Summary Effective interpretation of nuclear magnetic resonance (NMR) logs in shaly sands requires an understanding of the NMR contribution of clays. Of particular importance is the role of clays in the rapidly relaxing part of the NMR signal. In this study we measured the transverse relaxation time spectrum T2 of brine mixed with four clays (illite, smectite, kaolinite and glauconite) as a function of compaction. The Larmor frequency was 2 MHz and the echo spacing 0.16 ms. Mild compaction was achieved by centrifuging the clay slurry at three successive pressures ranging from 1 to 125 psi. Highly compacted samples were produced in a uniaxial press at six sequential pressures ranging from 500 to 16,000 psi. Each clay/brine slurry and its associated compacted sample showed a single peak in the T2 distribution spectrum. A second peak, which could be interpreted as the "clay-bound water," was never observed. The T2 peak position shifted to faster relaxation times as compaction increased, in proportion to the pore volume-to-surface ratio, Vp/As. The single peak and Vp/As proportionality are consistent with fast diffusion between the pore water and the monolayer of water on the clay surface. Surface relaxivity varied among the four clay minerals; glauconite, the clay with the highest magnetic susceptibility and iron content had the largest surface relaxivity. These results have important implications for the interpretation of NMR logs in shaly sands. Because of the effects of compaction and to a lesser extent the iron content on a clay's T2 peak position, it is not possible to independently determine clay type from some characteristic relaxation time. These data also imply that it is not feasible to estimate the cation exchange capacity from a single time cutoff of the T2 distribution without additional information such as laboratory measurements or other log data. Introduction Nuclear magnetic resonance logging has become an important tool in evaluating a formation's petrophysical properties. The unique and valuable advantage that NMR provides is pore size distribution information. No other logging method provides these data, which are the key component of log-based estimates of capillary-bound water volume, and permeability to flow.1 It has been proposed that NMR logging can be extended to estimate clay-bound water volumes, and identify clay minerals. Clay-bound water volume, important in determining water saturation from resistivity, has been correlated with the short-T2 less than 3 ms, porosity of 45 oilfield sandstones.2 Prammer et al.'s3 NMR clay/brine study found that the T2 distribution of clay-bound water associated with kaolinite and chlorite was greater than 3 ms, for illite it ranged between 1 and 2 ms and for smectite it was less than 1 ms. Observed T2 's were then used as an indicator of cation exchange capacity (CEC) because the number of available exchange sites is proportional to a clay's specific surface area. CEC is fundamental to converting bulk resistivity measurements into water saturation. The ability to estimate clay-bound water, and to identify the clay type, from NMR T2 distributions is not compatible with the ability to determine pore size distribution from the same data. In the first two cases the molecular diffusion rate of water in the pores must be slow, whereas in the latter case it is assumed to be fast. For example, consider a monolayer of water on the surface of room-dry clay. The monolayer has a short relaxation time, less than a millisecond, because of its interaction with the solid rock. Now fill the void space between the clay particles with water and consider the two extreme cases. In the first case, there is no molecular diffusion (exchange) between the surface-monolayer water and bulk water. Thus, the T2 spectrum will contain two separate peaks, one associated with the surface-monolayer water at less than a millisecond and one associated with the bulk water. In the opposite case, molecular diffusion is highly effective, and both the surface monolayer and bulk water have a common relaxation time, a single peak in the T2 spectrum with time constant: 1 T 2 = ρ s ( A s V p ) . ( 1 ) In this equation, which provides the fundamental connection between T2 and pore size, the term ?2 is the surface relaxivity parameter that indicates the capacity of the rock to cause the decay of magnetization in the water. Fig. 1 is a conceptual drawing of a T2 distribution for a sandstone that includes fluid in small pores that are typically associated with clays, capillary-bound and producible fluid. The objective of this study was to determine whether it is possible to infer a clay-bound water volume (peak or T2 cutoff), or clay type, and a pore size distribution from a NMR distribution spectrum. To achieve this goal we designed a set of experiments that examined the NMR relaxation of clay/brine mixtures at various compaction states. In contrast to shaly sands, the clay/brine mixtures provided the means to minimize the pore volume-to-surface ratio, so that any water-monolayer-related signal might be detected. The pore volume-to-surface ratio was easily varied through compaction, and the monomineralic samples enable the NMR response of individual clay types to be evaluated. We chose to study four clays commonly found in oil-bearing sedimentary environments: kaolinite, illite, smectite and glauconite. Experimental Procedures Samples of illite and glauconite were obtained from Wards Natural Science Establishment. Kaolinite and smectite (Ca Montmorillonite) were procured from ECC Intl., Georgia Kaolin Co. and the Source Clay Minerals Repository, respectively. Various physical properties of the clays were measured. Prior to the surface area and magnetic susceptibility measurements, the clays were dried overnight in a vacuum oven at 100°C. Surface area measurements were collected using the Micromeritics Gemini 2360 with nitrogen gas as adsorbate, and magnetic susceptibilities were measured on a Johnson Matthey MSB-AUTO magnetic susceptibility balance. CEC measurements were taken using the ammonium acetate/ammonium ion-specific electrode method by David K. Davies & Assoc., Inc. Table 1 is a summary of clay type, clay origin and clay physical properties. The samples were analyzed for mineralogy using dual-range Fourier transform infrared (FTIR) spectroscopy.4 In addition, the samples were sent to X-Ray Assay Laboratories for chemical analyses (Table 2). The mineralogy data indicate the presence of quartz in the illite sample. Consequently, a <2 ?m fraction of the sample was extracted by centrifuging the illite and drying the supernatant. The physical properties for these clays are in good agreement with those in the literature.5 The clay samples used for the room-dry and clay/brine mixture NMR experiments were kept at room temperature and at typical laboratory humidity conditions of 50%. To evaluate whether clay samples have a measurable NMR signal at room-dry conditions, samples were prepared by placing the clay in a test tube and sealing it off with a stopper and Teflon tape.
The interpretation of MWD and wireline logs of invaded formations can be enhanced by an understanding of the controls on invasion rates. Experimental investigation of dynamic filtration rates has shown that under certain conditions these are dependent only on the hydraulic shear stress at the mudcake surface for a given mud composition. They are virtually independent of differential pressure and independent of porous medium permeability (as opposed to mudcake properties) over porous medium permeability (as opposed to mudcake properties) over a 1:10 range. For very low permeability media however, the clear fluid Darcy flux may already be below the dynamic filtration rate. Then no mudcake forms and filtration rates obey Darcy's law. Comparison with theory suggests that there is a critical flux (filtration rate per unit area) above which day particles accrete irreversibly to the mudcake; below this flux they cannot stick at all. Invasion volumes at the time of wireline logging are usually dominated by fluid loss that occurred under conditions of dynamic filtration. The conjecture of a critical dynamic filtration rate, dependent only on mud composition and shear, means that invasion volumes should be independent of mud overbalance pressure and of formation properties (permeability, saturation, far field pressure, internal mudcake) unless formation permeability is so low that it limits the filtration flux to less permeability is so low that it limits the filtration flux to less than the critical flux. Then, formation properties control fluid loss. The permeability at which the behavior changes depends on both mud overbalance pressure and mud flow rate. These results are important for the interpretation of MWD and wireline logs in the presence of invasion. Field examples of multiple pass MWD and wireline resistivity logs and invasion volume computations from a single pass of wireline resistivity and porosity logs often indicate trends in the invasion volume within a zone or between zones. These trends may be interpreted as indicating differential formation resistance to invasion at low permeabilities but must be due to factors other than the loss rate at moderate to high permeabilities. One such factor is vertical movement of the filtrate under the Influence of gravity. Introduction Log analysts often seen to infer reservoir dynamic properties by interpreting the invasion profile at the time of wireline logging, or changes in the invasion profile between MWD and wireline logging times. An understanding of the influences on the rates at which filtrate is lost to the formation from the wellbore, particularly for those conditions under which the largest net volumes are lost, is a first step in this process. Fluid loss is conventionally divided into three categories: beneath-the-bit, static filtration and dynamic filtration. Fluid loss beneath the bit occurs as fresh rock is exposed. The loss rate is highest herb but ends within seconds, as pores near the wellbore are bridged, forming an internal mudcake. From published estimates one may conclude that the total invasion depths from volume lost in this phase are unlikely to much exceed half a wellbore radius. it is thus not of first importance for evaluating fluid loss at timescales exceeding half an hour. After the bit passage, further fluid loss involves mud filtration forming an external mudcake with concomitant invasion of the formation by mud filtrate. It is usual to distinguish two regimes: static and dynamic. These are really misnomers, since they refer only to mud flow in the well, not to the magnitude or time development of the loss rate. Dynamic filtration (referred to in chemical engineering as crossflow filtration) occurs when mud is flowing past the surface of the mudcake; static filtration takes over when the mud flow is stopped. For static filtration, the loss rate q(t) declines indefinitely as mudcake grows, in principle until the mudcake entirely plugs the wellbore. In dynamic filtration, q(t) declines until the mudcake stops growing and q(t) reaches (asymptotically) a steady state, with a nearly constant loss rate qD. Usually the majority of the fluid volume lost during time intervals of interest in formation evaluation is lost under dynamic conditions. This is because for most reservoir rocks, loss rates are greater for dynamic conditions than for static and because mud is circulating for the majority of the time for safety and hole deaning purposes. purposes. We describe in this paper simple semi-quantitative models for a complex process, concentrating on the principal features understood from existing literature and research, discounting effects and influences of secondary importance. Simple formulas are given; these deliberately sacrifice any attempt at high accuracy to serve as straightforward conceptual models for the major features of wellbore filtration. Most of what follows refers to orthodox water based muds, with bentonite as the primary viscosifier and fluid loss control material. More experimental data are required for other mud types to be confident in the applicability of these ideas to them. FLUID LOSS RATES IN DYNAMIC AND STATIC FILTRATION How the flow rate per unit area q(t) changes over time is significantly different between static and dynamic conditions for large times t. For small times however it turns out that the two regimes are closely similar. For this reason we consider static filtration first. P. 39
Effective interpretation of nuclear magnetic resonance (NMR) logs in shaly sands requires an understanding of the NMR contribution of clays. Of particular importance is the role of clays in the rapidly relaxing part of the NMR signal. In this study we measured the transverse relaxation time spectrum (T2) of brine mixed with four clays (illite, smectite, kaolinite and glauconite) as a function of compaction. The Larmor frequency was 2 MHz and the echo spacing 0.16 msec. Mild compaction was achieved by centrifuging the clay slurry at three successive pressures ranging from 1 to 125 psi. Highly compacted samples were produced in a uniaxial press at six sequential pressures ranging from 500 to 16,000 psi. Each clay-brine slurry and its associated compacted sample showed a single peak in the T2 distribution spectrum. A second peak, which could be interpreted as the "clay-bound water," was never observed. The T2 peak position shifted to faster relaxation times as compaction increased, in proportion to the pore volume-to-surface ratio, V/S. The single peak and VP'S proportionality are consistent with fast diffusion between the pore water and the monolayer of water on the clay surface. Surface relaxivity varied among the four clay minerals; glauconite, the clay with the highest magnetic susceptibility and iron content had the largest surface relaxivity. These results have important implications for the interpretation of NMR logs in shaly sands. Because of the effects of compaction and to a lesser extent the iron content on a clay's T2 peak position, it is not possible to independently determine clay type from some characteristic relaxation time. These data also imply that it is not feasible to estimate the cation exchange capacity from a single time cutoff of the T2 distribution without additional information such as laboratory measurements or other log data. P. 205
The compensated dual resistivity (CDRSM) tool is an electromagnetic propagation tool for measurement while drilling. The CDR tool provides two resistivity measurements with several novel features that are verified with theoretical modeling, test-tank experiments, and log examples. IntroductionThe CDR tool 1 is a 2 ~ 106-cycles/sec electromagnetic propagation tool 2 . 3 built into a drill collar. This drill collar is fully selfcontained and has rugged sensors and electronics. The CDR tool is borehole-compensated, requiring two transmitters and two receivers. The transmitters alternately broadcast electromagnetic waves, and the phase shifts and attenuations are measured between the receivers and averaged. Phase shift is transformed into a shallow measurement, R ps ' and attenuation is transformed into a deep measurement, Rad' The CDR tool has several new and important features. I. Rad and Rps provide two depths of investigation and are used to detect invasion while drilling. For example, in a I-a· m formation, the investigation diameters (50% response) are 30 in. for Rps and 50 in. for Rad' 2. ROil and Rps detect beds as thin as 6 in.; however, these measurements are affected differently by shoulder-bed resistivities and both require corrections in thin resistive beds. Rps has a better vertical response than Rad' Rad and Rps cross over at the horizontal bed boundaries; this crossover can be used to measure bed thickness.3. Both Rad and Rps are insensitive to hole size and mud resistivity in smooth boreholes. Borehole corrections are very small even for contrasts of 100: I between formation and mud resistivities. Rugose holes and salty muds together, however, can cause larger errors than indicated by the borehole-correction charts. In these conditions, borehole compensation is essential for an accurate measurement.An extensive research program was conducted to verif'y these features and to ensure that the CDR tool provides a high-quality log. To achieve wireline quality, the CDR tool's physics was studied thoroughly, and its environmental effects were modeled and experimentaly measured. Two theoretical models are used for the CDR tool. The first model treats the tool geometry in detail but assumes a homogeneous medium outside the tool. This model is verified by test-tank experiments and by air measurements. The second model assumes a simplified tool geometry but treats boreholes, caves, beds, and invasion in detail. This model is used to study environmental effects and to prepare correction charts. Experiments with artificial boreholes, caves, step-profile invasion, and horizontal bed boundaries verif'y the predictions of the second model. Finally, CDR logs are compared to wireline logs to demonstrate the new features.
TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractAn extensive petrophysical analysis has been carried out on a study well in a giant carbonate field in the Lower Cretaceous of Abu Dhabi. The well was cored, and a comprehensive log suite recorded including: Nuclear Magnetic Resonance (NMR), Borehole Imaging, Shear Sonic and Mineral Spectroscopy in addition to the more normal logs.A high-resolution permeability profile was measured on the core. Plugs taken every foot were supplemented by minipermeameter measurements every inch. While this profile captures high frequency variability it is limited in its applicability to predicting reservoir behavior. Analysis of large scale flow unit permeability is better assessed with dynamic data.A number of advanced log based permeability estimators were developed in an attempt to match the core permeability profile. The estimators are based on NMR, image analysis and Stoneley measurements among others. This paper presents a comparison of these predictors to the core and makes recommendations as to the preferred permeability prediction methods for this field.For this well the Stoneley permeability correctly predicted a trend decrease of permeability with depth but failed to honor the full extent of the trend. As expected, given its vertical resolution, it also failed to respond adequately in the more heterogeneous intervals.A method based on a mean NMR T 2 measurement gave a good general match to the data, but again lacked vertical resolution. An electrical image based method brings an important improvement in sensitivity to the small-scale variations in permeability, but in some places failed to follow the lower frequency trends.A composite method using the NMR to supply the overall trend, and the image to provide the higher resolution, gave the best match to the core permeability profile available from the tool suite, which was run. The method is a modification of the well-established SDR permeability equation. A new term is introduced to account for the contribution of connected vugs, which are responsible for the high permeability intervals in this well.The logs used in the methods described here do not make direct measurements of permeability. Instead permeability is inferred indirectly from measurements sensitive to parameters such as texture and pore size. It is doubtful that a single methodology based on such measurements can provide a "universal" permeability estimator fitting all cases. The methods presented here should be applied with care and will benefit greatly from calibration to spot permeability data.The method has particular application in high cost environments where a good permeability is required along the wellbore but the cost and rig time required for continuous whole coring cannot be justified.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.