The Application of Computed Tomography Scanning and Nuclear Magnetic Resonance for Rock Typing of Polymineral Clastic Reservoirs

Summary The main objective of this study was to provide rock typing of the producing formation based on high-resolution computed tomography (CT) scanning and nuclear magnetic resonance (NMR) data in combination with routine core analyses results. The target formation is composed of a shallowing up sequence of clastic rocks. Siltstones in its base are gradually replaced by sandstones toward its top. Initially, only sandstones were considered as oil-bearing, while siltstones were considered as water-bearing based on saturation calculation by means of Archie’s equation (Archie 1942) with the same values of cementation and saturation exponent for the whole formation. However, follow-up well tests detected considerable oil inflow also from the base of the reservoir composed of siltstones. Therefore, better rock typing was needed to improve the initial saturation distribution calculation. An applied approach that was based on integrated analysis of rock microstructural characteristics and derived from the NMR and CT techniques and conventional properties used for reserves calculation appeared to be an effective tool for rock typing polymineral clastic reservoirs. Measuring porous network characteristics and conventional properties in the same core plug enables a confident correlation between all measured parameters. Consequently, rock typing of samples based on flow units’ microstructural characteristics derived from NMR and CT scanning has shown a very good consistency with each other. As a result, four rock types were distinguished within a formation, which were previously interpreted as a single rock type. The detailed rock typing of the reservoir allowed more accurate reserves calculation and involvement of additional intervals into the production. Besides porous media characterization, CT scanning proved to be an effective tool for detecting minerals, such as pyrite and carbonates, characterizing depositional environments. Increasing content of pyrite in siltstones, detected by CT scanning and X-ray fluorescence spectroscopy, indicates deeper and less oxic conditions, while the presence of carbonate shell debris indicates shallower, more oxic depositional settings. The NMR test results show that the NMR signal distribution is affected by both pore size distribution and mineralogical composition. An increase of pyrite content caused shifting of the T2 distribution to the lower values, while carbonate inclusions caused shifting of the T2 distribution to higher values relative to the other samples not affected by these mineral inclusions. Because NMR distribution is affected by multiple factors, applying Т2cutoff values alone for rock typing can lead to ambiguous interpretation. Applying CT scanning next to NMR data increases the reliability of rock typing. The proposed laboratory workflow, including a combination of nonhazardous and nondestructive tests, allowed reliable differentiation of the rock samples based on multiple parameters that were interpreted in relationship with each other. Because the designed laboratory test workflow enabled both justified separation of the samples by rock type and determination of parameters used for reserves calculation, it can be recommended for further application in polymineral clastic reservoirs. Because the proposed techniques are nondestructive, the same samples can be applied for multiple tests including special core analysis (or SCAL).

Temperature Dependence of NMR Relaxation Time in Carbonate Reservoirs

NMR-based carbonate interpretation models are commonly calibrated using laboratory ambient core NMR measurements. For applying the core calibrated models to downhole NMR logging interpretation, the difference between the NMR responses measured at ambient and reservoir conditions needs to be evaluated. The temperature dependence of NMR relaxation time in high-quality carbonate reservoirs was investigated, and NMR temperature dependence models were determined using data analytic methods (Hursan et al, 2019). This paper focuses on temperature dependence of NMR relaxation time in low-quality carbonate formations. For more than 95% of the samples investigated, NMR relaxation time shows a positive correlation with temperature. The correlation is similar to that observed in high-quality carbonate rocks but slightly less significant. Temperature dependent correlations for predicting T2GM from a measured temperature to any other temperature are derived from high- and low-quality carbonate rocks independently first, then a unified T2GM correlation is derived including both the high- and low-quality carbonate reservoirs. Predicting T2 distribution from one temperature to other temperatures is achieved using dimension reduction approach involving principal component analysis (PCA) technique. It is found that the T2 distributions at any given temperature for both the high- and low-quality carbonate reservoirs can be predicted robustly from the T2 distributions at the ambient temperature by representing the T2 distributions with principal components (PCs) at the ambient temperature then using these PCs to predict the PCs at a different temperature. The optimal number of PC components depends on the multimodality of the T2 distribution. This work extends the validity range of a data analytic method that quantifies the temperature dependence of carbonate NMR properties. The new NMR temperature model enables the integration of NMR laboratory studies and dowhole measurements for advanced petrophysical analyses in a wide range of carbonate reservoirs.

Eliminating Diffusion Effects from NMR Logging Data for Enhanced Carbonate Pore Typing

Transverse relaxation (T2) times measured by multi-frequency, multi-gradient nuclear magnetic resonance (NMR) logging tools are affected by diffusion-induced enhanced relaxation which reduces the sensitivity to pore size in slow-relaxing formations such as macroporous carbonates and complicates the integration with zero-gradient core NMR data. We propose a solution for eliminating the diffusion-related uncertainties using intrinsic T2 distributions, obtained by a new inversion-forward modeling-inversion (IFMI) method, for carbonate pore typing applications. The NMR logs presented in this paper are based on data measured at five frequencies where the static magnetic field gradient varies from 26 to 55 G/cm. The high-quality echo signals are processed using a three-step IFMI differential signal analysis approach which nullifies diffusion effects due to the tool gradient and the potentially present internal gradient caused by paramagnetic minerals in the formation. The resulting diffusion-free intrinsic T2 distribution accentuates fine pore size variations and allows better discernment of micro-, meso-, and macropore systems of complex carbonate reservoirs. Multi-frequency NMR data, acquired in multiple wells, were processed and analyzed in several ways. First, apparent T2 distributions were obtained separately for individual frequencies. Discrepancies between the results of different frequencies clearly indicated that in macro- and mesoporous carbonates the diffusion effect is significant even with TE=0.3ms. This leads a peak broadening observed in the apparent T2 spectrum from conventional NMR processing, where echo trains from different frequencies are averaged in time-domain prior to the inversion. With the IFMI processing, individual-frequency echo trains are first pre-processed using a 2D NMR inversion whose results are used to forward model a diffusion-free echo train without prior assumptions on reservoir fluid diffusivity D. A second inversion, applied on the diffusion-free echo train, yields the intrinsic T2 distribution. The intrinsic T2 distribution has a noticeably higher spectral resolution in carbonate formations where diffusion effect is significant. The intrinsic T2 logs are expected to be more consistent with other gradient-free NMR measurements such as core NMR or LWD NMR data sets.

NMR Fluid Substitution–A New Method of Reconstructing T2 Distributions Under Primary Drainage and Imbibition Conditions

Most nuclear magnetic resonance (NMR)-based petrophysics models, such as pore structure characterization and permeability prediction, were developed using T2 distributions measured at fully water-saturated conditions (i.e., Sw = 1). The downhole implementation of those models across the hydrocarbon zones is disputable due to partial saturation (Sw < 1) conditions; hence, a correction to such effects on T2 distributions is required. This paper provides a critical review of the fluid substitution methods currently available in the industry and presents an improved method for enhanced formation evaluation. In the new method presented, an effective irreducible water saturation model is used to account for the pore structure and capillary pressure effects, which were barely considered by the currently available NMR fluid substitution methods. For water-wet reservoir rocks, the typical NMR T2 distribution at the partial saturation condition displays a clear separation between the wetting and nonwetting phases. The water phase can be classified as irreducible and movable fluid volumes. Then, using a T2 mapping relationship and a total porosity constraint, the T2 distribution of movable water at Sw < 1 is shifted and amplified to determine the T2 distribution of movable water at Sw = 1. To validate the new method, NMR measurements were conducted on sandstone samples at Sw = 1 as well as Sw < 1. The reconstructed T2 distribution at Sw = 1 was compared with the measured T2 distribution at Sw = 1. Results showed that the reconstructed T2 distribution matched very well with the T2 distribution measured at Sw = 1, confirming the robustness of the new technique. Parameters used in the reconstruction methodology are observed to be a good indicator of pore connectivity. During desaturation, the water T2 in large pores shifts to a shorter T2 because of the enhanced surface relaxation as the water volume decreases while the surface area remains constant. Therefore, the amplitude at the short T2 increases. The increased amplitude was remapped to large pores in reconstructing T2 spectra of full saturation.

Study on Magnetic Resonance Characteristics of Coal Sample under Progressive Loads

With the characteristics of gradual instability in the supporting pressure area of roadway as the engineering background, this paper aims to explore the evolution law of pore and fracture in the coal sample under progressive loads. The low-field nuclear magnetic resonance (NMR) test was designed and conducted with the coal sample under different axial loads (0, 3, 5, 7, 9, and 11 MPa). The characteristic parameters such as the porosity, the pore size distribution, the transverse relaxation time (T2) distribution curve, and the magnetic resonance image (MRI) were obtained. As the test results show, significant difference in the NMR characteristics of the coal samples can be observed throughout the compaction stage and the elastic stage. In the compaction stage, the porosity of the coal samples decreases slightly; the T2 distribution curve moves to the smaller value as a whole, and the percolation pore (PP) displays a tendency to transform to the adsorption pore (AP). In the elastic stage, the porosity of the coal samples rises gradually as the load increases; the T2 distribution curve moves to the larger value as a whole, and the AP tends to transform to the PP. The MRI shows that some pores and fissures in the coal sample close up and disappear as the load increases gradually, while the main pores and fissures expand and perforate till the macro failure occurs. Compared with one-time loading, the progressive multiple loads can ensure the fracture of the coal sample to develop more fully and the damage degree higher. It indirectly reflects that the instability and failure of the coal under the progressive load has the stage characteristics, verifying that the coal in the supporting pressure area needs to be controlled in advance.