In situ water content measurement of compacted highway subgrades using time domain reflectometry

The water content is a principle factor of subgrade soil stability in shallow foundation pavement systems. Currently data do not exist for seasonal subgrade water content changes in Tennessee pavement systems. Four test sites across the state were instrumented with equipment to monitor subgrade and flexible asphalt pavement water content and temperature, water infiltrated through the pavement layers, and climatic conditions. These data will later be used to develop a rational method of flexible asphalt pavement design that accounts for environmental factors and seasonal water content changes.

Five-segment TDR probes, manufactured by Environmental Sensors, were installed in the soil subgrade, the stone base, and the asphalt stabilized base. Single segment TDR probes, constructed in the laboratory, were installed in the asphaltic concrete. The multiple-segment TDR probes allow changes in water content to be tracked two-dimensionally, between the probes and along probe segments. Temperature sensors were installed at depths corresponding to the soil subgrade and the stone base TDR probes. Three temperature sensors were also installed in the top 20 cm of the pavement. A weather station was installed at each test site to monitor air temperature, relative humidity, solar radiation, wind speed, and rainfall.

The Moisture Point TDR system, manufactured by Environmental Sensors, was designed for use in agricultural soils. The materials in which the TDR probes were installed at each test site are quite dissimilar from agricultural soils. For this reason, a calibration study was performed with the Moisture Point TDR equipment using test site subgrade soils and a single crusher run gravel sample. Ten previously published TDR calibration equations were evaluated to determine which equation most accurately predicts water content for the subgrade soils and for crusher run gravel.

The relationship between inverse signal velocity and water content proposed by Herkelrath et al. (1991) most accurately predicted water content for all subgrade soils; however, this relationship requires the derivation of a soil specific slope and intercept. The three- and four-phase dielectric mixing models, proposed by Dobson et al (1985) and rewritten to calculate volumetric water content by Weitz et al. (1997), provided results within the accuracy of the Moisture Point equipment for subgrade soils. The equation proposed by Baran (1994) most accurately predicted water contents for the crusher run gravel.