Soil permittivity determination using a fringe field capacitor

In view of the fact that dry soil has a permittivity value of about 3 and water about 80, soil water content may be estimated from soil permittivity measurements. This study was undertaken to design, construct and test a sensor and instrument to determine soil permittivity. The research was partitioned into three sections: (1) design of the fringe field dielectric sensor, (2) development of instrumentation and equations required to calculate permittivity parameters and (3) testing of the system in soil. A double-cylinder configuration was selected and results from solving Laplace\u27s equation using finite difference methods identified a long, thin sensor that had a length to radius ratio of at least five as optimum. The simulated electric field extended laterally 4.8 times the radius and altitudinally 2.6 times the radius. Low frequency (1 to 20 KHz) measurements of the charging-discharging current through an epoxy coated sensor and the phase angle between voltage and current were recorded to estimate permittivity and the loss tangent. Permittivity of water from current and phase angle measurements were higher than from a commercial impedance bridge. Both estimates disagreed with values reported in the literature which suggests the assumed equivalent circuit model for the sensor is incorrect. Relative propagated error in permittivity ranged from 0.9 to 4% depending on frequency. This sensor along with the described instrumentation was then used to determine the influence of KCl concentration, soil water content and bulk density on permittivity and loss tangent. Samples of a sandy clay loam soil at different water contents and bulk densities were prepared. Measurements were collected at frequencies of 1 to 20 KHz. Permittivity and loss tangent were insensitive to changes in KCl concentration. Soil volumetric water content was statistically related to real permittivity with highest sensitivity at 1 KHz. Bulk density was related to the loss tangent with highest sensitivity at 7 KHz. Although variability was high, the results indicate that the sensor and instrumentation system merit further research