Browsing by Subject "Reservoir characterization"

Now showing 1 - 20 of 34

Application of Advanced Reservoir Characterization, Simulation and Production Optimization Strategies to Maximize Recovery in Slope and Basin Clastic Reservoirs, West Texas (Delaware Basin)
(2001) Dutton, Shirley P.; Flanders, William A.
The objective of this Class III project was to demonstrate that reservoir characterization and enhanced oil recovery (EOR) by CO2 flood can increase production from slope and basin elastic reservoirs in sandstones of the Delaware Mountain Group in the Delaware Basin of West Texas and New Mexico. Phase 1 of the project, reservoir characterization, focused on Geraldine Ford and East Ford fields, which are Delaware Mountain Group fields that produce from the upper Bell Canyon Formation (Ramsey sandstone). The demonstration phase of the project was a CO2 flood conducted in East Ford field, which is operated by Orla Petco, Inc., as the East Ford unit. Reservoir characterization utilized outcrop characterization, high-resolution sequence stratigraphy, subsurface field studies, and 3-D seismic data. Ramsey sandstones are interpreted as having been deposited in a channel-levee system that terminated in broad lobes; overbank splays filled topographically low interchannel areas. Porosity and permeability of the reservoir sandstones are controlled by calcite cement that can be concentrated in layers ranging from 2 to 16 inches in thickness.
Application of Advanced Reservoir Characterization, Simulation and Production Optimization Strategies to Maximize Recovery in Slope and Basin Clastic Reservoirs, West Texas (Delaware Basin)
(2001) Dutton, Shirley P.; Flanders, William A.; Mendez, Daniel L.
The objective of this Class III project is to demonstrate that detailed reservoir characterization of slope and basin elastic reservoirs in sandstones of the Delaware Mountain Group in the Delaware Basin of West Texas and New Mexico is a cost-effective way to recover oil more economically through geologically based field development. The project is focused on East Ford, a Delaware Mountain Group field that produces from the upper Bell Canyon Formation (Ramsey sandstone). The field, discovered in 1960, is operated by Orla Petco, Inc., as the East Ford unit. A CO2 flood is being conducted in the unit, and this flood is the Phase 2 demonstration for the project. Reservoir characterization was conducted using logs and pressure and production information from the East Ford unit, supplemented by Bell Canyon outcrop data and information from the nearby Geraldine Ford field. Characterization was enhanced this year by recovery of a core from the Ramsey reservoir interval in the EFU 41R well. Ramsey sandstones are interpreted as having been deposited in a basin-floor setting in a channel-levee system with attached lobes. Overbank splays are interpreted as being the main area of sand storage outside of the channels. Porosity and permeability of the reservoir sandstones are controlled by calcite cement that can be concentrated in layers ranging from 5 to 40 cm in thickness. These laterally extensive calcite-cemented layers form significant vertical permeability baffles in some areas of the reservoir.
Application of Advanced Reservoir Characterization, Simulation, and Production Optimization Strategies to Maximize Recovery in Slope and Basin Clastic Reservoirs, West Texas (Delaware Basin)
(2000) Dutton, Shirley P.; Flanders, William A.; Zirczy, Helena
The objective of this Class III project is to demonstrate that detailed reservoir characterization of slope and basin elastic reservoirs in sandstones of the Delaware Mountain Group in the Delaware Basin of West Texas and New Mexico is a cost-effective way to recover a higher percentage of the original oil in place through geologically based field development. Phase 1 of the project, reservoir characterization, was completed this year, and Phase 2 began. The project is focused on East Ford field, a representative Delaware Mountain Group field that produces from the upper Bell Canyon Formation (Ramsey sandstone). The field, discovered in 1960, is operated by Orla Petco, Inc., as the East Ford unit. A CO2 flood is being conducted in the unit, and this flood is the Phase 2 demonstration for the project. The depositional model of the East Ford unit was revised on the basis of analysis of pressure and production information and reexamination of the outcrop data. Overbank splays were recognized in outcrop as being the main area of sand storage outside of the channels, not levees. This model has now been applied to the East Ford unit. Deposits flanking the Ramsey 1 and 2 channels are interpreted as consisting of narrow levees and wider overbank-splay sandstones. Deposits at the south end of the field are interpreted to be lobe sandstones in both Ramsey 1 and 2 intervals.
Application of Advanced Reservoir Characterization, Simulation, and Production Optimization Strategies to Maximize Recovery in Slope and Basin Clastic Reservoirs, West Texas (Delaware Basin)
(1999) Dutton, Shirley P.; Flanders, William A.; Guzman, Jose I.
The objective of this Class III project is to demonstrate that detailed reservoir characterization of slope and basin elastic reservoirs in sandstones of the Delaware Mountain Group in the Delaware Basin of West Texas and New Mexico is a cost-effective way to recover a higher percentage of the original oil in place through geologically based field development. This year, the project focused on reservoir characterization of the East Ford unit, a representative Delaware Mountain Group field that produces from the upper Bell Canyon Formation (Ramsey Sandstone). The field, discovered in 1960, is operated by Orla Petco, Inc., as the East Ford unit and contained an estimated 19.8 million barrels (MMbbl) of original oil in place. Petrophysical characterization of the East Ford unit was accomplished by integrating core and log data and quantifying petrophysical properties from wireline logs. Most methods of petrophysical analysis that had been developed during an earlier study of the Ford Geraldine unit were successfully transferred to the East Ford unit. However, the approach used to interpret water saturation from resistivity logs had to be modified because in some East Ford wells, the log-calculated water saturation was too high and inconsistent with observations made during the actual production. Log-porosity to core-porosity transforms and core-porosity to core permeability transforms were derived for the East Ford reservoir. The petrophysical data were used to map porosity, permeability, net pay, water saturation, mobile oil saturation, and other reservoir properties.
Application of Advanced Reservoir Characterization, Simulation, and Production Optimization Strategies to Maximize Recovery in Slope and Basin Clastic Reservoirs, West Texas (Delaware Basin)
(1998) Dutton, Shirley P.; Barton, Mark D.; Malik, Mohammad Abdul
The objective of this Class III project is to demonstrate that detailed reservoir characterization of elastic reservoirs in basinal sandstones of the Delaware Mountain Group in the Delaware Basin of West Texas and New Mexico is a cost-effective way to recover more of the original oil in place by strategic infill-well placement and geologically based enhanced oil recovery. The study focused on the Ford Geraldine unit, which produces from the upper Bell Canyon Formation (Ramsey sandstone). Reservoirs in this and other Delaware Mountain Group fields have low producibility (average recovery <14 percent of the original oil in place) because of a high degree of vertical and lateral heterogeneity caused by depositional processes and post-depositional diagenetic modification. Outcrop analogs were studied to better interpret the depositional processes that formed the reservoirs at the Ford Geraldine unit and to determine the dimensions of reservoir sandstone bodies. Facies relationships and bedding architecture within a single genetic unit exposed in outcrop in Culberson County, Texas, suggest that the sandstones were deposited in a system of channels and levees with attached lobes that initially prograded basinward, aggraded, and then stepped around and stepped back toward the shelf. Channel sandstones are 10 to 60 ft thick and 200 to 3,000 ft wide. The flanking levees have a wedge-shaped geometry and are composed of interbedded sandstone and siltstone; thickness varies from 3 to 20 ft and length from several hundred to several thousands of feet. The lobe sandstones are broad lens-shaped bodies; thicknesses range up to 30 ft with aspect ratios (width/thickness) of 100 to 10,000. Lobe sandstones may be interstratified with laminated siltstones.
Characterization and modeling of paleokarst reservoirs using multiple-point statistics on a non-gridded basis
(2012-12) Erzeybek Balan, Selin; Srinivasan, Sanjay; Lake, Larry W; Bryant, Steven L; Janson, Xavier; Zahm, Christopher Kent
Paleokarst reservoirs consist of complex cave networks, which are formed by various mechanisms and associated collapsed cave facies. Traditionally, cave structures are defined using variogram-based methods in flow models and this description does not precisely represent the reservoir geology. Algorithms based on multiple-point statistics (MPS) are widely used in modeling complex geologic structures. Statistics required for these algorithms are inferred from gridded training images. However, structures like modern cave networks are represented by point data sets. Thus, it is not practical to apply rigid and gridded templates and training images for the simulation of such features. Therefore, a quantitative algorithm to characterize and model paleokarst reservoirs based on physical and geological attributes is needed. In this study, a unique non-gridded MPS analysis and pattern simulation algorithms are developed to infer statistics from modern cave networks and simulate distribution of cave structures in paleokarst reservoirs. Non-gridded MPS technique is practical by eliminating use of grids and gridding procedure, which is challenging to apply on cave network due to its complex structure. Statistics are calculated using commonly available cave networks, which are only represented by central line coordinates sampled along the accessible cave passages. Once the statistics are calibrated, a cave network is simulated by using a pattern simulation algorithm in which the simulation is conditioned to sparse data in the form of locations with cave facies or coordinates of cave structures. To get an accurate model for the spatial extent of the cave facies, an algorithm is also developed to simulate cave zone thickness while simulating the network. The proposed techniques are first implemented to represent connectivity statistics for synthetic data sets, which are used as point-set training images and are analogous to the data typically available for a cave network. Once the applicability of the algorithms is verified, non-gridded MPS analysis and pattern simulation are conducted for the Wind Cave located in South Dakota. The developed algorithms successfully characterize and model cave networks that can only be described by point sets. Subsequently, a cave network system is simulated for the Yates Field in West Texas which is a paleokarst reservoir. Well locations with cave facies and identified cave zone thickness values are used for conditioning the pattern simulation that utilizes the MP-histograms calibrated for Wind Cave. Then, the simulated cave network is implemented into flow simulation models to understand the effects of cave structures on fluid flow. Calibration of flow model against the primary production data is attempted to demonstrate that the pattern simulation algorithm yields detailed description of spatial distribution of cave facies. Moreover, impact of accurately representing network connectivity on flow responses is explored by a water injection case. Fluid flow responses are compared for models with cave networks that are constructed by non-gridded MPS and a traditional modeling workflow using sequential indicator simulation. Applications on the Yates Field show that the cave network and corresponding cave facies are successfully modeled by using the non-gridded MPS. Detailed description of cave facies in the reservoir yields accurate flow simulation results and better future predictions.
Characterization of the Cana-Woodford Shale using fractal-based, stochastic inversion, Canadian County, Oklahoma
(2016-05) Borgman, Barry Michael; Spikes, Kyle; Sen, Mrinal K; Wilson, Clark R
The past decade has seen a surge in unconventional hydrocarbon exploration and production, driven by advances in horizontal drilling and hydraulic fracturing. Even with such advances, reliable models of the subsurface are crucial in all phases of exploitation. This study focuses on the methods used for estimation of the elastic properties (density, velocity, and impedance), which play a key role in targeting reservoir zones ideal for hydraulic fracturing. Well-log data provides high-resolution vertical measurements of elastic properties, but a relatively shallow depth of investigation imposes spatial limitations. Seismic data provides broader horizontal coverage at lower cost, but sacrifices vertical resolution. Thin beds present in many unconventional reservoirs fall below seismic resolution. In addition, the band-limited nature of seismic data results in the absence of low-frequency content of the Earth model, as well as the high-frequency content present in well logs. Seismic inversion is a process that provides estimates of elastic properties given input seismic and well data. Stochastic inversion is a method that uses well-log data as a priori information, with an added aspect of randomness. The method generates many realizations using the same input model and takes an average of those realizations. We implement two separate stochastic inversion algorithms to estimate P-impedance in the Cana-Woodford Shale in west-central Oklahoma. First, we use a fractal-based, very fast simulated annealing algorithm that exploits the fractal characteristics found in well-log data to build a prior model. The method of very fast simulated annealing optimizes our elastic model by searching for the minimum misfit between observed and synthetic seismic traces. Next, we use a principal component analysis (PCA) based stochastic inversion algorithm to invert for impedance at all traces simultaneously. Comparison of the results with traditional deterministic inversion results shows improved vertical resolution while honoring the low-frequency content of the Earth model. The PCA-based inversion results also show improved lateral continuity of the elastic profile along our 2D line. The impedance profile from the PCA-based approach provides a better representation of the vertical and horizontal variability of the reservoir, allowing for improved targeting of frackable zones.
Characterizing reservoir quality for geologic storage of CO2 : a case study from the Lower Miocene shore zone at Matagorda Bay, Texas
(2021-05-10) Hull, Harry Lejeune; Meckel, Timothy Ashworth
The geologic storage of anthropogenic CO₂ through Carbon Capture, Utilization, and Storage (CCUS) is necessary to reduce the emissions produced as a biproduct of fossil fuel combustion. This process of injecting CO₂ into the subsurface is known as carbon sequestration and requires the assessment of geologic reservoirs. Depositional processes and the resulting facies and stratigraphic architectures have great influence over reservoir volumetrics and behavior. The objective of this study is to constrain the depositional controls on storage capacity. A subsurface Lower Miocene 2 strandplain/barrier bar complex of the Texas Gulf Coast at Matagorda bay is interpreted and modeled using well data and 3D seismic. These data reveal the presence of a major shore zone that experienced initial progradation through the late highstand and into the lowstand before later retrogradation. The LM2 is then capped by a thick regional shale. A stratigraphic framework is built that captures these changes in shoreline position at both the systems tract and parasequences level. Sediments were strike fed and wave-dominated processes are apparent. Petrophysical properties of this region including porosity are modeled from with machine learning from log data. Machine learning to predict porosity is carried out using a random forest regression in which porosity is a function of lithology and depth. Finally, a 3D reservoir model is built integrating the stratigraphic, facies, and petrophysical properties. Static storage capacity estimates and storage capacity maps are created from the 3D model. Storage capacity is observed to occur at a strike parallel geometry. This “axis” of highest storage capacity tracts with the position of the shore zone in vertical succession highlighting a dependence on the balance between the generation of accommodation and sediment supply. At a higher resolution storage capacity is observed highest within the foreshore where beach ridges are interpreted from seismic stratal slices. High wave energy processes at this position in the shoreline profile are known to create well sorted and therefore highly porous sandstones. Storage capacity is then a direct function of the high wave energy paleo-depositional processes occurring at the shoreline
Cluster Analysis in Reservoir Characterization
(1994-08) Muneta, Yasuhiro; Lake, Larry W.
Any raw data sampled in an oil field has a certain amount of noise; the sample may be called an obscure image of the real thing. We may eliminate the noise by a process of "image enhancement" in "statistical pattern recognition." Image enhancement is one of the important steps in processing large data sets to make them more suitable for classification than were the original data. In this work, cluster analysis, which is a method of image enhancement, is applied to some reservoir characterization problems such as permeability distributions of core samples, sand/shale sequences observed in wells, and pressure distributions in heterogeneous porous media to classify the sample data and find the intrinsic patterns (averaged images) from the original data sets. Cluster analysis is a multivariate statistical method. It is very general and can be applied to a wide area of scientific investigations. It is often called a tool of discovery or an unsupervised approach which doesn't depend on a priori information. It searches unknown-significant categories (patterns) themselves. Once we obtain typical patterns, we may analogously approach the real thing based on them. We find that cluster analysis is applicable to finding appropriate parent populations of a permeability distribution, theoretical indicator variograms of sand/shale sequences, and trends of effective permeability distribution.
Geostatistical Characterization of Naturally Fractured Reservoirs
(2004-05) Liu, Xiao Huan; Srinivasan, Sanjay
Natural fractures are commonly observed in many major reservoirs worldwide and contribute significantly to worldwide oil production. Characterization of fractures is necessary in order to make accurate forecast of reservoir performance. However, fracture reservoir characterization is not easy due to insufficient information generally derived from cores and logs. The main characterization tools are geological classification, geomechanical characterization, pattern recognition and stochastic simulation. Geological classification is based on analysis of paleo-stress conditions in the reservoir at the time of fracturing. Geomechanical characterization is based on utilization of the plausible stress state of the reservoir in order to predict the distribution and orientations of fractures. The disadvantage of this tool is the models are largely deterministic (i.e. uncertainty of fracture propagation might not be captured). Fracture patterns can also be classified using the information obtained from outcrop. Although some information supplementary to cores and logs can be obtained from outcrop and used to classify patterns in detail, environmental factors such as weathering would constrain the inference & application of outcrop patterns to model target reservoir. Besides, the extent to which the outcrop is analogous to the target reservoir is difficult to ascertain a priori. The stochastic simulation approach is based on application of some statistical interpolation algorithm such as kriging or cokriging that can be used to obtain estimates of necessary conditional distributions from which fracture patterns can be sampled. Since it does not consider any geomechanical criteria for fracture pattern generation, the model based on stochastic simulation is not physical. Therefore integration of the information derived from more realistically deterministic geomechanical models is necessary to develop physically realistic stochastic fracture models. This is a primary focus of this research. The spatial distribution of fractures in a reservoir affects the displacement of fluids and the prediction of future performance. Realistic characterization of fractured reservoirs requires quantification and classification of fracture patterns on the basis of the underlying geological characteristics and developing reservoir modeling algorithms that can integrate connectivity (multiple point) based statistics related to fracture patterns. Developing a methodology for summarizing the characteristics of fracture networks based on multiple point connectivity characteristics derived from analog models and all other available reservoir specific data in the form of well information, conditioning reservoir models, geomechanical models and seismic areal proportion maps is the primary objective of the research.
Geostatistical Reservoir Characterization and Scale-Up of Permeability and Relative Permeabilities
(1996-12) Malik, Mohammad Abdul; Lake, Larry W.
In this dissertation several unresolved issues related to reservoir characterization and simulation techniques are investigated so that improvements can be made in their reliability and efficiency. Geostatistical techniques are now commonly used to generate interwell permeability distributions. However, these techniques have not been adequately validated because of scarcity of reference or truth cases for subsurface reservoirs. This study uses detailed deterministic permeability distributions, and corresponding fluid-flow results from an outcrop study as a reference case, to validate stochastic permeability distributions generated by conditional simulation (CS). Only as much data from the outcrop are used as would normally be available in a subsurface reservoir. This is a first of its kind test on the reliability of stochastic permeability distributions. Results indicate that, although CS is highly flexible and generates realistic heterogeneity, it must be adapted to the specific geologic environment for best agreement with deterministic simulations.The reliability of a stochastic permeability distribution depends upon the correct statistical analysis of available data and careful inference of statistical parameters. This study uses core permeability data from an actual producing field to generate a three-dimensional CS permeability distribution. This example serves as a model with practical details for data evaluation, determination of autocorrelation structure and generation of multiple realizations of permeability distribution. The CS permeability distributions are evaluated for their conformity with reservoir geology so that the most realistic realizations can be selected. Millions of fine-scale blocks are required to adequately represent heterogeneity in a typical field. Because fluid-flow simulations on such a large number of blocks are still prohibitively expensive, fine-scale permeability distributions must be scaled-up. This work applies four methods for independent scale-up, of permeability from fine to a practical field-simulation scale. These methods include independent fine-scale simulation (IFS), the electrical network method (ENM), geometric averaging (GA) and the Cardwell-Parsons (CP) method. Permeability is also scaled-up by a dependent fine-scale simulation (DFS) scheme that acknowledges flow through neighboring coarse blocks in the process. Results from independent scale-up of realistic CS permeability cross sections show that IFS is an overall preferable method because of good accuracy and flexibility in application. Especially in low-permeability blocks, IFS and DFS estimates of scaled-up permeability can be significantly different. Closely related to permeability scale-up is the treatment of relative permeability. Laboratory-measured relative permeabilities or rock curves are assumed to apply to fine-scale simulations but they may or may not be applicable in a coarse-scale field simulation. In this research, relative permeabilities are scaled-up by steady-state as well as dynamic fluid-flow simulations. Results show that steady-state scaled-up curves are vii almost the same as the rock curves, while dynamic scaled-up curves are different. Comparison of results from fluid-flow simulations from fine and corresponding coarse cross sections shows that rock curves perform better than dynamic scaled-up curves in coarse simulations, provided the dimensions of the problem are not reduced (2D cross section remains 2D after scale-up) and adequate heterogeneity is retained in the scaled-up field. The small difference in effluent-water fractional flow results from fine and coarse simulations with rock curves appears to be caused by increased numerical dispersion in the coarse simulations.
Heterogeneity study of the Little Creek field from petrophysical data
(2018-06-20) Salazar Neira, Jose Julian; Lake, Larry W.
Reservoir characterization determines the quality and characteristics of a subsurface body to accumulate and produce oil. This characterization involves the study of reservoir heterogeneity, which has the objective to estimate the variability of geological properties directly related to flow and, consequently, they have a great impact in the overall production of the field. Although a thorough analysis of heterogeneity is performed, it is common to estimate the reservoir heterogeneity with screening tools that offer a brief assessment of the variability in the field. Nevertheless, there is a tendency of the screening methods to either i) consider the petrophysical data as an indivisible set, or ii) to not relate flow performance with heterogeneity. Otherwise, the decisions executed based on an inaccurate model would be wrong. Thus, this thesis focuses to overcome the previous drawbacks studying the interstitial velocity of the Little Creek field with two methods. First, statistical partitioning attempts to associate the producing sandstones of the field in terms of the assumed constituent parents. Second, the Koval theory is used to better estimate the heterogeneity as a function of the recovery process to be undertaken and to forecast the flow performance of an immiscible process as a function of the more representative heterogeneity appraisal. Even though the statistical partitioning did not correlate lognormal distributed parents with the producing facies, insightful results were obtained from this work. First, the overall interstitial velocity of the field, and from most of the wells studied, do not follow any of the probability distributions studied in this thesis. Second, roughly 73% of the wells studied are statistically different in their interstitial velocity means, standard deviations, or both. Third, the region within flow and storage capacities less than 90% are more meaningful for a practical recovery process. Finally, the Koval theory estimates the heterogeneity at the intra-well scale and predicts better production performance for low heterogeneous reservoirs, agreeing with field results. Hence, the Koval theory could be categorized as a reliable screening tool to estimate heterogeneity and accurately forecast the performance of recovery processes that are immiscible, like waterfloods. In other words, it is a helpful method to complement a successful reservoir characterization
Interrelationships between carbonate diagenesis and fracture development : example from Monterrey Salient, Mexico and implications for hydrocarbon reservoir characterization
(2012-05) Monroy Santiago, Faustino; Marrett, Randall; Laubach, Stephen E.; Fisher, William L.; Gale, Julia; Kerans, Charles
Many low matrix-porosity hydrocarbon reservoirs are productive because permeability is controlled by natural fractures. The understanding of basic fracture properties is critical in reducing geological risk and therefore reducing well costs and increasing well recovery. Unfortunately, neither geophysics nor borehole methods are, so far, accurate in the acquisition of key fracture attributes, such as density, porosity, spacing and conductivity. This study proposes a new protocol to predict key fracture characteristics of subsurface carbonate rocks and describes how using a relatively low-cost but rock-based method it is possible to obtain accurate geological information from rock samples to predict fracture attributes in nearby but unsampled areas. This methodology is based on the integration of observations of diagenetic fabrics and fracture analyses of carbonate rocks, using outcrops from the Lower Cretaceous Cupido Formation in the Monterrey Salient of the Sierra Madre Oriental, northeastern Mexico. Field observations and petrographic studies of crosscutting relations and fracture-fill mineralogy and texture distinguish six principal coupled fracturing-cementation events. Two fracture events named F1 and F2 are characterized by synkinematic calcite cement that predates D2 regional dolomitization. A third fracture event (F3) is characterized by synkinematic dolomite fill, contemporaneous with D2 dolomitization of host strata. The fourth event (F4) is characterized by synkinematic D3 baroque dolomite; this event postdates D2. The fifth fracture event (F5) is characterized by C3 synkinematic calcite, and postdates D3 dolomite. Finally, flexural slip faulting (F6) is characterized by C3t calcite, and postdates D3 dolomite. Carbon and oxygen stable isotopes were used to validate the paragenetic sequences proposed for the Cupido Formation rocks. The dolomite isotopic signatures are consistent with increasing precipitation temperatures for the various fracture cements, as is expected if fractures grew during progressive burial conditions. Three main groups of calcite cement can be differentiated isotopically. Late calcite cement may have precipitated from cool waters under shallow burial conditions, possibly during exhumation of the SMO. The development of the Structural Diagenetic Petrographic Study protocol, and its integration with geological, geophysical and engineering data, can be applied to oil fields in fractured carbonates such as those located in Mexico, to validate its applicability.
Investigation of Artificial Neural Networks, Alternating Conditional Expectation, and Bayesian Methods for Reservoir Characterization
(1998-12) Kapur, Loveena; Lake, Larry W.; Sepehrnoori, Kamy
The objective of reservoir characterization is to describe the complex distribution of properties of a reservoir based on available geological, petrophysical, and engineering data. Some of the reasons for the complexity are randomness or nonlinearity among petrophysical parameters and the subjective nature of geological interpretation. The nonlinear relationships among the reservoir properties can be quantified using artificial neural networks (ANN),· alternating conditional expectation (ACE), and Bayesian methods. First,· an approach is developed to correlate oil recovery efficiency with the petrophysical, engineering, and volumetric parameters reported in the Atlas of Major Texas Oil Reservoirs database compiled by the Bureau of Economic Geology at The University of Texas at Austin. Results are obtained by using the alternating conditional expectation (ACE) method on the database and dividing reservoirs according to drive mechanisms and/or reservoir classes. The categorical classification according to drive mechanism gives better predictions than classification by lithologies. This approach can be applied for prediction of oil recovery efficiency in a new reservoir. Second, an approach is developed for facies classification in a reservoir from wireline logs and core data using back-propagation artificial neural networks (BP-ANN) and Bayesian methods. The example facies selected from a sandstone reservoir are turbidities, debris flow, shallow marine, shoreface, and lower shoreface. Core and wireline logs (gamma ray, density, neutron porosity, and resistivity) are used for facies and facies pay prediction. The accuracy of the facies predicted from these methods usually ranges from 75 to 93%. Gamma ray and density logs are the most crucial for some types of facies while neutron porosity logs are most important for others. These results can be used where quantitative classification of a large number of logs by visual observation can be time-consuming and tedious. The approach can also be used to deteimine which logs are the most crucial for determining different types of facies. Third, the Bayesian approach is further extended for the prediction of facies pay and net pay using wireline log ·and core data. The facies pay is predicted based on the results from facies classification using Bayes theorem. The net pay is. predicted by Classifying the core data into permeability classes. Neutron porosity and density logs are usually important for prediction of facies pay. Sonic logs are usually important for net pay prediction.
Modeling and Application of Tracers for Reservoir Characterization
(1994-12) Maroongroge, Vichai; Pope, Gary A.; Sepehrnoori, Kamy
There were two main objectives of this research. The first objective was to develop an accurate and efficient tracer model to study gas tracers used in both single-well and interwell tracer tests. The second objective was to develop an advanced technique for the description of reservoir characteristics using an interwell tracer test. The first objective was achieved by implementing a tracer option in UTCOMP (an equation of state compositional simulator). The code was made efficient by separating the tracer components from the phase equilibria calculations. Data from two interwell perfluorocarbon gas tracer tests at Elk Hills Naval Petroleum Reserve (NPR) were analyzed and 3D simulation studies were conducted using the newly developed tracer features in UTCOMP. A new tracer test for estimating reservoir heterogeneity called vertical tracer profiling (VTP) was developed and tested. This interwell tracer test is done by injecting tracers or collecting tracer samples at different locations along the depth of a reservoir. One of the most important and difficult problems in analyzing interwell tracer data is the solution of the inverse problem, that is, the inference of the reservoir heterogeneity from the tracer production data. A combinatorial optimization method called simulated annealing was used to solve this problem. One of the advantages of simulated annealing is that more than one type of data can be used simultaneously to generate the reservoir description. In this case, both vertical variograms, which can be estimated from core data, and tracer data were used. This approach was applied to both conventional tracer production data and tracer production data using VTP. Results show that the new tracer test can give the same or better reservoir description compared to a conventional tracer test. Since a lot of computational time is required, an alternative procedure that takes much less computational time and yet produced superior results was developed. The permeability field generated from this improved procedure was shown to be effective in matching simulated waterflood data at high endpoint mobility ratios. We next investigated the accuracy of the swept volume calculated from the first-moment method using tracer production data from an interwell tracer test in a heterogeneous reservoir. Results of swept volume calculated from the first-moment method of tracer production data are compared with ones from reservoir tracer concentration contours from simulations. It is. shown that the accuracy of the first moment method depends on the detectable limit or the cutoff concentration of tracer production data. The first-moment method was extended to partitioning tracers to determine the residual oil saturation between wells. This method gives an accurate estimate of the residual oil saturation even for a very heterogeneous reservoir and a nonuniform (local) distribution of residual oil. A field application is illustrated using data from a gas tracer test in the Shallow Oil Zone (SOZ) reservoir of the Elk Hills NPR. The use of the Peng-Robinson equation of state to predict the partition coefficients of perfluorocarbon gas tracers (PFT's) at different conditions was also investigated. The first-moment method was applied to VTP tracer production data to obtain an estimate of the distribution of the residual oil saturation in the reservoir. These results show that VTP data can provide insight into the flow pattern of a reservoir and the accuracy of the estimate depends on the degree of vertical crossflow. The use of selective injection combined with VTP is shown to provide the most information about crossflow and reduce the chance of the tracers bypassing the residual oil.
Modeling complex spatial patterns in reservoir models using high order spectra
(2016-12) Elahi Naraghi, Morteza; Wheeler, Mary F. (Mary Fanett); Srinivasan, Sanjay; Sepehrnoori, Kamy; Sen, Mrinal K; Spikes, Kyle T; Foster, John T
One of the most challenging issues in reservoir modeling The important goal of reservoir modeling is to generate a map of geologic attributes that can yield predictions of hydrocarbon production. Mostly, the primary source of information for creating such a map is borehole measurements, which are only available at sparse locations. Reservoir modeling constrained to the available data along the wells allows us to generate multiple realizations for the whole reservoir. In order for these realizations to yield robust estimates of the uncertainty in reservoir performance prediction, it is imperative that they exhibit connectivity characteristics that are typical for the geological system being modeled. Different algorithms have been developed to stochastically simulate reservoir properties using sparse measured data. In these methods, the spatial variability represented by the underlying joint distribution is in the form of the spatial covariance. The major drawback of traditional variogram-based modeling is that they are not able to reproduce complex spatial patterns. Multiple-point statistical algorithms, however, can reconstruct such curvilinear features. In this study, we study the link between the multiple point spatial pattern connectivity and the Fourier spectrum. This will allow us to infer statistical functions describing reservoir connectivity more efficiently. This can be further sub-divided into two approaches based on the availability of data and information. We also propose methods for selecting an optimum training image when there is ambiguity associated with it, and integrating non-stationary secondary information into the simulation framework. Then, we develop a simulation algorithm in Fourier domain. We will show that the amplitude of Fourier transform can be calculated directly from power spectrum (Fourier transform of covariance function). The phase identification can be achieved by either solving an optimization problem to match the available conditioning data or from higher order spectra such as bispectrum or trispectrum. Finally, we present a new framework for integrating dynamic data and performing history matching. We show how polyspectra affect the production behavior and therefore, we can use the production measurements at well location to identify amplitude and phase within the proposed framework.
Modeling of recovery process characterization using magnetic nanoparticles
(2013-12) Rahmani, Amir Reza; Bryant, Steven L.; Huh, Chun
Stable dispersions of magnetic nanoparticles that are already in use in biomedicine as image-enhancing agents, also have potential use in subsurface applications. Surface-coated nanoparticles are capable of flowing through micron-size pores across long distances in a reservoir with modest retention in rock. Tracing these contrast agents using the current electromagnetic tomography technology could potentially help track the flood-front in waterflood and EOR processes and characterize the reservoir. The electromagnetic (EM) tomography used in the petroleum industry today is based on the difference between the electrical conductivity of reservoir fluids as well as other subsurface entities. The magnetic nanoparticles that are considered in this study, however, change the magnetic permeability of the flooded region, which is a novel application of the existing EM tomography technology. As the first fundamental step, the magnetic permeability change in rock due to injecting magnetic nanoparticles is quantified as a function of particle and reservoir properties. Subsequently, a new formulation is devised to compute the sensitivity of magnetic measurements to magnetic permeability perturbations. The results are then compared with the sensitivity to conductivity perturbations to identify the application space of magnetic contrast agents. Using numerical simulations, the progress of magnetic nanoparticle bank is monitored in the reservoir through time-lapse magnetic tomography measurements that are expected. Initially, simple models for displacement of injection banks are assumed and the level of complexity is gradually increased to incorporate the realities of fluid flow in the reservoir. The fluid-flow behavior of the nanoparticles is dynamically integrated with time-lapse magnetic response. Since the nanoparticles could help illuminate the flow paths, they could be used to indirectly measure reservoir heterogeneities. Therefore, numerous case studies are demonstrated where reservoir heterogeneity could potentially be inferred. Finally, fundamental pore-scale models are developed as a first step towards the multiple fluid phases extension of the EM tomography application. Using magnetic nanoparticles to improve electromagnetic tomography provides several strategic advantages. One key advantage is that the magnetic nanoparticles provide high resolution measurements at very low frequencies where the conductivity contrast is hardly detectable and casing effect is manageable. In addition, the sensitivity of magnetic measurements at the early stages of the flood is significantly improved with magnetic nanoparticles. Moreover, the vertical resolution of magnetic measurements is significantly enhanced with magnetic nanoparticles present in the vicinity of source or receiver. The fact that the progress of the magnetic slug can be detected at very early stages of the flood, that the traveling slug’s vertical boundaries can be identified at low frequencies, that the reservoir heterogeneities could potentially be characterized, and that the magnetic nanoparticles can be sensed much before the actual arrival of the slug at the observer well, provides significant value of using magnetic contrast agents for reservoir illumination.
New method for rock classification in carbonate formations using well-log-based rock fabric quantification
(2017-12) Purba, Sonia Arumdati; Heidari, Zoya
Challenges in rock classification of complex carbonate formations are often rooted in the failure to consider the spatial distribution of pore network or rock fabric. It is crucial to quantify rock fabric through different petrophysical properties that are affected by it. For example, Mercury Injection Capillary Pressure (MICP)-based pore typing uses the estimated pore-size distribution to cluster different pore types. Another example is well-log-based rock quality factor that takes advantage of the mud-filtrate invasion effects on well logs as proxy to the pore-size distribution. However, such methods require prior knowledge of the number of rock classes, which is mainly rooted in the differences between petrophysical rock classes and geological depositional facies. This thesis introduces a method for improving rock classification by determining the optimum number of rock classes, evaluating and quantifying pore network connectivity and geometry as rock fabric, as well as enhancing petrophysical evaluation. The proposed method involves an iterative procedure that starts with conventional well-log interpretation to obtain the petrophysical properties, such as volumetric concentration of shale, porosity, and permeability. Rock classification is performed using an unsupervised neural network with an initial assumption of the number of rock classes and well-log-based estimates of petrophysical and compositional properties as inputs. Permeability models are then developed in the pore-scale domain using core subsamples at different depths of interest. Models describing the correlation between electrical resistivity, conducting or effective porosity, and permeability are established in the pore-scale domain. These models are then applied in the log-scale domain to improve permeability estimates. Next, rock classification will be performed using the improved permeability estimates as inputs and updating number of rock classes. This process is repeated until a convergence in permeability estimates is achieved. outcomes of the iterative method include log-scale rock classification, permeability estimates, and the optimum number of rock classes. The method introduced in this thesis was successfully applied to two wells in a carbonate formation. outcomes of rock classification are in good agreement with the geological depositional facies. The iterative method results in 75% improvement in permeability estimates in the log-scale domain when compared against those obtained from conventional porosity-permeability correlations. Furthermore, this method effectively optimizes the number of rock classes, making it a promising approach for field cases with limited core measurements and no prior knowledge of rock types
Novel stochastic inversion methods and workflow for reservoir characterization and monitoring
(2013-12) Xue, Yang, active 2013; Sen, Mrinal K.
Reservoir models are generally constructed from seismic, well logs and other related datasets using inversion methods and geostatistics. It has already been recognized by the geoscientists that such a process is prone to non-uniqueness. Practical methods for estimation of uncertainty still remain elusive. In my dissertation, I propose two new methods to estimate uncertainty in reservoir models from seismic, well logs and well production data. The first part of my research is aimed at estimating reservoir impedance models and their uncertainties from seismic data and well logs. This constitutes an inverse problem, and we recognize that multiple models can fit the measurements. A deterministic inversion based on minimization of the error between the observation and forward modeling only provides one of the best-fit models, which is usually band-limited. A complete solution should include both models and their uncertainties, which requires drawing samples from the posterior distribution. A global optimization method called very fast simulated annealing (VFSA) is commonly used to approximate posterior distribution with fast convergence. Here I address some of the limitations of VFSA by developing a new stochastic inference method, named Greedy Annealed Importance Sampling (GAIS). GAIS combines VFSA with greedy importance sampling (GIS), which uses a greedy search in the important regions located by VFSA to attain fast convergence and provide unbiased estimation. I demonstrate the performance of GAIS on post- and pre-stack data from real fields to estimate impedance models. The results indicate that GAIS can estimate both the expectation value and the uncertainties more accurately than using VFSA alone. Furthermore, principal component analysis (PCA) as an efficient parameterization method is employed together with GAIS to improve lateral continuity by simultaneous inversion of all traces. The second part of my research involves estimation of reservoir permeability models and their uncertainties using quantitative joint inversion of dynamic measurements, including synthetic production data and time-lapse seismic related data. Impacts from different objective functions or different data sets on the model uncertainty and model predictability are investigated as well. The results demonstrate that joint inversion of production data and time-lapse seismic related data (water saturation maps here) reduces model uncertainty, improves model predictability and shows superior performance than inversion using one type of data alone.
Pre-injection reservoir characterization for CO₂ storage in the inner continental shelf of the Texas Gulf of Mexico
(2017-05) Sabbagh, Reinaldo Jose; Meckel, Timothy Ashworth
The injection of CO₂ into the subsurface (carbon capture and storage; CCS) is the most viable approach to significantly reduce industrial emissions of greenhouse gasses to the atmosphere. The inner continental shelf of the northern Gulf of Mexico has incredible potential for CO₂ storage. This study quantitatively evaluates the CO₂ storage capacity of the Lower Miocene brine-filled sandstones in the inner continental shelf of the Texas Gulf of Mexico using 3D seismic and well log data. The first part of this work investigates the relationship between elastic properties and reservoir properties (e.g., porosity, mineralogy, and pore fluid) of the Lower Miocene section using rock physics modeling and simultaneous seismic inversion. The elastic properties are related to porosity, mineralogy and pore fluid using rock physics models. These rock physics transforms are then applied to the seismically derived elastic properties to estimate the porosity and lithology away from the wells. The porosity and lithology distribution derived using this quantitative method can be interpreted to predict the best areas for CO₂ storage in the inner continental shelf of the Texas Gulf of Mexico. The second part of this work studies the effect that CO₂ has on the elastic properties of the Lower Miocene rocks using fluid substitution, amplitude variation with angle (AVA), and statistical classification to determine the ability of the seismic method to successfully monitor CO₂ injected into the subsurface. The velocities and density well logs were modeled with different fluid saturations. To characterize the seismic properties corresponding to these different fluid saturations, the AVA responses and probability density functions were calculated and used for statistical classification. The AVA modeling shows a high sensitivity to CO₂ due to the soft clastic framework of the Lower Miocene sandstones. The statistical classification successfully discriminates between brine and CO₂ saturation using Vp/Vs and P-impedance. These results shows that the Lower Miocene sandstones have the capacity to host CO₂, and that the CO₂ injected in these rocks is likely to be successfully monitored using seismic methods.