Wettability, Saturation, and Viscosity Using the Magnetic Resonance Fluid Characterization Method and New Diffusion-Editing Pulse Sequences

This paper discusses a new nuclear magnetic resonance (NMR) method that can provide wettability, saturation, and oil viscosity values in rocks partially saturated with oil and brine. The method takes advantage of two new technological advances in NMR well logging — the MRF Magnetic Resonance Fluid Characterization Method and NMR “diffusion-editing” (DE) pulse sequences. We discuss the principles underlying the fluid characterization method and the pulse sequences. The fluid characterization method is used to provide robust inversions of DE data suites acquired on fully brine-saturated and partially saturated rock samples. The outputs of the inversion are separate diffusion-free brine and oil T2 distributions for the fluids measured in the rocks. NMR measurements on partially saturated rocks are sensitive to wettability because of surface relaxation of the wetting phase fluid. The surface relaxation rate, however, must be significant compared to the bulk relaxation rate in order for wettability to noticeably affect the NMR response. We present results showing that the surface relaxation rate at lower wetting phase saturations is enhanced compared to that measured at higher saturations. The consequence of wettingphase saturation on NMR-based wettability determination is discussed. Wettability affects the relaxation rates of both the wetting and nonwetting phases in partially saturated rocks. Surface relaxation of the wetting phase in a rock results in shorter relaxation times than would otherwise be observed for the bulk fluid. The nonwetting phase fluid molecules do not * Mark of Schlumberger come into contact with the pore surfaces and therefore their relaxation rate in the rock is the same as in the bulk fluid. We present accurate and robust computations of diffusionfree T2 relaxation time distributions for both the wetting and nonwetting phases in four rocks that include two sandstones and two dolomites. A DE data suite was acquired on each rock measured in two different partial saturation states and also fully brine saturated. Wettability is determined by comparing the oil and brine T2 relaxation time distributions measured in the partially saturated rocks with the bulk oil T2 distribution and with the T2 distribution of the fully brine-saturated sample. The brine and oil T2 distributions are used to compute saturation and oil viscosity values. A general discussion elucidating the sensitivity range and T2 limits of diffusion-based NMR methods is given in the appendix. The appendix also derives and displays the gain in signal-to-noise ratio that is achieved by using DE data sequences for fluid characterization in place of Carr-PurcellMeiboom-Gill (CPMG) data suites. Introduction This paper discusses a new nuclear magnetic resonance (NMR) method for determining wettability, saturation, and viscosity values in partially saturated reservoir rocks. It has potential applications to wettability interpretation in nativestate cores measured in the laboratory as well as to measurements made downhole by an NMR logging tool. Previous methods for determining wettability of partially saturated rocks, including NMR methods, are limited to laboratory measurements. NMR wettability determination of partially saturated rocks is based on comparing either T1 or diffusion-free T2 distributions of oils measured in rocks with the distributions measured on the bulk oils (i.e., outside the rocks). Previous NMR methods of measuring rocks partially saturated with water and oil are only capable of measuring the composite T2 distribution of both the water and oil phases in the rock. The oil distribution is sometimes measured in restored-state cores by replacing the water phase by D2O (heavy water), which does not have an NMR signal at the proton Larmor frequency. The latter approach works well but is not useful for studying SPE 77397 Wettability, Saturation, and Viscosity Using the Magnetic Resonance Fluid Characterization Method and New Diffusion-Editing Pulse Sequences R. Freedman, SPE, and N. Heaton, SPE, Schlumberger; M. Flaum, SPE, and G.J. Hirasaki, SPE, Rice University; C. Flaum, SPE, and M. Hürlimann, Schlumberger-Doll Research Center 2 R. FREEDMAN, N. HEATON, M. FLAUM, G.J. HIRASAKI, C. FLAUM, AND M. HÜRLIMANN SPE 77397 native-state cores in the laboratory or for downhole NMR measurements. We take advantage of recent innovations in NMR welllogging technology that provide the capability to measure robust and accurate diffusion-free brine and oil T2 distributions in partially saturated rocks. These innovations are the MRF characterization method and DE pulse sequences. The innovations are discussed in detail in the following sections. The experiments reported on in this paper were conducted at the Schlumberger-Doll Research (SDR) Center. The NMR data were acquired in the fringefield of a superconducting magnet at a proton Larmor frequency of 1.764 MHz with a magnetic field gradient of 13.2 G/cm. The experiments included measurements on four rocks — Bentheim and Berea sandstones and two dolomite samples from the Yates oil field in west Texas. The samples were partially saturated with a 33°API gravity North Sea stock tank oil. The samples were measured first fully brine saturated and then at two partial saturations. The first partial saturation state was at very high oil saturation achieved by drainage of the brine phase close to residual water saturation. The second partial saturation state was at a lower oil saturation achieved by spontaneous imbibition of water for the sandstones and forced imbibition for the dolomites. Wettability. Wettability is the tendency of a fluid to spread on and preferentially adhere to or “wet” a solid surface in the presence of other immiscible fluids. Knowledge of reservoir wettability is critical because it influences important reservoir properties including residual oil saturation, relative permeability, and capillary pressure. An understanding of the wettability of a reservoir is crucial for determining the most efficient means of oil recovery. This is becoming increasingly important as more secondary and tertiary recovery projects are being undertaken to recover remaining reserves after primary production. It is generally believed that most reservoirs are water wet or mixed wet. The concept of mixed wettability was first introduced by Salathiel. In mixed-wet rocks the brine phase occupies the smaller pores, which are therefore water wet. In the larger oiland brine-filled pores the oil wets part of the pore surfaces. Two widely used laboratory indicators of wettability are contact angles measured in water-oil-solid systems and the Amott wettability index. The definition of contact angles and their relationship to wettability is shown in Fig. 1. Contact angles less than 90, measured relative to the water phase, are indicative of a preferentially water-wet surface, whereas angles greater than 90 indicate a preferentially oil-wet surface. A practical limitation of contact angle measurements is that they are restricted to special geometries and cannot be made on reservoir rocks. Fig. 1: The definition of wettability of a brine-oil-solid surface according to the contact angle. If θc < 90 , water exhibits an affinity for the surface that is said to be preferentially water wet. If θc = 0, the surface is strongly water wet. If θc > 90 , water exhibits an aversion for the surface that is said to be preferentially oil wet. If θc = 180 o , the surface is strongly oil wet. The Amott wettability index is determined from the amount of oil displaced from a core, starting at some initial oil saturation, by spontaneous imbibition of brine divided by the amount of oil displaced by both spontaneous and forced imbibition. Amott defined an analogous index by also considering the displacement of water by oil. The Amott indices vary linearly on a scale from 0 to 1. The endpoints for the displacement of oil by water are 0 for a neutral to oil-wet system and 1 for a strongly water-wet system. Imbibition measurements like the Amott index provide the most quantitative indicators of wettability, however, they are limited to the laboratory. NMR measurements on fluid-saturated porous media are sensitive to wettability because of the enhanced relaxation rate caused when fluid molecules come into contact with pore surfaces that contain paramagnetic ions or magnetic impurities. Surface relaxation of nuclear magnetism is usually the dominant relaxation mechanism for the wetting phase in a partially saturated rock. The nonwetting phase is unaffected by surface relaxation because the pore surface is coated by the wetting fluid. The other relaxation mechanisms, bulk and diffusion relaxation, affect both the wetting and nonwetting phases. The relaxation rate of the transverse magnetization measured in a spin-echo experiment is the sum of the relaxation rates from all three mechanisms. The bulk relaxation rates for liquids are proportional to their viscosities. The surface relaxation rate of the wetting-phase in a single pore can be written in the form,