Use of Ground-Penetrating Radar to Detect Cement Content in Cement-Stabilized Subgrade †

: Cement stabilization has been successfully used to improve poor-quality subgrade soils by increasing the soil support to remedy these soils useful for pavement construction. Cement stabilization has the potential to reduce initial construction costs through improved subgrade stability in the pavement structure. Cement stabilization also provides greater long-term stability of the pavement structure and lower pavement life-cycle costs through reduced pavement maintenance. Unfortunately, ﬂexible pavements over cement-stabilized subgrade are experiencing reﬂective cracking originating from the shrinkage cracks on top of cement-stabilized subgrade due to poor construction. In this study, ground-penetrating radar (GPR) was used to capture the inconsistent layer thickness of cement-stabilized subgrade and its cement content. The results show that GPR is capable of capturing different dielectric constants along with different percent cement contents in subgrade soils.


Introduction
The primary function of pavement structures is to protect the subgrade by reducing stresses and strains due to the loading to a tolerable level. Pavement structures are designed to decrease the stresses as they propagate through the layers above the subgrade. In the case of flexible pavements where close access to quarries is not feasible, the subgrade is stabilized to support and distribute the stresses.
Flexible pavements over stabilized subgrades are highly dependent on the subgrade elastic modulus. The compressive strains in the pavement structure greatly decrease as the subgrade modulus increases. When the compressive strains in the pavement become too high, pavement failures begin to occur. Likewise, the subgrade soil will experience deformation failure due to the high compressive forces. A weak soil underlaying the pavement structure can lead to accelerated pavement deterioration. Cement stabilization can be used to improve subgrade stability, leading to reduced initial construction costs and reduced repair costs [1]. GPR has been used to detect the thickness and cement content of the cement-stabilized layer [2]. GPR is widely used to evaluate the pavement structure using electromagnetic (EM) waves which find the dielectric constant, the primary material property obtained from GPR surveys. For homogenous layers, the speed of light is proportional to the EM waves passing through the layer. The relative permittivity, or dielectric constant, of a homogeneous medium connects the EM velocity in a substance to the speed of light in empty space [3].

Research Approach
The dielectric constant can be used to characterize the microstructure of cement-based materials and, thus, was used as the method of strength prediction [4]. The percentage of capillary pores in a soil sample was directly proportional to the compressive strength of the soil. This relationship can be detected with the GPR due to the higher percentage of cement-stabilized soil samples possessing a greater water content, which led to an increased dielectric constant. In this study, mathematical models were developed to detect the different percentages of cement in cement-stabilized subgrade soils through the use of the relative dielectric constant.

Ground-Penetrating Radar (GPR) System
The GSSI 2 GHz air-coupled antenna was used for this study due to the wide range of possible applications in the field. The air-coupled antenna was ideal for pavement scanning, rather than the GSSI 400 MHz ground-coupled antenna, which must remain very close to the ground, as seen in Figure 1. The GSSI 2 GHz air-coupled antenna can be driven safely at normal road speeds without the need to worry about unlevel surfaces in the road. This model also more clearly captures the first layers of the pavement, as opposed to the GSSI 400 MHz ground-coupled antenna, due to the higher frequency not penetrating the road layers as deeply [5].
where = dielectric constant; = speed of light in free space of 3 × 10 m/s; = EM velocity in the material.

Research Approach
The dielectric constant can be used to characterize the microstructure of cementbased materials and, thus, was used as the method of strength prediction [4]. The percentage of capillary pores in a soil sample was directly proportional to the compressive strength of the soil. This relationship can be detected with the GPR due to the higher percentage of cement-stabilized soil samples possessing a greater water content, which led to an increased dielectric constant. In this study, mathematical models were developed to detect the different percentages of cement in cement-stabilized subgrade soils through the use of the relative dielectric constant.

Ground-Penetrating Radar (GPR) System
The GSSI 2 GHz air-coupled antenna was used for this study due to the wide range of possible applications in the field. The air-coupled antenna was ideal for pavement scanning, rather than the GSSI 400 MHz ground-coupled antenna, which must remain very close to the ground, as seen in Figure 1. The GSSI 2 GHz air-coupled antenna can be driven safely at normal road speeds without the need to worry about unlevel surfaces in the road. This model also more clearly captures the first layers of the pavement, as opposed to the GSSI 400 MHz ground-coupled antenna, due to the higher frequency not penetrating the road layers as deeply [5].

Sample Preparation
The southern region of the state of Georgia does not have close access to quarries and, thus, relies on the use of cement stabilization for roadway construction. The GDOT uses plant-mixed cement and soil to provide an accurate mixture for cement-stabilized subgrades. Previous mix data show that the most common range of percentage of cement

Sample Preparation
The southern region of the state of Georgia does not have close access to quarries and, thus, relies on the use of cement stabilization for roadway construction. The GDOT uses plant-mixed cement and soil to provide an accurate mixture for cement-stabilized subgrades. Previous mix data show that the most common range of percentage of cement used is 5% to 7% [6]. These data are based on the road loading design requirements and are closely monitored by the mixing plant to meet the needs of the traffic.
Soil cement test specimens were prepared in a standard proctor mold and compacted to 100% of the maximum dry density to simulate site conditions in accordance with the American Society for Testing and Materials (ASTM) D1633-17 [7]. Varying soil mixes were provided by the GDOT to compare the data of different soil types. Three samples of 6% cement-stabilized soil were tested at different time intervals. The dielectric constant is known to decrease rapidly within the first 7 days due to the chemical properties of cement hydration [4]. Therefore, samples were scanned daily for the first week to monitor this chemical process.

Laboratory Test Results
A series of laboratory tests was conducted on soil samples of sand using the GPR setup. For the test, two 5 gallon buckets of sand were assembled and one was stabilized with 6% Type I Portland cement. Type I Portland cement was selected due to it being the type of cement used by the GDOT and 6% cement was used to provide a middle range of the cement percentages to be tested. The other bucket was not stabilized to provide a dielectric constant reading of the plain soil. The two buckets were then scanned at days 0, 7, 14, 21, and 28 using the GPR.
The preliminary soil cement test results show that the dielectric constant of the cementstabilized soil decreased with time. This trend can be seen in Table 1. The decrease in the dielectric constant with time was due to the cement hydration process [4]. Water has a dielectric constant value of 81, which led to the large change in dielectric constant. It is important to note that the results shown in Table 1 were calculated using the amplitudes from the GPR scan. The two buckets were also subject to changing temperatures. The same sand soil sample was also stabilized with 6% cement and prepared according to the sample preparation standards of ASTM D1633-17. The proctor mold-shaped cylinder was then scanned with the GPR every day for 7 days to detect the decrease in the dielectric constant. The soil sample was allowed to cure in a controlled environment at a constant temperature and humidity. The results of the scan are shown in Figure 2. The dielectric constant decreased the most within the first 2 days of the sample being constructed and then maintained a constant value.
used is 5% to 7% [6]. These data are based on the road loading design requirements and are closely monitored by the mixing plant to meet the needs of the traffic.
Soil cement test specimens were prepared in a standard proctor mold and compacted to 100% of the maximum dry density to simulate site conditions in accordance with the American Society for Testing and Materials (ASTM) D1633-17 [7]. Varying soil mixes were provided by the GDOT to compare the data of different soil types. Three samples of 6% cement-stabilized soil were tested at different time intervals. The dielectric constant is known to decrease rapidly within the first 7 days due to the chemical properties of cement hydration [4]. Therefore, samples were scanned daily for the first week to monitor this chemical process.

Laboratory Test Results
A series of laboratory tests was conducted on soil samples of sand using the GPR setup. For the test, two 5 gallon buckets of sand were assembled and one was stabilized with 6% Type I Portland cement. Type I Portland cement was selected due to it being the type of cement used by the GDOT and 6% cement was used to provide a middle range of the cement percentages to be tested. The other bucket was not stabilized to provide a dielectric constant reading of the plain soil. The two buckets were then scanned at days 0, 7, 14, 21, and 28 using the GPR.
The preliminary soil cement test results show that the dielectric constant of the cement-stabilized soil decreased with time. This trend can be seen in Table 1. The decrease in the dielectric constant with time was due to the cement hydration process [4]. Water has a dielectric constant value of 81, which led to the large change in dielectric constant. It is important to note that the results shown in Table 1 were calculated using the amplitudes from the GPR scan. The two buckets were also subject to changing temperatures. The same sand soil sample was also stabilized with 6% cement and prepared according to the sample preparation standards of ASTM D1633-17. The proctor mold-shaped cylinder was then scanned with the GPR every day for 7 days to detect the decrease in the dielectric constant. The soil sample was allowed to cure in a controlled environment at a constant temperature and humidity. The results of the scan are shown in Figure 2. The dielectric constant decreased the most within the first 2 days of the sample being constructed and then maintained a constant value.

Conclusions and Discussion
This research aims to identify a GPR method to detect the causes of failure in flexible pavements over cement-stabilized subgrades. This method of failure originates from the shrinkage cracks on top of cement-stabilized subgrade due to poor construction.
The application of this research is to provide a nondestructive testing method capable of predicting possible areas of roadways failures and to provide a method of asset management and quality control. This can be accomplished through the detection of cement content which is directly proportional to the compressive strength of the subgrade soil. The thickness of cement-stabilized subgrade also provides a method to ensure design field validation. Cement stabilization has been proven to provide greater long-term stability of the pavement structure and lower pavement life-cycle costs through reduced pavement maintenance.