Fit-For-Purpose Model for Dielectric Data Interpretation for Enhanced Formation Evaluation

Abstract

AbstractGreat progress in technology development of dielectric logging has been made in the last decade, mainly by establishing the capability of multi-frequency measurements for more robust determination of near wellbore water-filled porosity and flush zone resistivity, in addition in providing the potential of characterizing reservoir petrophysical properties such as rock tortuosity. To interpret acquired dielectric data, several models have been developed and used in the industry. The objective of this laboratory study was to evaluate the commonly used models, identify conditions under which each model is applicable, and establish a workflow to select the most suitable model for targeted applications.This study was performed on limestone core samples. Routine and special core analysis including resistivity and dielectric measurements were conducted. Three commonly used dielectric models (CRIM, Bimodal, and SMD) were evaluated. Properties having major influence on these models are brine salinity, brine saturation, and Archie parameters m and n. Effects of these properties on the dielectric data interpretation models were studied systematically, by comparing the models’ outputs to direct laboratory measurements. For core samples fully saturated with fresh water of salinity ranging from 2 to 23 kppm NaCl, all three models provide consistent results of inverted water-filled porosity when compared with core weight porosity. The inversion results are robust and independent of whether the salinity is constrained or not. In a saline environment of 200 kppm NaCl, the inverted water-filled porosities are found to be slightly higher than core porosity. Comparing between inverted dielectric textural parameter MN and Archie resistivity parameter m shows that the Bimodal model performs better than the SMD model at this high-water salinity environment.Based on these results, a simple linear regression correction scheme is proposed for each salinity range investigated. The proposed algorithm is structured to correct the output of each model using statistical mean at specific salinity conditions. Finally, we introduce a fit-for-purpose model selection for dielectric data interpretation for enhanced formation evaluation.