Facies characterization and classification play critical roles in reservoir description and modeling, facilitating the prediction of 3D reservoir extent and rock properties distribution. The term facies originates from early sedimentologic work and describes rock types with shared traits. The

Understanding the linkages between grain mineralogy and diagenetic and sedimentary processes enhances the reliability of petrophysical models to predict reservoir deliverability from permeability. Petrographic data within well-defined depositional facies reveal the diagenetic evolution of porosity-permeability relationships. Formation evaluation methods relying solely on petrophysical rock typing are seriously limited when predicting ultimate reservoir performance in complex pore structures. The Almond Formation, Wyoming, is characterized by three depositional facies associations — shoreface, deltaic (bay head and flood tide), and fluvial-coastal plain — which present three distinctive porosity-permeability trends. Textural features resulting from depositional processes, such as grain size and sorting, vary little between facies associations, yet permeability can vary by up to four orders of magnitude for the same porosity value. Differences between petrophysical facies are primarily driven by diagenetic (cementation and grain dissolution) effects on different framework grain compositions (petrographic facies). Therefore, the main difference between the facies associations is diagenetic, due to provenance and transport mechanisms. The characterization of depositional and diagenetic controls on pore geometry allows the narrowing of uncertainty in absolute permeability prediction. We have quantified the relationship between depositional facies, with their specific mineral composition and diagenetic overprint, and the steepness functions in porosity-permeability space. This analysis allowed us to effectively reduce the uncertainty in the prediction of initial gas production from wireline logs.

Tight-gas reservoirs undergo unique and often complex burial, diagenetic, structural, fluid pressure and saturation histories. Porosity alteration from compaction, cementation and grain leaching can continue after hydrocarbon charge, further complicating saturation modeling. Many reservoirs have gone through multiple cycles of drainage and imbibition, often at different stages on the diagenetic pathway to current pore-scale morphologies. The understanding of saturation distribution and state is not only desired but required for predicting reservoir performance, estimating realistic recoverable volumes, and optimizing costs for development and production.The Almond Formation is characterized by three depositional facies associations: shoreface, deltaic and fluvial-coastal plain. These groups are commonly fine grained and well sorted. The differences in pore architecture arise from differences in primary depositional fabric and rock-frame mineralogy and their subsequent diagenetic alteration; yielding predictive trends in porosity permeability space.Drainage and imbibition saturation-height models have been developed from core studies and integrated with logs to verify that reservoirs are at primary drainage and to highlight any potential imbibition due to trap tilting or leaking. Centrifuge and multicycle mercury injection data were integrated to produce composite drainage capillary pressure curves. Stressed mercury extrusion tests are commonly used for modeling water saturation through the imbibition process. These tests display no correlation with rock quality at low capillary pressures. To circumvent these problems, mercury extrusion was integrated with maximum-trapped-gas measurements obtained by countercurrent imbibition experiments.Using the resistivity-derived water saturation model as reference, the free-water level for drainage and imbibition models was optimized by matching saturation height models in reservoirs free of resistivity shoulder bed effects. The accuracy of the match in different rock qualities provided insights on the likely saturation state of reservoirs. Such observations were used to develop successful interpretations of the special distribution of free-water level, reservoir architecture, and hydrocarbon charge.

Abstract Saturation Height Functions (SHF) have a key role in reservoir description and in quantifying oil in place. The function(s) must compare well with other sources of water saturation (Sw) when available, such as core measurements and well-log interpretations. We have reconciled the different Sw sources through a reliable SHF function based on the Brooks-Corey model with parameters optimised for Thunder Horse (TH) Field. The function is used to populate oil volumes in the 3-D static and dynamic models. Thunder Horse is one of the largest deep-water fields in Gulf of Mexico and is operated by BP, with ExxonMobil as a co-owner. TH is divided in two trapping structures: TH North (THN) and TH South (THS). The field has multiple Miocene turbidite reservoirs commonly grouped in units known as the Pink, Brown, and Peach. Thunder Horse is fortunate to have Dean-Stark (D-S) and well-log data sets that detail the Sw changes at and above the oil-water contact (OWC). We developed a continuous log-derived Sw after appropriate model calibration with D-S measurements. Subsequently, we applied the Brooks-Corey SHF model with parameters defined for each rock type in TH area based on an iterative regression to the log-interpreted Sw following a workflow we developed. This case study illustrates methods and results of reconciling saturations determined from a number of techniques and validation of the outcomes with independent borehole and direct measurements from core. Two cases where differences were observed are addressed by combining resistivity modeling and borehole images with core-log integration. A comparison of binary vs. continuous net-to-gross (NTG) on Hydrocarbon Pore Volume, based on SHF and log-derived Sw, proved the volume impact is not significant for the highest NTG TH Pink reservoirs. However, the deliverable provides significant value through a consistent and defendable linked saturation and net-to-gross suitable for use in reservoir modelling.

Formation evaluation in thinly bedded sands has always been a challenge especially as individual sand beds thin below the resolution of logging tool measurements. Evaluation of these reservoirs is often further complicated since the various log measurements have differing vertical resolutions and depths of investigations and the wells normally have somewhat disparate log datasets.In this study of a field located in the Gulf of Mexico (GoM) deep water we demonstrate how a consistent and crossverified approach was taken to evaluate thinly bedded sands, which resulted in increasing interpretation confidence and a more realistic net pay when compared with prior techniques that use conventional log interpretations.The field used in this study has a total of five wells, including one well with whole core, three wells with nuclear magnetic resonance (NMR) logs, including the cored well, and two wells with standard log suites. The methodology applied begins with creating a continuous sand-count curve from the visual evaluation of whole-core photographs. This sand-count curve is then compared with the net-to-gross (NTG) curve obtained from a Thomas-Stieber analysis and then crossverified via NTG obtained from a NMR log. The calibration from the three wells with whole core or NMR was applied to the Thomas-Stieber analysis performed on the two wells with standard suites of logs. Whole-core porosities and water saturations are then used to constrain the log-derived porosity and water saturation values. Finally saturation-height modeling has been perfoimed using porous-plate capillary pressure data and compared to log-derived water saturations.

History matching and appraisal of reservoirs usually include a partitioning of the reservoir for each physical property that is to be sensitised. Different engineers or geo-scientists often regionalise the reservoir differently. However, the choice of regionalisation can significantly affect critical results such as reserves estimates. Two auto-regionalisation methods were developed. The first method utilises small scale dimensional models to generate new property arrays compatible with those from the original geo-modelling. This approach also provides the option to regionalise using statistical fitting algorithms and the method of Principal Variables. The second method employs user-tailored distance measures to generate regions automatically. In trials the results were superior to the results obtained by conventional methods. Fewer runs were required to obtain a history match, with the second method performing the best.

Abstract Modeling seismic responses in reservoir rock requires the accurate determination of fluid density, modulus or its reciprocal, the adiabatic compressibility and acoustic velocity. The geophysics community has developed methods to estimate these properties for subsurface oil and gas accumulations at reservoir conditions. Similarly, the engineering community has developed methods to estimate density and isothermal compressibility. The goal of both groups is the accurate characterization of these PVT properties at elevated pressure and temperature especially at the extreme conditions encountered in frontier reservoir environments and newer discoveries and for the quantitative interpretation of 4D seismic. A large and diverse database consisting of 1099 worldwide oil PVT reports with 11,960 density measurements was used to evaluate the various fluid property correlations. The database encompasses oil with API gravity ranging from 10.6-63 along with solution gas-oil ratios of 5-4,631 SCF/STB resulting in saturation pressures ranging 60-10,326 psia at temperatures over the range 50–332 °F. The PVT reports provide a measure of isothermal compressibility; therefore, heat capacity ratio must also be addressed in the comparison with adiabatic compressibility methods. In this work, gas property correlations were tested against methane PVT data derived from the National Institute of Standards and Technology (NIST) which ensured consistent results over a wide range of pressure and temperature conditions. The principle of corresponding states can then be used to examine the suitability of using these correlations for gas-condensate systems. Results of the study detailing the accuracy of the oil and gas PVT correlations are presented.

Abstract This paper describes an integrated workflow that determines the minimum number of petro elastic models required to adequately predict elastic properties for a given flow scale. The workflow involves the formulation of scale-dependent petro-elastic models (PEM) and petro-elastic facies from well logs. Then through an error evaluation technique, determines the minimum number of PEMs required and the maximum vertical grid size limit for the flow scale. The validity of the fine scale PEMs across different scales is also tested in the workflow. Elastic properties of reservoir rocks can be predicted through petro-elastic models which relate fluid and intrinsic reservoir properties to elastic properties through mathematical functions or rock physics models. Although we would like to generate PEMs at flow scale, very often the PEMs are generated from logs that are consistent with core data. A PEM is lithology and rock fabric specific. In an upscaled flow model, the lithologies or textures are mixed. Fine scale PEMs may no longer adequately describe the elastic properties of the reservoir at flow and seismic scale. Prediction is improved in a flow model that integrates all available measurements. Scale inconsistencies of the measurements need to be addressed for better integration and prediction. In this paper we will show that through a unique and iterative approach, PEMs must be generated at flow scale for improved integration. The methodology has been tested with synthetic data and resulted in a better predictive flow scale model, which is consistent across all scale resolutions.

Abstract The second Archie equation relates water saturation to formation resistivity index through a power function. This important relationship has been widely used to evaluate hydrocarbon saturation. However, many rocks don't obey this empirical rule. The best Archie fit to these data may not comply with physical bounds and create significant bias in the computed hydrocarbon saturation. This paper develops a new method for water saturation estimation based on the equivalent rock element model (EREM) that has been demonstrated to work well for rock transport properties to the first order effect. EREM contains two orthogonal pore components, one parallel and the other perpendicular to the direction of transport. Pore structure efficiency is defined as the volume ratio between the two components. When pores contain conductive and non-conductive fluid phases, the pore structure efficiency for the conductive phase changes with saturation due to fluid and mineral properties and pore structure variations. The conductive phase may become discontinuous below a certain critical saturation due to pore structure, wetting characteristics, interfacial tension, etc. The proposed algorithm incorporates this critical saturation phenomenon and relates capillary pressure data with electrical measurements. Approximating the pore structure efficiency of conductive phase as a function of saturation, we formulated a theoretical relationship between resistivity index and water saturation. For some rocks, additional conductive meachanisms may exist in addition to electrolytic conductivity, in which case the Archie's equation would not apply. Many proposed formulas can be found in the literature regarding non-Archie rocks, but each only works well for a particular type of rocks. This paper demonstrates that the EREM approach can account for the additional conductivity and work well regardless of rock type. This new approach expands the applications of the equivalent rock element model. It has several advantages over existing methods. (1) The approach is based on a simple physical model that reflects the two main components in a pore structure and accounts for the first order transfer effect. This ensures compliance with physical boundary conditions and increases the predictiveness. (2) The innovative inclusion of critical water saturation avoids the underestimation of water saturation at low water saturations. It links electrical measurements with capillary pressure measurements and allows the two observations to cross-validate and complement each other. (3) The proposed method is demonstrated to match core measurements from different rock and pore types with a single model.

In this paper, we present a method for the direct stratigraphic inversion of lithology from pre-stack seismic data. This approach is a generalization to discrete variables of a methodology for petrophysical seismic inversion of continuous attributes.

Abstract We introduce a Stratigraphic inversion method that simultaneously integrates pre-stack seismic data with petrophysical and geological data. We use simulated annealing to invert directly for reservoir properties such as porosity, lithology and fluid content in a 3D geocellular model. Well and seismic data are integrated in their respective domains along with physical constraints at different vertical scales to produce an optimal solution. Application of user-defined Petro-Elastic Models (PEM) is a key element of the proposed methodology. In addition to connecting the inverted properties to the seismic response, the PEMs are used to maintain consistency between the time, depth and derived velocities throughout the inversion process. The proposed methodology overcomes the limitations faced by many existing techniques with regards to vertical resolution, time-to-depth conversion and the link between seismic response and reservoir properties. The result of our petrophysical seismic inversion is a fine-scale shared earth model in depth that is consistent with both log and seismic data and can be used for reservoir performance prediction. After demonstrating the robustness of the method on synthetic data, we present a result from a real dataset. The proposed methodology has been successfully applied to porosity inversion on one of the largest undeveloped oil fields in the North Sea. A fine-scale reservoir model has been obtained which reveals previously undetected geological structures and leads to a better understanding of the reservoir zone.

H039 RELATING ELASTIC AND ELECTRICAL PROPERTIES VIA AN EQUIVALENT ROCK ELEMENT MODEL Abstract 1 This paper proposes an equivalent rock element model to correlate elastic properties with electrical properties. The pore structure of this model is comprised of two orthogonal components that capture the first order pore structure effects. Given the measured porosity bulk and shear moduli we can establish the corresponding equivalent rock element model and consequently compute its formation resistivity factor. Our real data example shows that the correlation between porosity and bulk or shear modulus alone can be very poor but the formation resistivity factor derived from

A015 A CASE STUDY OF INTEGRATED RESERVOIR CHARACTERIZATION AND FINE-SCALE SIMULATION 1 B. Z. SHANG H. YANG J. G. HAMMAN D. H. CALDWELL J. MILLIKEN Marathon Oil Company 5555 San Felipe St. Houston TX 77056 Abstract This paper presents an integrated reservoir characterization and simulation study which applies a combination of deterministic and stochastic approaches depending on the amount of data and degree of understanding. The workflow incorporates all available data types like 3D seismic data well logs geologic data core analysis well tests and PVT analysis. Seismic data is first inverted into acoustic impedance which is again inverted into

The second Archie equation relates water saturation to formation resistivity index by a power function. This important relationship has been widely used to evaluate hydrocarbon saturation. However, many rocks don't obey this empirical rule. The best Archie fit to these data may not comply with physical bounds and create significant bias in computed hydrocarbon saturation.This paper develops a new method for water saturation estimation based on the equivalent rock element model (EREM) that has been demonstrated to work well for rock transport properties to the first order effect. EREM contains two orthogonal pore components, one parallel and the other perpendicular to the direction of transport. Pore structure efficiency is defined as the volume ratio between the two components. When pores contain conductive and non-conductive fluid phases, the pore structure efficiency for the conductive phase changes with saturation due to fluid and mineral properties and pore structure variations. Approximating conductive phase pore structure efficiency as a function of saturation, we formulated a theoretical relationship between resistivity index and water saturation. The conductive phase may become discontinuous below a certain critical saturation due to pore structure, wetting characteristics, interfacial tension or a combination of different factors. The proposed algorithm readily incorporates this critical saturation phenomenon and explains the empirical relationships between resistivity index and water saturation for both Archie and non-Archie rocks. It fits core measurements better than Archie's second equation as demonstrated in our real examples and provides more accurate estimation for water saturation. This new approach expands the applications of the equivalent rock element model.

B027 IOI – A METHOD FOR FINE-SCALE QUANTITATIVE DESCRIPTION OF RESERVOIR PROPERTIES FROM SEISMIC 1 Abstract Inversion of inversion IOI is a method for predicting fine scale reservoir properties relevant to flow through the integration of seismic petrophysical and geological models. The technique uses constrained deterministic physical relationships of rock and fluid compressibility to predict porosity and fluid saturation. When aggregated and forward modeled these reservoir properties will reproduce the original experimental observations found in the well logs and seismic surveys. The reservoir scale layers are adjusted and combined in a defined depositional stacking sequence to produce an effective medium

A novel method for the integration of multi-scale rock and fluid information was performed on a single well deep-water exploration discovery. The method integrates diverse data types through physically constrained models into a single earth model. Development scenarios were optimized through multiple flow simulations. The development scenarios were designed to investigate the range of uncertainty in the earth model. This technique utilizes a single integrated earth model and multiple simulations in contrast to techniques that cannot reproduce the seismic, or reproduce the seismic but require a flow simulation for each geologic "realization". The pending business decisions required reservoir characterization with the greatest precision possible and an understanding of the inherent uncertainties. The discovery was made based on 3-D seismic amplitudes, but a portion of the potential reservoir appeared to be obscured by a shallower seismic event. The available borehole data with which to construct the earth model consisted of conventional wireline logs and a wireline-conveyed formation tester from a single penetration. No flow tests, core or fluid samples were available. The technique honors all the observed data through the alteration of an initial geologic model until, through up-scaling, an acceptable match is made with acoustic impedance determined from the observed seismic. The link between the geology and seismic is a rock physics model consistent with the well data. The rock physics model is used to estimate the equivalent seismic properties from the porosity and fluid saturations in the fine scale geologic model. The objective is to minimize the difference between acoustic impedance generated from the 3-D seismic data and the forward-modeled very-fine scale geologic model. The uncertainties associated with each measurement and derived variables were assessed at each step. The resulting model will reproduce the original seismic amplitude data and all of the well data, although the well data were not explicitly used in the seismic inversion or the "inversion-of-the-inversion" estimation of petrophysical properties. Each cell in the resulting fine-scale geologic model contains porosity, permeability, and oil saturation. The static model was up-scaled for dynamic flow simulation to estimate productivity and evaluate various development schemes. Numerous simulations were performed to address uncertainties in the geologic model and to account for risks associated with the geophysical imaging problems, and flow capacity. Two sets of distributions were assigned to porosity and permeability in the simulator for separate areas of the reservoir. A distribution was also assigned to fault transmissibility. The results of hundreds of simulator realizations were then used to estimate the range of uncertainty in reservoir performance under varied development scenarios.