machine. The machine was able to accommodate the entire width of the pavement. Agitator trucks delivered the concrete, and a conveyor belt distributed it to the center of the pavement lane. The crews finished the surface of the pavement slab immediately after the paving machine had passed.

Following the paving machine, a texturing/curing machine conducted two additional operations. The machine dragged burlap fabric to create microtextures on the finished surface and then tined the surface to obtain macrotextures to provide adequate friction for dry and wet weather. The texturing/curing machine then sprayed a curing compound on the textured surface to slow water evaporation from the concrete.

This first in a series of four photos shows the reinforcement layout of a sleeper slab during construction of a wide-flange beam terminal joint. Photo: WVU.
This first in a series of four photos shows the reinforcement layout of a sleeper slab during construction of a wide-flange beam terminal joint.


Here, the wideflange beam is partially embedded in the sleeper slab. Photo: WVU.
Here, the wideflange beam is partially embedded in the sleeper slab.


This third photo shows continuous steel (left) and GFRP (right) reinforcement assemblies placed on the sleeper slab. Photo: WVU.
This third photo shows continuous steel (left) and GFRP (right) reinforcement assemblies placed on the sleeper slab.


This fourth photo shows the completed wide-flange beam terminal joint system connecting the steel-CRCP and GFRPCRCP sections. Photo: WVU.
This fourth photo shows the completed wide-flange beam terminal joint system connecting the steel-CRCP and GFRPCRCP sections.

Experimental Instrumentation and Monitoring

The WVU researchers and the contractors tested the concrete mix to measure its properties in both fresh and hardened states. The contractors took concrete samples from the field and immediately measured the temperature, slump, and air content. At the same time, the WVU researchers cast 30 cylindrical concrete specimens for testing of compressive strength, tensile splitting strength, and elastic modulus at various ages, while casting three prismatic specimens for a drying shrinkage test.

WVDOT engineers also took cores about 4 months after construction; the average core compressive strength was almost 40 percent higher than that of the 28-day sample for both the steel-CRCP section (two core samples) and the GFRP-CRCP section (three core samples), although the GFRP compressive strength was slightly higher than that of the steel.

At about midlength of both CRCP sections, the researchers installed thermocouples and strain gages to investigate the first 3-day behaviors of each CRCP in terms of concrete temperature, concrete strain, and reinforcement strain. To set up a reference point and measure the strains in the longitudinal direction, the researchers created an artificial known transverse crack location. The WVU researchers placed a crack-inducer across each CRCP lane at a location where a set of thermocouples and strain gages was installed. The researchers attached an inverse T-shaped plastic crack-inducer on the subbase surface.

For the in situ temperature measurements, the WVU team installed 18 thermocouples at various depths and longitudinal locations. A thermocouple set consisted of three thermocouples and a metal stand. The researchers tied the thermocouples vertically to the stand, enabling temperature measurements to be made at 5, 13, and 20 centimeters (2, 5, and 8 inches) from the top of the pavement slab and glued the four legs of the metal stand to the subbase surface.

The research team placed five of the thermocouple sets at various longitudinal locations in the GFRP-CRCP section and one set in the steel-CRCP section. The researchers monitored the ambient temperature (with a standard thermometer), the surface (with an infrared thermometer), and interior temperatures (with thermocouples) of the concrete every 2 to 4 hours to attain a comprehensive understanding of the temperature variations at different locations over time under the influence of concrete hydration and ambient temperature.

The researchers installed eight concrete embedment strain gages to measure the concrete strain changes over time. The sensing grid of embedment gages, encased in polymer concrete, has an active gage length of about 10 centimeters (4 inches). A set of embedment gages included two gages and a metal stand tied together to measure the strains at two vertical locations: 5 centimeters (2 inches) from the top and bottom of the pavement slab.

The slip-form paving machine. Photo: WVU
The slip-form paving machine.

To avoid any effects from the slab edge, the researchers placed all the gage sets about 1.2 meters (4 feet) from the edge of the slab. Two data acquisition systems were used, one for the steel-CRCP section and the other for the GFRP-CRCP section, to collect the concrete strain data every 10 minutes throughout the first 3 days after the concrete placement.

The researchers attached a total of 10 general-purpose resistance strain gages to the reinforcements for measuring the longitudinal reinforcement strains in the steel- and GFRP-CRCP sections. The strain gages were self-temperature-compensated with respect to steel or GFRP rebar materials, so that the undesirable thermal outputs resulting from the mismatch in thermal expansion between the strain gage and the rebar material could be minimized. In each section, to avoid potential loss of field data because of gage malfunction, the researchers installed three reinforcement strain gages at the induced transverse crack location where the maximum reinforcement stress developed. The researchers also installed two gages at 25 centimeters (10 inches) and 0.9 meter (3 feet) longitudinally from the induced transverse crack location.

To protect the wires from the paving machine track, the researchers gathered them into an electrical conduit and embedded the conduit in a trench dug into the subbase. The conduit led the wires into electrical enclosures connecting to a data acquisition station. The thermocouple wires from two additional locations near the main data acquisition station in the GFRP section also were gathered into small electrical enclosures, which were embedded in the shoulder subbase. When the wire connectors were not in use, the researchers kept them inside the enclosures.

Here, workers distribute the concrete on the subbase. Photo: WVU.
Here, workers distribute the concrete on the subbase.

The researchers conducted visual surveys for transverse crack spacing and width over the first 3 days and then 1 month after the concrete placement. The team monitored a 122-meter (400-foot) midsection length and a 55-meter (180-foot) end-section (joint-section) in each CRCP section. They classified all cracks within the survey areas according to the location and date of their occurrence.

The researchers observed the cracks at the smooth face of the pavement edge, which had much clearer crack appearances. They measured the crack widths particularly from the upper corner of the pavement edge, which provides overestimated (or conservative) values compared to those from the driving surface. The largest concrete volume changes usually occurred at the upper corner of the pavement edge, where there was less restraint from reinforcement and subbase friction. The changes in crack width at this location should be larger than if measured at other locations.

Experimental Results at 7, 28, and 38 Days, and 4 Months

Test

Steel-Reinforced CRCP

GFRP-Reinforced CRCP

Average Compressive Strength at 7 Days (tested at WVU)

19.7 MPa (2,850 psi)

19.7 MPa (2,850 psi)

Average Compressive Strength at 28 Days (tested at WVU)

26.9 MPa (3,900 psi)

26.9 MPa (3,900 psi)

Average Compressive Strength at 4 Months (tested at WVDOT and WVU)

37.6 MPa (5,450 psi)

37.9 MPa (5,500 psi)

Midsection Cracks at 3 Days

45

19

Midsection Cracks at 38 Days

75

40

Midsection Average Crack Spacing at 3 Days

2.88 meters (9.44 feet)

6.91 meters (22.67 feet)

Midsection Average Crack Spacing at 38 Days

1.71 meters (5.61 feet)

3.31 meters (10.86 feet)

Midsection Average Crack Width at 3 Days

0.025 centimeter (0.01 inch)

0.043 centimeter (0.017 inch)

Midsection Average Crack Width at 38 Days

0.028 centimeter (0.011 inch)

0.053 centimeter (0.021 inch)

Maximum Crack Width on January 31, 2008

0.058 centimeter (0.023 inch)

0.086 centimeter (0.034 inch)

To measure the crack width, the researchers used a magnifying glass and a crack comparator, which is a transparent ruler printed with graduations at different widths. A GFRP-CRCP crack observed on the third day and again 125 days later showed the maximum crack width in the GFRP-CRCP test section as 0.058 centimeter (0.023 inch) on the third day and 0.086 centimeter (0.034 inch) on the 125th day.

Observation of Crack Spacing and Width at Early Age

All of the cracks in the concrete were transverse, with no longitudinal cracks observed. The anticipated absence of longitudinal cracks is due to designing for a longitudinal joint that limits the lane width to 3.7 meters (12 feet) and thus reduces the likelihood of cracking in that direction. The researchers evaluated maximum, average, and minimum values of transverse crack spacing and crack width for each CRCP section for each date when a measurement was made. After the construction, the team traced the width of each crack at four different ages in order to observe the changes in crack width over time.

WVU researchers installed a crackinducer and strain gages in the GFRPCRCP section, as shown here.
WVU researchers installed a crackinducer and strain gages in the GFRPCRCP section, as shown here.

A drastic decrease in the average crack spacing occurred between the first and second days because a number of cracks were generated due to a combination of a large change in the concrete volume and low concrete strength, which are both inherent at this early age. The crack spacing for the GFRP-CRCP section was larger than that for the steel-CRCP section, due to the lower stiffness of the GFRP reinforcement. Using GFRP rebars as reinforcement can reduce undesirable stress development in concrete caused by mismatches in stiffness and thermal expansion between steel reinforcement and the concrete. Steel stiffness is about six times larger than concrete or GFRP. The improved compatibility can be beneficial depending on other CRCP-design factors to control the crack width and spacing, such as reduction of stresses surrounding the reinforcement at a crack location that may cause spalling or punchout failure in CRCP. As expected with the terminal joints, which allow movement of the free end of the CRCP slab, the average crack spacing at the joint-section was larger than at the midsection.

As for the crack width, the researchers observed mixed results. The widths remained unchanged or even became smaller during the second day and then started increasing. The cracks found on the first day generally had the larger widths, while additional cracks found at later ages had smaller widths due to less change in the concrete volume. "We believe that the restraining stress in the concrete was probably released when the additional cracks occurred, narrowing the widths of the existing cracks," says William "Bill" Shanklin, area construction engineer, West Virginia Division of Highways.

The researchers placed this embedment strain gage set directly above a crack-inducer in the steelreinforced CRCP section. Photo: WVU.
The researchers placed this embedment strain gage set directly above a crack-inducer in the steelreinforced CRCP section.

The researchers found more new cracks on the second day than on the later days. From the third day and beyond, crack width started slowly increasing due to continuous, yet less drastic, concrete shrinkage. Even though the crack width for the GFRP-CRCP section was larger due to the larger crack spacing and lower stiffness of reinforcement, it still meets the American Association of State Highway and Transportation Officials' (AASHTO) limiting criterion for crack width -- # 0.1 centimeter (0.04 inch) -- which is of utmost importance in providing adequate aggregate interlock and ensuring the integrity of the pavement. In addition, the crack widths in the joint-section appear to be smaller than those in the midsection, due to the lower restraining stresses developed in the joint-section.

The thermocouple set shown here is installed in a GFRP-CRCP section. Photo: WVU.
The thermocouple set shown here is installed in a GFRP-CRCP section.

Currently, both CRCP sections are open to traffic. According to field observation on January 31, 2008, the maximum crack width for the GFRP-CRCP section and the steel-CRCP section met the current AASHTO limiting criterion, even though the guideline was based on experience and understanding gained from steel-reinforced CRCP. Limiting criteria, such as crack spacing, crack width, and reinforcement stress level for GFRP-reinforced CRCP still need to be developed.

Suggestions for Future Research

Additional studies on the performance of GFRP-reinforced CRCP in response to traffic loading are needed. The lessons taken from this short-term field study suggest that future research is needed to make further improvement of the design for GFRP-reinforced CRCP, if such improvement proves necessary after long-term traffic loading. Periodic observations of the load transfer efficiency at cracks, crack spacing and width under traffic loading, a crack width profile throughout the slab depth under loading, and pavement distresses are essential to obtain a comprehensive understanding of the overall performance of GFRP-reinforced CRCP. This understanding eventually will assist in developing standard design guidelines for future GFRP-reinforced CRCP.

The researchers placed these three embedment strain gage sets in the GFRP-CRCP section. Photo: WVU.
The researchers placed these three embedment strain gage sets in the GFRP-CRCP section.

In terms of CRCP life-cycle costs, the current expectation is that GFRP section costs will be substantially lower than those of steel sections. The long-term maintenance cost would be lower for the GFRP-CRCP than the steel-CRCP because there will be no structural distresses caused by reinforcement corrosion.

Two WVU researchers are measuring the temperature recorded by a small electrical conduit embedded in the shoulder subbase. Photo: WVU.
Two WVU researchers are measuring the temperature recorded by a small electrical conduit embedded in the shoulder subbase.

The data on early-age performance from the GFRP-CRCP field test section compares favorably with those from the steel-CRCP section. With additional construction experience using GFRP-reinforced CRCP and improvements in GFRP-CRCP design, even better performance should be achievable.

The researchers placed a ballpoint pen on the GFRP-CRCP section to show the scale of this crack width of 0.058 centimeter (0.023 inch) on the third day (September 28, 2007) after concrete placement. The rusty-orange paint beneath the pen marks the location of each crack found on the first day after concrete placement, and an identification number is written on the red tape.
The researchers placed a ballpoint pen on the GFRP-CRCP section to show the scale of this crack width of 0.058 centimeter (0.023 inch) on the third day (September 28, 2007) after concrete placement. The rusty-orange paint beneath the pen marks the location of each crack found on the first day after concrete placement, and an identification number is written on the red tape.

 

The GFRP-CRCP section showed a crack width of 0.086 centimeter (0.034 inch) 128 days after placement (January 31, 2008). This crack is the same one pictured in the previous photo. When the shoulder was placed, the contractor removed all the paint and tapes, but the locations were recorded by construction station number.
The GFRP-CRCP section showed a crack width of 0.086 centimeter (0.034 inch) 128 days after placement (January 31, 2008). This crack is the same one pictured in the previous photo. When the shoulder was placed, the contractor removed all the paint and tapes, but the locations were recorded by construction station number.

Roger H. L. Chen, Ph.D., is a professor of civil engineering at West Virginia University (WVU), Morgantown. He has been involved extensively with research in structural dynamics, nondestructive evaluation (NDE), dynamic soil-structure interaction, and material characterization of concrete, composites, and timber and ceramic materials for about 25 years and has ongoing research projects in GFRP-reinforced CRCP, self-consolidating concrete, evaluation of bridges for transporting coal, and diagnostics of thermal barrier coatings. He serves on several technical committees for the American Concrete Institute, American Society of Civil Engineers, and American Society for Nondestructive Testing (ASNT) related to concrete, NDE, FRP, dynamics, and experimental analysis. He received his Ph.D. from Northwestern University and is a fellow of ASNT.

Jeong-Hoon Choi is a graduate research assistant in the Department of Civil and Environmental Engineering at WVU. He received his undergraduate degree in civil engineering from Hanyang University, Republic of Korea, and a master's degree in civil engineering from WVU. His Ph.D. research is related to the design and application of GFRP-CRCP.

Hota V. GangaRao is a professor of civil engineering and director of the Constructed Facilities Center at WVU. He is a fellow of ASCE and serves on many technical committees of professional societies.

Peter A. Kopac is a senior research highway engineer on the Pavement Design and Performance Modeling team of FHWA's Office of Infrastructure Research and Development. He has almost 40 years of highway-related experience, including 31 years with FHWA. Kopac has managed, monitored, and contributed to numerous research studies dealing with concrete and concrete pavements.

This research is funded by FHWA through the Constructed Facilities Center at WVU. For more information, contact Roger H. L. Chen at 304-293-3031, ext. 2631, or roger.chen@mail.wvu.edu, Jeong-Hoon Choi at 304-293-3031, ext. 2434, jchoi1@mix.wvu.edu, Hota V. GangaRao at 304-293-3031, ext. 2634, hota.gangarao@mail.wvu.edu, or Peter A. Kopac at 202-493-3151, peter.kopac@fhwa.dot.gov. Also see www.fhwa.dot.gov/pavement/pccp/pubs/05081/05081.pdf.

ResearchFHWA
FHWA
United States Department of Transportation - Federal Highway Administration