Fluid Typing Using Magnetic Resonance Data with High Diffusivity

Abstract

Abstract Magnetic resonance (MR) data were acquired in a six-frequency PoroPerm + Light Oil mode in the study well. The acquired data had a low signal-to-noise ratio from midpoint to the top of the logged interval. This ratio could adversely affect the suitability of this data for hydrocarbon fluid typing and saturation computation. Examination of the detailed quality control plot showed high B-pulse ringing on the two lowest and the third highest frequencies, A-pulse ringing on the two lowest frequencies, and noise on the two lowest frequencies. The affected frequencies could not be used for fluid typing analysis. The reduction in the total number of available frequencies resulted in a high diffusivity, especially in the gas-bearing reservoir sand. High diffusivity caused some of the movable fluids to appear as irreducible bound water. The low hydrogen index of the gas also caused a low-permeability profile. Despite the failure of some of the acquired frequencies, there was a need to identify and quantify the fluid type in the reservoir. Several fluid-typing techniques where employed to find a suitable technique for this dataset. After trying Multiple Gradient Inter-Echo Time (MGTE) analysis, the Simultaneous Inversion of Multiple Echo Trains (SIMET) module, and a 2D NMR (D-T2) technique, we decided to use D-T2 analysis because it gave results consistent with the gamma ray and density neutron logs. The D-T2 Analysis module is a software implementation of the 2D/3D NMR technique that solves for the T2 and D, or T1 /T2, T2, and D distributions from wireline magnetic resonance data. Often the results are shown in 2D image form by summation of the individual images having different T1 /T2. The fluid typing result from D-T2 analysis showed the presence of gas and light hydrocarbon. The gas had an average saturation of 70%, while the oil had an average saturation of 65%. The oil viscosity varied slightly with the overburden gradient, increasing from 0.36cP at the top, to 0.55cP at the bottom. Although the resistivity-based saturation was not compared with the MR-derived saturations, the saturation result and other computed results match the average reservoir properties of offset wells within the vicinity of the study well. The results are also consistent with the gamma ray and density neutron logs.