In June 1994, a 150-m-long secondary road pavement section was built as part of the realignment of route 616 and 757 in Bedford County, Virginia to evaluate the performance of geosynthetically stabilized flexible pavements. The California Bearing Ratio (CBR) of the subgrade after construction was approximately 8%. The pavement section is was divided into nine individual sections, each approximately 15 m long. Sections one through three have a 100-mm-thick limestone base course (VDOT 21-B), sections four through six have a 150-mm-thick base course, and sections seven through nine have a 200-mm-thick base course. Three sections were stabilized with geotextiles and three with geogrids at the base course-subgrade interface. The remaining three sections were kept as control sections. One of each stabilization category was included in each base course thickness group. The hot-mix asphalt (HMA), SM-2A, wearing surface thickness was 78-90 mm. The outside wheel path of the inner lane was instrumented with strain gages, pressure cells, piezoelectric sensors, thermocouples, and moisture sensors. Section performances based on the instrumentation response to control and normal vehicular loading indicated that geosynthetic stabilization provided significant improvement in pavement performance. Generally, the measured pressure at the base course-subgrade interface for the geotextile-stabilized sections was lower than the geogrid-stabilized and control sections, within a specific base course thickness group. This finding agreed with other measurements, such as rut depth, ground penetration radar survey, and falling weight deflectometer survey. The control section (100-mm-thick base course) exhibited rutting that was more severe than the geosynthetically stabilized sections. Falling weight deflectometer back-calculation revealed consistently weaker subgrade strength for the geogrid-stabilized and control sections than for the geotextile-stabilized sections over the three year evaluation period. To quantitatively assess the extent of contamination, excavation of the first three sections in October 1997 revealed that fines present in the base course were significantly greater in the control and geogrid-stabilized section than in the geotextile-stabilized section. These findings led to the conclusion that the subgrade fine movement into the base layer when a separator is absent jeopardizes its strength. Further analysis of the field data showed that geotextile-stabilization may increase the service life of flexible secondary road pavements by 1.5 to 2 times.

Finally, a new mechanistic-empirical flexible pavement design method for pavements with and without geosynthetics has been developed. Elasto-viscoelastic material characterization is used to characterize the HMA layer. The field results from Bedford County, Virginia project have been used to calibrate and validate the final developed design procedure. The concept of transition layer formed at the interface of base course and subgrade is also incorporated into the design approach. Powerful axisymmetric linear elastic analysis is used to solve the system of equations for mechanical and thermal loading on the pavement structure. Elasto-viscoelastic correspondence principle (EVCP) and Boltzman superposition integral (BSI) are used to convert the elastic solution to its viscoelastic counterpart and also to introduce the dynamic nature of vehicular loading. Pseudo-elastoplasticity is introduced into the problem by determining the extent of plastic strain using laboratory experimentation results and estimating the failure mechanisms, based on accumulated strains as opposed to the total strain (recoverable and non-recoverable). The pavement design approach presented in this dissertation is a hybrid of already existing techniques, as well as new techniques developed to address the visco-plastic nature of HMA.