ABSTRACT

The design of a given highway project is based on the engineering parameters, such as strength or stiffness. Unfortunately, the construction specifications are not based on these properties. The acceptance criteria are typically based on adequate thickness, and adequate density of the placed and compacted materials. A procedure to measure the modulus of each pavement layer shortly after placement and after the completion of the project is presented. Stress wave or seismic technology has been used for this purpose. Simplified field and laboratory tests are suggested that can be rapidly and nondestructively performed and interpreted, so that any problems during the construction process can be adjusted. The field and laboratory methods are incorporated in a manner that their results can be readily reconciled without any scaling or simplifying assumptions. Therefore, the simplified laboratory tests can be used to develop the ranges of acceptable properties for a given material. Nondestructive field tests are performed to determine whether the contractor has achieved these levels.

INTRODUCTION

The current state of practice in quality assurance-quality control (QA/QC) of earthwork for transportation and infrastructure projects is mostly concerned with the constructability and durability of the materials. To implement performance-based specifications or warranty-based construction, the current practice has to be supplemented by methods that will provide continuity between the design, laboratory and construction. In this paper, we have attempted to demonstrate simple test methods that can be added to the existing procedures to address such concerns.

A comprehensive QA/QC procedure has to incorporate several inter-related items. First, the major parameters considered in this process should be closely correlated to the parameters used in the design. Pre-construction laboratory tests for determining the suitability of a material should also be, to a large extent, based on the selected design parameters. In addition, the acceptable values for the selected parameters should be based on the same laboratory tests. Finally, the quality control during construction has to ensure that the acceptable levels are achieved.

Procedures to measure the modulus of different flexible pavement layers in the laboratory and during construction are briefly discussed. By taking advantage of the existing specimen preparation techniques used for determining optimum maximum density, the optimum modulus is determined. An NDT device used to measure the modulus in situ is also included. Test protocols for developing a project-specific QA/QC plan are proposed.

QUALITY MANAGEMENT PROCEDURE

The goal in any earthwork project is to produce a durable material that will perform satisfactorily throughout its expected design life. Aside from environmental parameters, the primary parameters that affect the performance of a compacted layer are the modulus or strength of that layer. Based on this discussion, the goal of performance-based quality management is to ensure that the moduli or strengths throughout the project are similar to the design values, with a small variation. As such, the quality is defined as meeting a structural-related target variable with a minimal variance. In this section, the proposed laboratory and field test methods are described and different steps, from the completion of design to the completion of project, are discussed.

Laboratory Testing
The laboratory test used in this study is the free-free resonant column test. The method is comprehensively described in Nazarian et al. (2003) and is briefly discussed here. A schematic of the test set-up is shown in Figure 1. The specimen is prepared using the Proctor (ASTM D-698), modified Proctor (ASTM D-1557) or any other procedure adopted by the agency. Since the test is nondestructive, a membrane can be placed around the specimen so that the specimen can be tested later for strength (static triaxial test) or stiffness (resilient modulus or cyclic triaxial test). An accelerometer is securely placed on one end of the specimen, and the other end is impacted with a hammer instrumented with a load cell. In less than 3 minutes, a specimen can be tested, and the test result can be obtained. The process has been automated and simplified so that a technician can perform the test, interpret the results, and generate a report in almost real time.

Fig 1: Free-free Resonant Column Test.

An ultrasonic laboratory device using a v-meter (see Figure 2) is particularly useful for testing AC briquettes. In this device, a transmitting transducer is securely placed on the top face of the specimen. The transducer is connected to the built-in high-voltage electrical pulse generator of the device. The electric pulse transformed to mechanical vibration was coupled to the specimen. A receiving transducer is securely placed on the bottom face of the specimen, opposite the transmitting transducer. The receiving transducer, which senses the propagating waves, is connected to an internal clock of the device. The clock automatically displays the traveltime of compression wave. By dividing the length of the specimen by the traveltime, the compression wave velocity and as such modulus of the material is determined.

Fig 2: Ultrasonic Device.

Field Testing
The Young's modulus of a certain layer is nondestructively measured by using the high-frequency surface waves (a.k.a. ultrasonic surface waves, USW or SASW). The theoretical and experimental background behind this method can be found in Baker et al. (1995). An impact source and at least two receivers are typically adopted (see Figure 3). The surface of the medium is impacted and the transmitted waves are monitored with the receivers. By conducting a spectral analysis, the average modulus of the top layer can be obtained.

Fig 3: Portable Seismic Pavement Analyzer.

Quality Control Steps
The proposed QA/QC procedure consists of several steps. The first step is to determine the maximum modulus for a layer. In the second step, the impact of environmental parameters (such as moisture and temperature) is determined. The third step consists of determining the desirable modulus for the material. The final step is to compare the field modulus with the acceptable laboratory modulus. Each step is described below.

Step 1: Selecting Most Suitable Material
For the last century, the focus of the highway agencies has been towards developing the most durable pavement layers. For the most part, the characteristics of a durable material for a given layer depend on the collective experience of a large and diverse group of scientists and practitioners. Parameters such as angularity of the aggregates, the hardness of aggregates, percent allowable fines, allowable plasticity and degree and method of compaction impact the modulus and strength of a base layer. However, the selection of acceptable levels for these parameters is for the most part experienced based. Very little effort has been focused to routinely define the impact of these parameters on the modulus of the layer.

Similar procedures are also followed for the AC layers. The volumetric design, from the simplest from (Marshall Method) to the most sophisticated one (SHRP Method) ensures a constructible and durable material. Despite the recent recommendations by SHRP and academic circles, the desirable modulus of the material is not defined during or after mix design. Wheel tracking tests can provide an experimental way of ensuring durability but they do not provide any insight into the parameters that the designer requires for estimating the adequate thickness to ensure performance.

Based on this discussion, durability and performance go hand in hand. The material selection and mix design should be based on know-how acquired by the highway community. However, the design should be carried out based on the measured modulus of the material.

Step 2: Selecting Most Suitable Modulus
After the material is selected and its constructability is ascertained, the next step is to determine the most feasible modulus for the given material. To do so, the modulus has to be related to one of the primary construction parameters. For example, for base materials the modulus can be related to moisture content. For AC materials, the modulus can be related to the compaction effort (i.e. VTM).

To develop the moisture density curve, several specimens with different moisture contents are prepared using the same compaction energy. The same specimens can be used for determining the modulus with the free-free resonant column device. Similar to the moisture-density curve, the moisture-modulus curve is developed. In this way, the moisture content at which the maximum seismic modulus is obtained can be determined. Alternatively, the seismic modulus at the traditional optimum moisture content can be estimated.

As an example, the moisture modulus curve for a typical base is shown in Figure 4. The optimum moisture content of the material is about 6.5%. The maximum seismic modulus occurs at a moisture content closer to 6%. As such, the maximum modulus may occur at a lower moisture content than at the optimum moisture content determined by the Proctor method.

Fig 4: Variation in Modulus with Moisture under Constant Compaction Effort.

Fig 5: Variation in Modulus with Voids in Total Mix.

The variation in modulus with VTM corrected for frequency for one AC mixture is shown in Figure 5. A linear trend can be observed between the modulus and the VTM. As the VTM increases, the modulus decreases.

Step 3: Simulating Seasonal Variation in Modulus
After the compaction of a layer is completed, it may be exposed to environmental factors that could impact its behavior. One of the major input parameters in many pavement design procedures for most base and subgrade materials is the seasonal variation in modulus with exposure to moisture. In terms of performance, the water retention potentials of some materials have shown to have detrimental impact on the strength and stiffness parameters and, as such, their performance (Saarnketo and Scullion, 1997). For the AC pavement, the primary parameter is the variation in modulus with temperature. The new mechanistic design procedures, such as the AASHTO 2002 Design Guide, require this input.

To address this concern for base and subgrade materials, a specimen is prepared at the optimum moisture content and placed in an oven normally used for moisture content specimens and dried until all the moisture is removed. The specimen is weighed, and the FFRC test is performed on it daily. Since the test is nondestructive, the same specimen can be used repeatedly. When the moisture content approaches to zero, the specimen is placed on a saturated porous stone in a pan filled with water. The gain in weight of the specimen and the change in modulus with time are then monitored until the moisture content is about the optimum moisture content. By inspecting the change in modulus with moisture content, the behavior of the material can be judged. For a more comprehensive analysis, these tests can be combined with the tube suction test advocated by Saarnketo and Scullion (1997). The variation in modulus with moisture for a base material is shown Figure 6. A significant difference (about an order of magnitude) in modulus can be detected.

Fig 6: Variation in Modulus with Moisture during Drying and Soak Tests.

The drying cycle can be potentially associated with the change in the properties of the exposed soil during hot summer days after the completion of compaction. The soaking cycle can be related to the occasional rainstorms experienced by the soil. To incorporate the annual seasonal variation in modulus in the pavement design in a more systematic way, this modulus-moisture curve could potentially be used. For structural pavement management during the life of the pavement, the moisture content of the material in the field can be measured and considered in the analysis during the periodic pavement evaluation.

For the AC layer, the simplest way of relating modulus to temperature consists of preparing a specimen at the job mix formula and subjecting it to a sequence of temperatures suitable for the region being considered. At the end of each temperature sequence, the specimen is then tested. An example for the variations in modulus with temperature at a 4% and an 8% VTM for one mixture are shown in Figure 7. In both cases, the modulus decreases with an increase in temperature. A more comprehensive procedure consists of developing a master curve from the complex modulus tests and seismic tests performed on the same specimen as discussed in Nazarian et al. (2003).

Fig 7: Variation in AC Modulus with Temperature.

Step 4: Determining Desirable Modulus of Material
For the base and subgrade layers, knowing the moisture-density curve and moisture-modulus curve, one can then make a decision on the desirable modulus for the material. If the main concern is to achieve the maximum stiffness, the moisture content at which maximum modulus is achieved should be used. The downside of this selection is that the compacted material will be brittle and may be too permeable. On the other hand, if the moisture content is somewhat above the optimum, the modulus may decrease but the permeability of the material and the potential for cracking also decreases. Different base materials exhibit different levels in the reduction in modulus with increase in moisture content. For some materials, the reduction is dramatic (as much as a factor of four reduction in modulus when moisture is about 2% above optimum). For some other materials, the reduction in modulus may be as low as a factor of 1.5. In short, depending on the goals of the project, the engineer has to decide on the most desirable moisture content to use.

For the AC layer, the modulus can be based on the modulus VTM curve. After deciding on the degree of compaction at placement, the desired modulus is determined.

When the decision on the desirable modulus for the selected moisture content (for base and subgrade) and VTM (for AC layer) is made, the seismic modulus should be translated to a design modulus. This step is necessary because seismic moduli are low-strain, high-strain-rate values; whereas the design moduli are based on high-strain, low-strain-rate values.

For the base and subgrade materials, the state of stress under representative loads should be determined in order to calculate the design modulus. A summary of the approach is described in Abdallah et al. (2002). It is important to emphasize that the design modulus is dependent on the thickness of the structure and the nonlinear behavior of each layer. For example if 300 mm of the base material shown in Figure 4, is overlain with 75 mm of a typical AC, over a typical subgrade with a modulus of 70 MPa, the representative design modulus for the base will be about 170 MPa if the moisture content of 6% (moisture content at which maximum modulus is obtained as per Figure 4) is specified and will be about 155 MPa if the traditional optimum moisture content of 6.5% is used.

If the modulus assumed by the designer and the one obtained from this analysis are significantly different, either an alternative material should be used, or the layer thickness should be adjusted. In that manner, the design and material selection can be harmonized. One of the attractive attributes of this process is that the feasibility of using thicker, lower quality local materials can be explored.

For the AC layer, the most desirable way of calculating the design modulus is to develop the master curve based on the recommendations of Witczak et al. (1999). Then the design modulus can be readily determined from the design vehicular speed and the design temperature as discussed in Chapter 3.

Step 5: Field Quality Control
Field quality control is then carried out using the portable devices described above. Tests are carried out at regular intervals or at any point that the construction inspector suspects segregation, lack or excess moisture, or any other construction related anomalies. Similar to the statistical-based acceptance criteria used for moisture-density measurements, the field moduli should be greater than the representative seismic modulus (not design modulus) determined in the previous step. It is important to make a distinction between the design modulus reported to the designer and the lab seismic modulus used as a guideline for quality control. As mentioned above, these two are inter-related. However, the design modulus is also a function of the pavement structure and nonlinear properties of the base and subgrade and viscoelastic behavior of AC. As indicated before, a close relationship between the laboratory seismic modulus and field seismic modulus should be anticipated, provided the field moisture and compaction efforts are similar to those obtained in the field.

Fig 8: Typical Results from Field Measurement of Base Moduli.

Results from a base section of a road during construction are shown in Figure 8. Each data point is approximately 10 m apart. On the average, the field seismic moduli are greater than the designated lab modulus. However, at several points, the field values are lower than the designated lab values. These points that passed the density and moisture criteria during the inspection process happened to correspond to a finer than anticipated area. This area can be reworked to achieve the design modulus.

Fig 9: Typical Results from Field Measurement of AC Moduli.

The variation in modulus of the AC layer along a road, adjusted to a temperature of 25^oC and a frequency of 15 Hz, is shown in Figure 9. The modulus values were fairly constant, and within the acceptable range of moduli of 4 GPa and 5 GPa, except for an area between 27 m and 33 m. This area coincides with the entrance to a business entity where a new drainage pipe was installed. On an average, the modulus of the AC layer was about 4 GPa when all points were included and 4.3 GPa when the results from the area between 27 m and 100 33 m were ignored. Therefore, the area between the 27 m and 33 m is substandard and should be considered for some type of improvement or penalty.

VALIDATION

Once these technologies were developed, a study was performed to determine how seismic properties relate to other fundamental material properties and to assess how seismic technologies can indicate, measure and/or predict performance of pavement systems. This validation is best demonstrated and illustrated by how seismic testing was evaluated for base material.

Fig 10: Correlation between Low-strain Resilient and Seismic model.

Fig 11: Correlation between Angle of Internal Friction and Seismic Velocity.

For the base layer, the focus was on relating the performance to two traditional performance indicators, the resilient modulus of the material and the angle of internal friction. Relationships between resilient modulus and shear strength parameters and seismic modulus were established in the laboratory through extensive research. The relationship between seismic and low-strain resilient moduli can be seen in Figure 10, and between strength parameters and seismic modulus in Figure 11.

CONCLUSIONS

In this paper, the foundation for a comprehensive quality management of flexible pavement layers is laid down. Practical seismic-based laboratory and field tests are introduced that can be combined to evaluate the appropriateness of a compacted layer from modulus point of view. Since seismic moduli are low-strain moduli, an algorithm has been proposed to relate them to design moduli that are used by the pavement engineer for highway projects.

The Texas Department of Transportation is evaluating the system for implementation on a limited basis. Work still has to be done to develop the acceptance criteria and validate the procedures. However, our experience thus far clearly indicates that the methodology can be of value in developing performance-based specifications for highway construction.

ACKNOWLEDGMENT

This work was supported by the Texas Department of Transportation. The authors would like to express their sincere appreciation to Steve Smith, Mike Arellano and Mark McDaniel for their ever-present support and valuable advice.

REFERENCES

1. Abdallah, I., and Nazarian, S. (2002), "Program SMART: Determination of Nonlinear Parameters of Flexible Pavement Layers from Nondestructive Testing," Research Report 1780-4, Center for Highway Materials Research, The University of Texas at El Paso, TX.
2. Baker, M. R., Crain, K., and Nazarian, S. (1995), "Determination of Pavement Thickness with a New Ultrasonic Device," Research Report 1966-1, Center for Highway Materials Research, The University of Texas at El Paso, El Paso, TX, 53 p.
3. Nazarian, S., Yuan, D., Tandon, V. and Arellano, M. (2003), "Quality Management of Flexible Pavement Layers with Seismic Methods," Research Report 1735-3, Center for Highway Materials Research, The University of Texas at El Paso, TX.
4. Saarenketo, T and Scullion, T (1997) "Using Suction and Dielectric Measurements as Performance Indicators for Aggregate Base Materials," Transportation Research Record 1577, pp 37-44.
5. Witczak, M.W., Bonaquist, R., Von Quintus, H., and Kaloush, K. (1999) "Specimen Geometry and Aggregate Size Effects in Uniaxial Compression and Constant Height Shear Tests," Journal of Association of Asphalt Paving Technologist, Volume 69, pp 733-793.