Pore-Network Connectivity and Permeability Assessment by Use of Nuclear Magnetic Resonance Log-Inject-Log Method

Abstract

Summary Nuclear-magnetic-resonance (NMR) logging is a well-established technology for estimating porosity, pore-size distribution, and permeability in conventional reservoirs. However, the uncertainty associated with these estimates can be significant in complex heterogeneous-carbonate reservoirs, such as those with variable pore size and pore-network connectivity. In such cases, distinguishing isolated-pore space is often impossible by use of well-log measurements and conventional well-log-interpretation methods, which makes permeability evaluation unreliable. This paper proposes a new application of the NMR log-inject-log method to improve assessment of permeability and to distinguish isolated pores from connected pores. We propose injecting manganese-bearing solution in the rock samples and simultaneously analyze NMR measurements before and after injection of the contrast solution. Injection of contrast agents enables eliminating the effect of isolated-pore space from the NMR T2 (spin-spin relaxation time) distribution, which is then used to improve permeability assessment. To confirm the feasibility of the proposed method for field applications, we conduct NMR laboratory measurements in two carbonate-rock types. We inject manganese-bearing solution into rock samples by use of a coreflood experimental procedure and measure NMR T2 distributions before and after injection. We then estimate isolated porosity and interconnected porosity by taking into account the difference between the NMR T2 distributions acquired before and after the injection of the contrast solution. We introduce a method to calculate the geometric mean of the T2 distribution for interconnected pores, and also to obtain effective free-fluid volume and bound-fluid volume. Finally, we use the T2 distribution corrected for the effect of isolated pores in conventional NMR-based permeability models to improve permeability assessment. We cross validate the NMR-based permeability estimates against Klinkenberg permeability measured by an unsteady-state gas permeameter. The results confirm that the proposed method enables quantifying the isolated and connected pore volume (PV), and finally improves NMR-based permeability assessment. The new method provides estimates of permeability with up to 10% average error in eight carbonate-rock samples, which was a significant improvement compared with the average errors of up to 500% when T2 distribution is not corrected for the effect of isolated pores.