VERIFICATION OF NMR MODELED PERMEABILITY USING CONVENTIONAL ESTIMATES IN TARANAKI BASIN

Abstract

Permeability measurements are of paramount importance in understanding reservoir flow behaviors and ultimate production. For different rock types and formation, various models have been developed to predict permeability of the reservoirs. As most such models are empirically derived, they are good for site specific studies. The limitations in terms of rock type, rock particle size, boundary conditions, etc. make them redundant in unexplored areas. The best among the available method to measure permeability is using the cores. In the absence of core data, permeability can be best estimated using NMR logs (with calibration using core data). Not only that NMR is less expensive than obtaining core data and doing the measurement, NMR provides information on a number of related topics for characterizing lithology fluid and pore size characterization. In the absence of core data and NMR data, ability to predict permeability using the conventional well logs(Gamma ray, Mudlog, Gamma-Gamma, Resistivity) is of great value in any field once calibrated in a well having NMR data. In this thesis, I explored the convergence of permeability values obtained using empirical relations and other physical models with that of NMR derived permeability values. The permeability using NMR was obtained using Timur-Coates and Schlumberger Doll Research Model. The Timur-Coates model depends on porosity and T2cutoff whereas Schlumberger Doll Research model depends on porosity and T2gm (geometric mean). The empirical relations (Carmen-Kozney, k-Vclay) obtained from gamma ray log and Mudlog) were applied to estimate permeability in terms of Vclay, porosity, specific surface area and Kozney constant. The permeability estimated using Carman-Kozney and k-Vclay turns out to be almost same as the NMR driven permeability. A similar practice in the field will provide easier methods of estimating permeability at low cost operations.