Formation water content is one of the key petrophysical quantities provided by dielectric logging. However, to determine water content from formation permittivity measurements, the rock matrix permittivity must be known. Uncertainty in the rock matrix permittivity values translates into uncertainty in the water-content estimate, which is especially important in low-porosity formations or complex lithologies. Matrix permittivity values are not well-known for a number of minerals and can also vary for the same type of mineral in different formations. Thus, a laboratory methodology for the accurate determination of matrix permittivity at dielectric logging frequencies is required to facilitate accurate log interpretation.
One can measure matrix permittivity values on solid plugs (Seleznev et al. 2011). However, the plug-based methodology can be challenging in very-low-permeability or unconventional reservoirs because of difficulties with plug drying. In addition, it is not readily applicable to unconsolidated formations. Finally, it may be impossible to cut solid plugs because of limited availability of rock material. Matrix permittivity measurements made on rock powders are capable of addressing all these issues.
We introduce a methodology for laboratory measurements of matrix permittivity on rock powders at 1 GHz. The methodology is based on conducting dielectric measurements on mixtures of rock powders and liquids with variable permittivities in a dielectric resonator. The permittivity of the rock matrix is inverted from a series of measurements obtained on pure liquids and powder/liquid mixtures. The methodology was benchmarked on a collection of samples representing common oilfield lithologies with matrix-permittivity values between 4.6 and 8.6. The reference matrix-permittivity values were first measured on solid plugs. Then, the plugs were crushed into powders, and the matrix permittivity values were determined on powders following the proposed methodology. The values obtained on powders matched the ones measured on solid plugs within 0.2 dielectric units, resulting in accuracies better than 1% for the water-filled porosity and better than 1,000 ppm for water salinity.
This new methodology was applied to a number of core samples from a carbonate reservoir offshore Sarawak, where dielectric logging was performed along with conventional core analysis. The resulting measured matrix permittivity values were then used to interpret the dielectric log measurement. Results showed a better estimation of water-filled porosity and of the textural MN parameter, equivalent to the Archie's cementation exponent in a water-bearing zone, than would have resulted from using “chartbook” values of matrix permittivity. A consistent and optimized interpretation was obtained in porosities ranging from 5% to more than 30%.