An important element of dielectric logging is validating the mixing models that can predict a formation's electromagnetic response over a wide frequency range. The main petrophysical outputs from dielectric log interpretation include water saturation and water resistivity. The resistivity of water as an output of the interpretation provides an advantage for formations where this information is not available or varies over short depths. The electromagnetic properties of a formation show a variation with respect to frequency that is called dispersion. To relate the dispersive behavior of a formation to petrophysical properties, certain mixing models are used that relate the properties of the aforementioned elements of the rock to that of their collective effective behavior. Interpretation success depends on the application of a suitable mixing model. This work provides a comparative study of different mixing models and their validation with respect to experimental results on sand and carbonate rocks. Results of experiment on dielectric dispersion of rock cores are used to study the performance of commonly used mixing models. Experimental results for various values of water saturation and water resistivity at different frequency ranges are used in pursuing two goals. First, the ability of the mixing models to recreate the rocks' dielectric dispersive behavior is considered. It is shown that selecting an appropriate mixing model requires experimental data on a wide frequency range. Second, the fitted mixing models are used to invert for petrophysical parameters. These parameters are compared with known values from experiment to quantify the error associated with using each mixing model. Dielectric Logging service is poised to play an important role in Formation Evaluation of unconventional and/or non-Archie environments, only if appropriate mixing models are developed and applied. Paper provides guidelines on choosing the appropriate mixing model to improve accuracy of dielectric data processing.
