Investigation of In Situ Thermal Properties and Early-Age Pavement Behavior in the Design and Performance Evaluation of Roller Compacted Concrete Pavement in Louisiana,U.S

Abstract

The proven durability and high load-carrying capacity of roller compacted concrete pavement (RCCP), combined with its simple and cost-effective construction method and high placement speed, has created a great deal of interest from many states and local transportation agencies in the U.S. Roller compacted concrete (RCC) mixture uses less cement content and less water compared with conventional concrete mixtures, which reduces the total shrinkage strain and RCC set temperature during the hardening stage, resulting in a reduction of early-age deformation and stress developments in the RCCP. Currently, research on the early-age behavior and other thermal properties of RCCP has not been well documented. Therefore, investigations of early-age behavior and thermal properties such as coefficient of thermal expansion, ultimate shrinkage, and built-in curling are needed to understand the true behaviors of RCCP under real climatic conditions. On the other hand, the currently available RCC thickness design procedures, for example, Street-Pave and Pavement Designer, are in general short of flexibility in consideration of the combined effect of wheel and environmental loading on RCCP performance. The primary objective of this work is to investigate the early-age behavior of RCCP and to evaluate the effect of thermal properties on the design and performance of RCCP constructed in Louisiana.

Keywords

infrastructure pavements design and rehabilitation of concrete pavements roller compacted concrete pavements

Roller compacted concrete (RCC) is a zero-slump concrete mixture placed with asphalt-type paving equipment and compacted by vibratory rollers. RCC is an economical, fast, and durable candidate for many pavement applications. In the early application, RCC has traditionally been used for pavements carrying heavy loads in low-speed areas, such as parking, storage areas, ports, airport service areas, intermodal, and military facilities ( 1 ). With improved paving and compaction methods as well as surface texturing techniques, recent applications of RCC are found for interstate highway shoulders, city streets, and other highways ( 2 – 6 ). According to a survey conducted between 2014 and 2016 in United States, over 265 projects have been paved covering more than 5.1 million square yards (4.3 million square meters), and the past 3 years have seen the placement of over 1.6 million square yards (1.3 million square meters) of RCC each year ( 7 ).

Even though the long-term behavior and performance of RCCPs have been reported, studies on the early-age behavior of RCCPs are rare. The early-age load-carrying capacity of RCCP is mainly achieved by roller compaction in addition to cement hydration. Because of the early strength gain, the early-age behavior of RCCP will be different from conventional concrete pavement ( 1 ).

Moreover, because of low water-to-cement ratio, RCCP has reduced shrinkage as compared with conventional concrete pavements ( 8 ). The shrinkage cracks that occur in RCCP are usually small (less than 0.1 in.) and very good load transfer exists across the crack through aggregate interlock. This aggregate interlock is enhanced through the use of the dense-graded aggregate structure specified for RCC mixtures ( 2 ). The lower water-to-cement ratio, along with lower coarse aggregate content in the RCC mixture as compared with conventional concrete pavements, will also result in a lower coefficient of thermal expansion (CTE), since CTE is greatly influenced by the types and volumetric proportion of coarse aggregate in the mixture. Several studies found that increased CTE generally resulted in increasing cracking, joint faulting, and International Roughness Index (IRI) in concrete pavements, even in an increase at the smaller CTE values (9–11). It was also widely accepted that concrete pavements constructed with concrete that has a high CTE would experience a larger amount of curling because of temperature effects. As a result of having reduced shrinkage and lower CTE, the built-in curl (also known as construction curl), along with the curling stresses of RCCP under daily temperature fluctuation, will be different from conventional concrete pavements, which will also affect the RCCP design in relation to pavement performance. Several studies were conducted only on the material properties of RCC in a laboratory (12, 13). Therefore, field evaluation of early-age behavior and thermal properties of RCCP subjected to environmental loads is needed to better understand the behavior of RCCP from the design point of view.

In this study, the early-age behaviors of RCCP constructed in Louisiana, U.S., conditions were investigated using a computer program—HIPERPAV, developed under Federal Highway Administration (FHWA) research—to predict the early-age behavior of concrete pavements ( 14 ). In addition, a mechanistic-empirical (ME)-based RCC thickness design procedure is considered based on the modified Portland Cement Association (PCA) thickness design method to quantify the combined effect of traffic and environmental loading for RCCP. For the RCC design procedure, a recently developed fatigue model based on in situ saw-cutting beam samples is being considered as a replacement for the current American Concrete Pavement Association (ACPA) PCC fatigue model ( 15 ). The effect of drying shrinkage, built-in curling, CTE, and joint spacing was evaluated in the RCCP damage analysis.

Project Details

Louisiana has many miles of low-to-medium-volume roadways currently used by a significantly large amount of heavy and over-loaded trucks from shale oil/gas industries, logging, and agricultural activities. Because of the heavy trafficking, the pavements experienced significant pavement distress that typical maintenance activities cannot sufficiently address. An accelerated pavement testing (APT) study in the Pavement Research Facility (PRF) in Louisiana showed outstanding structural performance and load-carrying capacity of RCCP over soil cement, and proposed RCC as a design alternative for those low-to-medium-volume roadways undergoing heavy trafficking. Based on the findings from the APT study, RCC was used for a low-volume road undergoing heavy trafficking in Lafayette Parish, marking the first usage of RCCP in the roadway system in Louisiana. The RCCP (Case I: APT RCC test sections and Case II: RCCP in Lafayette Parish) were investigated for early-age behavior and other thermal-related properties subjected to environmental and traffic loadings. More details about both cases can be found in the Louisiana Transportation Research Center (LTRC) final report project no 12-7P and 19-1P ( 16 ).

Case-I: APT Test Section

Two 8 in. RCC sections (each 72 ft long and 12 ft wide) over treated and stabilized soil bases were constructed at the PRF in Louisiana and assessed under accelerated load testing. The test sections had saw-cut joints spaced at 20 ft intervals with a joint depth of 1.5 in. The RCC mixture used in the APT sections includes a type I Portland cement, a #67 crushed limestone, and manufactured sand. The proposed mixture contained 43% coarse aggregate and 57% fine aggregate by weight. The mix design was developed for a 4,000 psi 28-day design compressive strength. The mixture proportions used in the construction of the test lanes are shown in Table 1 below:

Table 1.

Mixture Proportion of Roller Compacted Concrete (Case I)

Material	Quantity (lb/yd³)
Cement	450
Fine aggregate	2,107
Coarse aggregate	1521
Water	154
Water-to-cement ratio	0.34

Case II: RCCP in Lafayette Parish

The project is located in an industrial area of Lafayette Parish under the jurisdiction of Lafayette Consolidated Government (LCG). The project was introduced as a connectivity project, to connect several dead-end roads to existing roadway facilities. There were three different locations, Decal Street, Sage Glenn Lane, and Denbo Street, where the RCCPs were constructed. All the lanes were paved with a 6 in. RCC on top of an 8 in. soil cement base layer. The saw-cut joints were cut 1/8th of an inch wide by 1/3rd of the pavement depth spaced at 10 ft intervals for all three pavement sections. No joint sealing was performed on any of those saw-cut joints. The RCC mixture considered in this project includes a type I Portland cement, #689 Mexican limestone, and natural sand. The proposed mixture contained 54% coarse aggregate and 46% fine aggregate by weight. The mix design was developed for a 4,000 psi 28-day design compressive strength. The mixture proportions used in the construction of the RCCPs are shown in Table 2 below:

Table 2.

Mixture Proportion of Roller Compacted Concrete (Case II)

Material	Quantity (lb/yd³)
Cement	550
Fine aggregate	1,483
Coarse aggregate	1,797
Water	174
Water-to-cement ratio	0.32

Early-Age Behavior of RCCP

Cracking is one of the primary design and construction concerns for concrete pavements. Uncontrolled cracks shorten the effective life of the pavement. Therefore, it is particularly important to prevent early-age uncontrolled cracks. In this context, early-age is defined as the first 72 h after placement. Because of the low water-to-cement ratio, different construction techniques, roller compaction, and early strength gain, the early-age behavior of RCCP is expected to be different from the that of conventional concrete pavement. To evaluate the early-age behavior of RCCP, both cases were investigated using the computer program HIPERPAV. HIPERPAV is a user-friendly, powerful analysis tool to assess how a variety of factors, such as environment, construction, design, and mixture proportion, affect strength and stress development in the concrete pavement. Both cases have different mixture proportions and are constructed at different times of the day and under different seasonal conditions. The APT RCC test sections were constructed in wintertime with a high temperature of 60°F and a low temperature of 37°F on the construction day. The RCCP in Lafayette Parish was constructed in springtime with a high temperature of 86°F and a low temperature of 64°F on the construction day. Figure 1 shows the early-age cracking prediction of RCCPs using HIPERPAV for both cases. The solid blue line noted on the graphs indicates the tensile strength development of concrete in the first 72 h after placement. The lighter yellow area indicates the critical tensile stress development at the top of the slab and the lighter blue area represents the critical tensile stress development at bottom of the slab during the same time. If the tensile stresses exceed the tensile strength of concrete, cracks could be expect.

Figure 1.

Early-age behavior of roller compacted concrete pavement (RCCP): (a) Case I: tensile stresses in the pavement exceed the tensile strength of concrete (early-age cracking is expected) and (b) Case II: tensile stresses in the pavement exceed the tensile strength of concrete (early age cracking is expected).

A sensitivity analysis of the construction time and initial mix temperature was also examined. Figure 2 shows the HIPERPAV sensitivity summary screen for both cases.

Figure 2.

Sensitivity analysis of roller compacted concrete (RCC) early-age behavior: (a) sensitivity analysis of initial mix temperature for Case I, (b) sensitivity analysis of initial mix temperature for Case II, (c) sensitivity analysis of construction time for Case I, and (d) sensitivity analysis of construction time for Case II.

According to the HIPERPAV output results, early-age cracking of RCCP is expected for both cases. From the sensitivity analysis it can also be seen that, no matter what the initial mix temperature was, the pavement will exhibit early-age cracking for Case II, unless it was constructed in the evening or nighttime. However, in reality, no early-age cracking was observed in any of the RCCP for either case. The reason could be because of the early strength gained through roller compaction, which created friction between the confined particles (aggregate interlock), whereas conventional concrete is in a plastic state after placement until hydration begins to harden the paste and bind the aggregates together, as shown in Figure 3. This implies that RCCP may not be susceptible to early-age cracking. Further studies need to be conducted to confirm the early-age behavior of RCCP.

Figure 3.

Strength development and load-carrying capacity of roller compacted concrete pavement (RCCP) and Portland Cement Concrete (PCC) pavement after placement ( 1 ).

Permanent Built-In Curl or Equivalent Temperature Difference of RCCP

Built-in curl (also known as construction curl) is caused by the presence of a temperature gradient at the time of the initial set of concrete. Proper estimation of the amount of built-in curl is very important to evaluate fatigue damage and faulting which can be expected to occur in the pavement. Small changes in the amount of built-in curl can significantly change the predicted pavement performance. Built-in curl is usually quantified by the temperature gradient required to deform a similar, but theoretically flat slab, to the same shape as the actual slab; this temperature gradient is known as the “equivalent temperature difference.” In most cases, concrete pavements are cast during the daytime, in the presence of a positive temperature gradient. For this reason, the pavement is usually said to have a built-in negative temperature gradient indicating the pavement will only resume a flat shape when a sufficient positive temperature gradient is present to counteract the effective negative gradient that was built into the slab. Several studies showed that the distress predicted by the design software PavementME strongly depends on the value of the equivalent temperature difference. Currently, a default value of −5°C is used in the design software for the permanent curl/warp effective temperature difference between the top and the bottom of the slab for jointed plain concrete pavement ( 17 , 18 ). For RCCP, the built-in curl or the equivalent temperature gradient is expected to be different than that of conventional concrete pavement, for the same reason of having lower water-to-cement ratio, lower CTE, and drying shrinkage.

The equivalent temperature difference or built-in curl of RCCP was investigated from the APT RCCP test sections instrumentation results. APT RCCP test sections were retrofitted with a thin polymeric plate positioned perpendicularly to the loading direction. The plate was instrumented with fiber optic gages and fixed inside a thin saw cut in the RCC layer with a slow-curing epoxy glue. Each plate has 16 fiber optic strain gages and 3 temperature gages at different depths. The sensors were positioned apart along the plate to measure critical strains under dual tire accelerated loading. More details about the instrumentation can be found elsewhere ( 19 ).

The equivalent linear temperature gradient (ELTG) within the RCC slab was obtained from the measured pavement temperature based on the moment equilibrium concept, which considers the actual nonlinear temperature distribution ( 20 , 21 ). Temperature moment was calculated using Equations 1 and 2. The temperature moment was then used to find the ELTGs using Equation 3:

WAT = \sum_{i = 1}^{n} \frac{0.5 (T_{i} - T_{i + 1}) \times (D_{i} - D_{i + 1})}{D_{slab}}

(1)

T_{mo} = \sum_{i = 1}^{n} [0.5 (T_{i} - T_{i + 1}) \times (D_{i}^{2} - D_{i + 1}^{2})] - 2 (D_{i}^{2} - D_{i + 1}^{2})

(2)

ELTG = \frac{- 12 \times T_{mo}}{D_{slab}^{3}}

(3)

where

WAT = weighted average temperature,

$T_{mo}$ = temperature moment,

$T_{i}$ = temperature at depth i,

D = slab thickness, and

n = number of computation nodes.

As the RCC slab is subjected to a set of temperature and moisture gradients throughout the testing period, the slab strains vary along the thickness causing the slab to curl either upward or downward. Based on the top and bottom fiber optic strain responses, the daily changes in curvature are also calculated using Equation 4:

Curvature = \frac{(ε_{bot} - ε_{top})}{D * (1 + ε_{bot} * ε_{top})}

(4)

The daily curvature change, along with calculated ELTG, was then investigated. Figure 4 shows that a hysteresis loop occurs on the ELTG versus curvature plot. This phenomenon has been experienced by other researchers for rigid pavement and explained by the possible moisture movement along the slab ( 22 ). It can be also seen that the equivalent temperature gradient is approximately −1.18°C/in. when the slab curvature is zero (assumed to be flat).

Figure 4.

Change in curvature with equivalent linear temperature gradient (ELTG).

The equivalent temperature gradient for RCCP was also calculated using the current PavementME equivalent temperature gradient estimation model shown in Equation 5. Based on the equation, the equivalent temperature gradient was estimated to be −1.38°C/in., which is slightly higher than the in situ estimation. This indicates that the RCC slabs have relatively fewer slab deformations resulting from daily temperature fluctuations compared with conventional concrete pavement. That is because of the relatively low water-to-cement ratio of the RCC mixture.

\begin{matrix} deltaT / inch = - 5.27805 - 0.00794 * TR \\ - 0.0826 * SW + 0.18632 * PCCTHK \\ + 0.01677 * uw + 1.14008 * w / c + 0.01784 * latitude \end{matrix}

(5)

where

deltaT/inch = predicted gradient in RCCP slab (^oF/in.),

TR = difference between the maximum and minimum temperature in construction month (^oF),

SW = slab width (ft),

PCCTHK = slab thickness (in.),

uw = unit weight of PCC (lb/ft³),

w/c = water-to-cement ratio, and

latitude = latitude of the project location (degrees).

To further investigate the equivalent temperature gradient for RCCP, a 3D finite element (FE) model was developed using ABAQUS software to investigate the strain responses of RCCP under accelerated loading. Two analysis steps are included in the FE numerical simulation: 1) in the first step, the temperature measured from the field is applied as boundary conditions on the top and bottom surface of the RCC slab—the slab is curling or wrapping because of the thermal gradient; 2) in the second step, duel-tire load is applied with the existing temperature and strain distribution obtained from the previous step. The difference of the strain results from the two steps is used to calculate the strain response and verified with the instrument responses. The strain responses at the top and bottom RCC layers on the RCC test sections with various tire load levels were predicted and plotted against the measured responses in Figure 5.

Figure 5.

Measured versus predicted strain responses under accelerated loading: (a) bottom transverse strain and (b) top transverse strain.

The simulation results matched very well with the measured bottom and top transverse responses, especially for the critical transverse strain located under the tire loading area. The simulation results also showed that, with no temperature load, the FE model cannot predict the true in situ pavement responses. For the temperature effect, a constant CTE of 6.85 με/°C and a negative temperature gradient of −9°C were considered to match with the in situ responses. This also validates the field estimated equivalent temperature gradient of −1.18°C/in. for an 8 in. RCCP test section at a zero-curvature slab condition.

In Situ Coefficient of Thermal Expansion (CTE) of RCCP

The CTE of concrete pavement is a very important parameter in pavement design because the temperature-related pavement deformations are directly proportional to this value during early ages as well as during the pavement design life. A higher concrete CTE would generate relatively large thermal stresses and wide crack widths. It is difficult to cast RCC laboratory samples to conduct laboratory CTE testing following AASHTO TP 60 test procedure ( 23 ). The in situ strain responses subjected to only environmental loading were used to predict the in situ CTE for RCCP. Figure 6 shows the in situ strain responses at the top and bottom of RCCP subjected to environmental loading.

Figure 6.

Static strain response with temperature variation: (a) top strain and (b) bottom strain.

Figure 6 confirms that the top of the RCC slab was subjected to a higher temperature and has a more rapid cooling rate than the bottom of the slab. However, the bottom layer of the RCC slab, for both sections, showed a lower thermal expansion coefficient than the top of the slab, which can be explained by the effects of moisture conditions and cooling rate. CTE in concrete pavement usually changes with depth because of the saturation and non-linear temperature profile ( 22 ). Several studies indicated that the CTE value is usually higher in a partially saturated condition (usually top of slab) compared with a fully saturated condition (bottom of slab), which aligns with the findings from the in situ strain responses for RCC slabs ( 24 ). The in situ thermal expansion coefficient values for RCCPs ranged from 5.0 to 7.5 µε/°C with an average of 6.11 µε/°C. The FE model can predict the pavement responses under the combined effect of wheel and environmental loading very closely for a constant CTE value of 6.85 µε/°C, which is within the range of the measured in situ CTE values. The measured in situ CTE value for RCCP is significantly lower than the current PavementME default value of 9.45 µε/°C for the conventional concrete pavement with limestone aggregate.

Optimum Joins Spacing for RCCP

Joint spacing for RCCP may be crucial to ensure adequate crack opening and load transfer across the cracks. Excessive large crack spacing will result in wider opening of the crack and reduce the load transfer. Currently, PavementDesigner does not include joint spacing as a design input variable for RCCP, but instead calculates the recommended joint spacing as 5.25 times the radius of relative stiffness. As an alternative, ACPA recommends joint spacing based on slab thickness and support condition. None of the above recommendations considered shrinkage or thermal expansion as a variable ( 25 ). According to Darter and Barenberg, joint spacing in concrete pavements depends more on the shrinkage characteristics of the concrete rather than on the stress in the concrete ( 26 ). The joint spacing can be calculated based on the following Equation 6 considering an allowable joint opening of 0.1 in.

L = Δ L / (C * (α Δ T + ε))

(6)

where

△L = joint opening,

L = joint spacing or slab length,

△T = temperature range (temperature at placement minus the lowest mean monthly temperature), and

C = adjustment factor because of slab-subbase friction (0.65 for stabilized base).

The ultimate drying shrinkage of RCCP can be calculated based on the water content and 28-day compressive strength using the following empirical Equation 7:

ε = C 1 * C 2 {26 W^{2.1} (f c')^{- 0.28} + 270}

(7)

where

ε= ultimate drying shrinkage strain*10^-6,

W = water content for the mix under consideration (lb/ft³),

fc′ = 28-day compressive strenght (psi),

C1 = cement type factor (defined as 1.0 for Type I cement, 0.85 for Type II cement, and 1.1 for Type III cement), and

C2 = type of curing factor (defined as 0.75 if steam cured and 1.0 if cured in water).

Considering the in situ measured CTE value and calculated drying shrinkage of 360* $10^{- 6}$ based on mixture proportion, an optimum joint spacing of 20 ft can be predicted for an allowable joint opening of 0.1 in.

Effect of Temperature Curling and Moisture Warping

It is very important to consider curling stress and moisture warping in RCCP thickness design, because curling stress may be quite large and cause the slab to crack when combined with only very few numbers of load repetitions. For daytime curling conditions, compressive curling stresses are induced at the top of the slab, whereas tensile stresses occur at the bottom, or vice versa for night-time curling conditions. The moisture gradient in concrete slabs also results in additional warping stresses. Therefore, the effect of thermal curling stress needs to be included in the RCCP design.

To incorporate the thermal curling effect in the RCCP design, the modified equivalent stress concept proposed by Lee et. al. has been adopted for this study based on the following Equations 8 to 10 ( 27 ):

σ_{eq} = (σ_{w} * R_{1} * R_{2} * R_{3} * R_{4} * R_{5} + R_{T} * σ_{c}) * f_{3} * f_{4}

(8)

\begin{matrix} σ_{w} = \frac{3 (1 + μ) P}{π (3 + μ) h^{2}} [[\log \frac{E h^{3}}{100 * a^{4}} + 1.84 - \frac{4}{3} μ + \frac{1 - μ}{2} + 1.18 (+ 2 μ) \frac{a}{l}] \end{matrix}

(9)

σ_{c} = \frac{CE α Δ T}{2} {1 - \frac{2 \cos λ \cosh λ}{\sin 2 λ + \sinh 2 λ} (\tan λ + \tanh λ)}

(10)

λ = \frac{W}{(\sqrt{8} l)}

where

$σ_{eq}$ = modified equivalent stress (psi),

$σ_{w}$ = Westergaard’s closed-form edge stress solution (psi),

h = slab thickness,

λ = $\frac{W}{(\sqrt{8} l)}$ ,

C = curling stress coefficient,

R ₁ = adjustment factor for gear configurations,

R ₂ = adjustment factor for the finite slab length and width,

R ₃ = adjustment factor for a tied concrete shoulder,

R ₄ = adjustment factor for a widened outer lane,

R ₅ = adjustment factor for a bonded/unbonded second layer, and

R _T = adjustment factor for the combined effect of loading plus daytime curling.

In this procedure, the temperature-induced stresses are analyzed separately from the load-induced stresses. However, these stresses cannot be simply added together unless the curling and load stress analyses have the same boundary conditions. Because the slab curls off its support, there is some error associated with the superposition assumption and therefore a correction factor must be applied. The determination of the correction factor R to account for temperature curling stress has been studied by several researchers for rigid pavements ( 28 – 31 ).

The temperature stresses and R factor can be determined from an equivalent effective temperature gradient with a percent time of occurrence, or with a temperature, differential distribution based on temperature fluctuation data. The input can be either a cumulative damage analysis using the full temperature differential distribution or a single effective temperature gradient value and the equivalent percent time of occurrence. To illustrate the importance of incorporating the effect of thermal curling in the thickness design of RCCPs, a similar set of design inputs was used for load only and load plus curling for an equivalent effective temperature gradient with a percent time of occurrence. It is assumed that an equivalent effective temperature gradient of 0.5°C/in. along the RCC thickness with a 10% time of occurrence will represent the full temperature differential distribution during the design period. Because of the necessity to have a stronger base under RCCP, an unbonded base is also considered in this analysis. Table 3 shows the effect of load only and load plus curling on the pavement damage on RCCPs.

Table 3.

Effect of Load only and Load plus Curling on Roller Compacted Concrete Pavement (RCCP) Damage

Thickness (in.)	Average daily truck traffic	RCC modulus, E (ksi)	RCC flexural strength (psi)	Base thickness (in.)	Base modulus (psi)	Subgrade, k (pci)	% Damage (load only)	% Damage (load + curling)
4	2,000	4,000	800	8	200,000	150	92.3	106.3
6	2,000	4,000	800	8	200,000	150	13.6	34.9
8	2,000	4,000	800	8	200,000	150	0.0023	0.04

Note: RCC = roller compacted concrete; psi = pounds per square inch; pci = pounds per cubic inch.

The results of this fatigue analysis show that, for a 4 in. RCCP, a total of 92.3% fatigue damage was caused by 100% load repetition only, whereas a total of 106.3% of fatigue damage could be induced by 90% of load repetitions plus only 10% daytime curling. In this case, an additional 1/2 in. of slab thickness may reduce the total cumulative fatigue damage to less than 100%. Similarly, the damage on 6 in. and 8 in. RCCP also showed increased damage for considering both load and curling stresses. However, the increase in damage caused by curling stresses in the thicker pavement is more significant.

It can be seen that the CTE and saw-cut joint spacing have significant effects on the RCCP damage as well. As shown in Table 4, the cumulative fatigue damage increases significantly with an increase in CTE value. Therefore, it is very important to select the appropriate CTE for damage analysis of RCCP.

Table 4.

Effect of Coefficient of Thermal Expansion (CTE) on Roller Compacted Concrete Pavement Damage

Thickness (in.)	Average daily truck traffic	RCC modulus, E (ksi)	RCC flexural strength (psi)	Base thickness (in.)	Base modulus (psi)	Subgrade, k (pci)	CTE (/^oC)	% Damage
4	2,000	4,000	800	8	200,000	150	4.00E-06	106.33
							5.50E-06	113.34
							7.00E-06	126.98
6	2,000	4,000	800	8	200,000	150	4.00E-06	34.90
							5.50E-06	67.18
							7.00E-06	148.17
8	2,000	4,000	800	8	200,000	150	4.00E-06	0.031
							5.50E-06	0.116
							7.00E-06	0.436

Note: RCC = roller compacted concrete; ksi = kips per square inch; psi = pounds per square inch; pci = pounds per cubic inch.

Table 5 shows the effect of joint spacing in the RCCP critical stress analysis. For a given pavement structure, traffic condition, and CTE value, the longer joint spacing will exhibit higher critical stress compared with a shorter joint spacing, indicating higher cumulative fatigue damage.

Table 5.

Effect of Joint Spacing on Roller Compacted Concrete Pavement Design

Thickness (in.)	Coefficient of thermal expansion (/^oC)	Joint spacing (in.)	Increased critical stress (%)
4	0.000004	240	5.71
		180	3.85
		120	1.83
	0.0000055	240	7.95
		180	5.47
		120	2.76
	0.000007	240	10.54
		180	7.44
		120	3.98
6	0.000004	240	17.51
		180	16.96
		120	14.65
	0.0000055	240	24.41
		180	23.44
		120	20.11
	0.000007	240	31.86
		180	30.82
		120	26.15
8	0.000004	240	22.37
		180	21.1
		120	17.99
	0.0000055	240	30.68
		180	28.61
		120	24.04
	0.000007	240	39.98
		180	36.84
		120	30.42

It is clear that a lower joint spacing will reduce the thermal related stresses, but it is necessary to know the optimum saw-cut joint spacing for RCCP to prevent shrinkage cracking and to be cost-effective.

Summary and Conclusion

The principal objective of this paper was to investigate the early-age behavior of RCCP constructed under Louisiana climatic conditions and to evaluate the effect of thermal properties of RCC in the RCCP thickness design. In summary, the following observations were made from the study:

Because of lower water-to-cement ratio, reduced shrinkage, and early strength gain through roller compaction, RCCP may not be susceptible to early-age cracking.

The in situ CTE of the RCCP was successfully obtained using the field instrumentation responses. The measured in situ CTE value of RCCP was found to be lower than that of conventional concrete pavement with limestone aggregate in the mixture proportion. Thermal-induced strain responses also indicated the variation of CTE along the depth of the RCC slab because of partially and fully saturated conditions.

The daily slab curvature was computed from the static strain response from the top and bottom strain sensors. The daily slab curvature values revealed a smaller amount of slab deformation under temperature fluctuation. This could be because of the lower CTE value and less drying shrinkage characteristics of RCCP. The curvature results also revealed the hysteresis loop occurrence with an equivalent temperature gradient because of the possible moisture movement along the slab.

Based on the RCC slab curvature analysis and FE simulation results, it was further confirmed that the RCC slab exhibited a lower negative effective built-in temperature gradient at zero-stress time.

In conclusion, for RCCP design, it is necessary to consider the thermal-related stresses for damage analysis. The shrinkage strain, CTE, and built-in curling temperature of RCC are the key contributors to the fatigue cracking potentials and in determining an optimum joint spacing. An accurate estimation of CTE and appropriate joint spacing needs to be considered in the thickness design to achieve an optimum thickness for RCCP.

Footnotes

Acknowledgements

The authors would like to express thanks to all those who provided valuable help in this study.

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: M. Mahdi, Zhong Wu; data collection: M. Mahdi; analysis and interpretation of results: M. Mahdi, Y. Liu, G. Sobhani; draft manuscript preparation: M. Mahdi, Z. Wu, Y. Liu. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This paper was prepared under the research project 19-1P, “Application of Mechanistic-Empirical Pavement Design Approach into RCCP Thickness Design,” sponsored by the Louisiana Department of Transportation and Development (LADOTD) and the Federal Highway Administration (FHWA).

ORCID iD

Moinul Mahdi

References

Harrington

Abdo

Adaska

Hazaree

Guide for Roller-Compacted Concrete Pavements. National Concrete Pavement Technology Center, Institute for Transportation, Iowa State University, Ames, 2010.

Nanni

Ludwig

Shoenberger

Roller Compacted Concrete for Highway Pavements. Concrete International, Vol. 18, No. 5, 1996, pp. 33–38.

Palmer

W. D.

Paving with Roller Compacted Concrete. Public Works, January 2005, pp. 41–45.

Kim

Y. S.

Roller-Compacted Concrete Shoulder Construction on an Interstate Highway in Georgia. Transportation Research Record: Journal of the Transportation Research Board, 2007. 2040: 71–79.

U.S. Army Corps of Engineers (USACE). Roller-Compacted Concrete Pavement Design and Construction. Technical Letter ETL 1110-3-475. Department of Army, Washington, D.C., 1995.

Delatte

Simplified Design of Roller-Compacted Concrete Composite Pavement. Transportation Research Record: Journal of the Transportation Research Board, 2004. 1896: 57–65.

Zollinger

Recent Advances and Uses of Roller Compacted Concrete for Pavement Construction in United State. Proc. 13th International Symposium on Concrete Roads, Berlin, Germany, 2018.

Piggot

R. W.

Roller Compacted Concrete Pavements—A Study of Long Term Performance. RP366. Portland Cement Association, Skokie, IL, 1999.

Mallela

Abbas

Harman

Chetana

Liu

Darter

Measurement and Significance of the Coefficient of Thermal Expansion of Concrete in Rigid Pavement Design. Transportation Research Record: Journal of the Transportation Research Board, 2005. 1919: 38–46.

10.

Won

Improvements of Testing Procedures for Concrete Coefficient of Thermal Expansion. Transportation Research Record: Journal of the Transportation Research Board, 2005. 1919: 23–28.

11.

Mindess

Young

J. F.

Darwin

Concrete, 2nd ed. Prentice-Hall, Inc., Englewood Cliffs, NJ, 2003.

12.

Chhorn

Lee

S. W.

Influencing Compressive Strength of Roller-Compacted Concrete. Construction Materials, Vol. 171, No. 1, 2016, pp. 3–10.

13.

Chhorn

Hong

S. J.

Lee

S. W.

A Study on Performance of Roller-Compacted Concrete for Pavement. Construction and Building Materials, Vol. 153, 2017, pp. 535–543

14.

Federal Highway Administration. HIPERPAV Software, Validation Model Summary. Report. https://www.fhwa.dot.gov/publications/research/infrastructure/pavements/pccp/execsumm/execsumm.pdf.

15.

Mahdi

Z. Wu

Sobhani

Liu

Xie

Development of a Fatigue Damage Model for Roller Compacted Concrete Pavement Based on In Situ Saw-Cut Beams and Accelerated Pavement Performance. Transportation Research Record: Journal of the Transportation Research Board, 2022. 2676: 279–290.

16.

Rupnow

Mahdi

M. I.

Roller Compacted Concrete over Soil Cement under Accelerated Loading. LTRC Project Number: 12-7P. Louisiana Transportation Research Center, Baton Rouge, 2017.

17.

Coree

Halil

Harrington

Implementing the Mechanistic-Empirical Pavement Design Guide: Implementation Plan. Center for Transportation Research and Education, Report No. Iowa Highway Research Board; Project TR-509. Iowa State University, 2005.

18.

Vandenbossche

J. M.

Gutierrez

J. J.

Sherwood

An Evaluation of the Built-in Temperature Difference Input Parameter in the Jointed Plain Concrete Pavement Cracking Model of the Mechaninstic-Empirical Pavement Design Guide. International Journal of Pavement Engineering, Vol. 12, No. 3, 2011, pp. 215–228.

19.

Mahdi

Liu

Rupnow

Experimental Investigation of Wheel Load Induced Strain Response in Roller Compacted Concrete Pavement. Proc. 6th APT Conference, Nantes, France, 2020. https://doi.org/10.1007/978-3-030-55236-7.

20.

Wang

Roesler

J. R.

One-Dimensional Rigid Pavement Temperature Prediction Using Laplace Transformation. Journal of Transportation Engineering, Vol. 138, No. 9, 2012, pp. 1171–1177.

21.

Tan

Zhou

Pavement Temperature Estimation Model Based on Field Temperature Data. Journal of Tongji University, Vol. 41, No. 5, 2013, pp. 700–704.

22.

Alland

K. D.

Vandenbossche

J. M.

Melo de Sousa

Daily Cycles of Temperature-Independent Curvature in Jointed Plain Concrete Pavements. Journal of Transportation Engineering, Part A: Systems, Vol. 143, No. 8, 2017, p. 04017034. https://doi.org/10.1061/JTEPBS.0000042.

23.

AASHTO. Standard Test Method for the Coefficient of Thermal Expansion of Hydraulic Cement Concrete. TP60-00. American Association of State and Highway Transportation Officials, Washington, D.C., 2000.

24.

Pittman

D.W.

Development of a Design Procedure for Roller-Compacted Concrete. US Army Corps for Engineer, Washington, D.C., 1994.

25.

Pavement Designer.org (ACPA). https://www.pavementdesigner.org/projectTypes. Accessed July 28, 2021.

26.

Darter

M. I.

Barenberg

E. J.

Design of Zero-Maintenance Plain Jointed Concrete Pavement. Vol. I - Development of Design Procedures. Report No. FHWA-RD-77-111. Federal Highway Administration, Washington, D.C., 1977.

27.

Lee

Y. H.

Bair

J. H.

Lee

C. T.

Yen

S. T.

Lee

Y. M.

Modified Portland Cement Association Stress Analysis and Thickness Design Procedures. Transportation Research Record: Journal of the Transportation Research Board, 1997. 1568: 77–88.

28.

Salsilli-Murua

R. A.

Calibrated Mechanistic Design Procedure for Jointed Plain Concrete Pavements. PhD dissertation, University of Illinois at Urbana-Champaign, 1991.

29.

Barenberg

E. J.

ILLICON Computer Code. University of Illinois, Urbana, 1994.

30.

Zollinger

D.G.

Barenberg

E. J.

Proposed Mechanistic Based Design Procedures for Jointed Concrete Pavements. Report No. FHWA/IL/UI 225. University of Illinois at Urbana-Champaign, 1989.

31.

Lee

Y.H.

Darter

M. I.

Loading and Curling Stress Models for Concrete Pavement Design. Transportation Research Record: Journal of the Transportation Research Board, 1994. 1449: 101–113.