Dynamic monitoring of dielectric properties during two phase immiscible displacements in sand packs using frequency domain electromagnetic sweeps

Electromagnetic heating is a promising soil remediation method especially in thin formations. The lack of a wide-spread adoption of this method stems from insufficient knowledge of how the complex dielectric properties, that govern propagation of the electromagnetic waves through porous media, change with changing frequency, water saturation, displacement types and flow regimes. To breach these gaps several sets of spontaneous deionized (DI) water imbibition experiments, followed by the primary drainage floods, that were followed by the secondary DI water imbibition floods in confined uniform sand packs were performed. The frequency domain relative dielectric constant and conductivities were extracted from the two-port complex S-parameter measurements taken with the vector network analyzer during these immiscible displacements at various water saturation levels at ambient conditions. A novel coaxial transmission line core holder was designed and commissioned, and a modified version of a plane-invariant dielectric extraction algorithm was developed for this purpose. Series, parallel and semi-disperse mixing models were applied to fit the water saturation dependent relative dielectric constant and conductivity values sampled at 500 MHz from the extracted frequency domain spectra. The Maxwell-Garnett parallel model was proved to be the most flexible because it could capture the sampled conductivity values in all secondary imbibition floods before and after the breakthroughs, where the inflection points were observed. These inflection points were attributed to silica production and a potential shear-stripping flow. This observation was further confirmed by conducting a single-phase Darcy's law analysis of two DI water imbibition floods. [Display omitted] •Dielectric properties were extracted for the DI water and oil floods of sand packs.•Series, parallel and semi-disperse mixing models were fitted to the dielectric data.•In-situ emulsification and silica production were detected in imbibition floods.•1D pressure distributions were generated for two secondary imbibition floods.•Dynamic capillary pressures were generated for two secondary imbibition floods.