The simulation of fluid flow in fractured reservoirs is mostly based on the Warren-Root formulation in which the matrix is dissected into blocks by fractures(1). In modern simulators, the Warren-Root formulation has been extended to account for fluid flow in the matrix as well as in the fractures(2). These two media of contrasting transmissibility and storage capacity constitute the basis of dual porosity dual permeability modeling. Hydraulic connection between the two media is managed via a transfer function. This formulation is appropriate for cases with diffuse interconnected fractures. Alternatively, fractures can be modeled explicitly by capturing their geometries, shapes and sizes in the form of numerical grid cells. But this would be prohibitively costly in terms of run time and memory. This paper presents a method that is suitable for giant fields that are dominated by clusters of sub-vertical fractures called fracture corridors. It is based on a hybrid approach of the two alternatives mentioned. The effective Warren-Root and fracture parameters are adjusted to mimic explicit fracture modeling thereby capturing the advantages of both. The method was applied to a giant carbonate field in Saudi Arabia, which has both fracture corridors and super-permeable bodies. These bodies are typically horizontal and they inter-connect with the fracture corridors to form a combined high conductivity medium which is responsible for the unusual water movement observed in some parts of the field. The full field simulation model contains a homogeneous matrix and a fracture grid comprised of fracture corridors and super-permeable bodies. Fluid segregation is assumed in the fracture system in agreement with the physics inside conductive vertical fractures. The matrix-fracture corridor and the matrix-super permeable body exchanges are represented in a manner similar to matrix-fracture transfers in the Warren-Root dual-porosity system. Special attention was paid to matrix block size and to imbibition capillary pressure. This approach led to a reliable history match that captures the water arrival time and water production profiles in the reservoir. Introduction Fluid flow in naturally fractured reservoirs primarily takes place via high permeability and low porosity fractures surrounding matrix blocks. The simulation of such reservoirs is challenging both in terms of characterization and numerical modeling. This paper addresses some challenging aspects of numerically modeling a special kind of fractured reservoir where fractures cluster together to form the so called fracture corridors, which are also known as fracture fairways or fracture swarms. Unlike diffuse fracture networks whereby fractures are distributed within reservoir matrix rock, reservoirs with fracture corridors are not readily approximated by the Warren-Root, so-called "sugar-cube" model. This is because fracture corridors (a) are large scale features that generally cut through the reservoir thickness, (b) exist along corridors while the vast part of the reservoir may or may not be free of fractures. Such large scale but clustered fractures are akin to major faults in some ways and could be modeled explicitly by approximating them as such. At the same time, we know they actually consist of a large number of individual and relatively small high conductivity openings that are clustered and aligned together. It is possible to model such configurations by using a modified set of Warren-Root parameters. In this paper, they are in effect approximated by the judicious choice of Warren-Root parameters as in dual media modeling so as to respect the fact that they are akin to major fault lines. This approach is applied in the Arab-D full field modeling. The Arab D reservoir is carbonate and oil bearing. The model is constructed using Saudi Aramco's POWERSTM simulator and contains a total of about 9 million cells.

This paper presents an innovative and promising, multi-discipline integrated approach that includes geology (BHI, cores, wireline logs), geophysics (seismic facies analysis), and reservoir engineering data (production data, PLT, welltest) that were combined to identify the main types of fractures, to predict their occurrence in the reservoir and to determine the hydraulic properties of the different fractures sets The Najmah - Sargelu of West Kuwait is an oil bearing reservoir made of tight carbonates where porosity and permeability is mainly provided by the fracture network. In this paper, we will first introduce the method used to identify and predict the two main scales of fractures: joints and large-scale fractures (faults and fracture swarms). The shale content (Vshale) and mechanical beds thickness were found to be the two main geological drivers on joints occurrence. Thickness of individual beds were recorded from BHI acoustic images which enabled to measure an S/T ratio (fracture spacing to bed thickness) for each fracture set and for different shalyness. Secondly, we used an innovative solution to deliver an accurate map of large-scale fractures location. This approach uses concurrently a set of selected fracture relevant attributes in a multi-variable statistical process called Seismic Facies Analysis (SFA). A 3D stochastic fracture model was then generated incorporating the two scales of fractures and constrained by the reservoir shalyness, the S/T ratio and the seismic facies map. The calibration of the hydraulic properties of the fractures was achieved through the second innovation presented in this paper: the simulation of a synthetic well test using the 3D fracture model and matched with the real data. This resulted in the calibration of the hydraulic fractures conductivity for each fracture type. These values were combined with the 3D stochastic fracture model to produce 3D fracture properties models (porosity, permeabilities and block size) for the Najmah - Sargelu of West Kuwait. Introduction A detailed geological and hydraulic characterisation of the fracture network occurring within the Upper Jurassic Najmah - Sargelu reservoir of West Kuwait was planned in 2003–2004 by Kuwait Oil Company (KSC). The objective of the study was to identify the main geological drivers on natural fractures occurence, to measure their hydraulic properties and eventually using discrete fracture modeling (DFN) approach to compute the equivalent fracture properties (porosity, permeability and block sizes) required for the reservoir simulation. This was achieved through a close integration of geological, geophysical, petrophysical and dynamic data carried out using workflows and methods implemented in a fracture analysis and modeling software (see Ref. 1). The main tasks performed during the project and presented in this paper are the following:Fracture analysis from coresFracture analysis from BHI logsIntegration of 3D seismic data set3D fracture modelingHydraulic characterization of the fracture networkComputation of the fracture properties in the reservoir grids Background The study area is approx. 2000 Km[2] and covers four fields namely A, B, C and D from North to South, (Fig. 1). The structure of the reservoir is characterized by gentle, rather elongated anticlines plunging mainly in the NNE and SSW directions at A, B and D fields. Field C and the west branch of field D are NNW - SSE oriented. The Top reservoir depth ranges between 11,000ft to 12,000ft.

This paper presents an innovative and promising, multidisciplinary approach that includes geology (borehole images, cores, and wireline logs); geophysics (seismic facies analysis), and reservoir engineering data (production data, production logs, and well test) that were combined to identify the main types of fractures, to predict their occurrence in the reservoir, and to determine the hydraulic properties of the different fractures sets.The Najmah-Sargelu of west Kuwait is an oil-bearing reservoir made of tight carbonates where porosity and permeability are provided mainly by the fracture network. In this paper, we first introduce the method used to identify and predict the two main scales of fractures: joints and large-scale fractures (faults and fracture swarms). The shale content (V shale ) and mechanical-beds thickness were found to be the two main geological drivers on joints occurrence. The thickness of individual beds was recorded from borehole acoustic images, which enabled us to measure a fracture spacing/bed thickness (S/T) ratio for each fracture set and for different shaliness. Second, we used an innovative solution to deliver an accurate map of the location of large-scale fractures. This approach concurrently uses a set of relevant attributes per selected fracture in a multivariable statistical process called seismic facies analysis (SFA).A 3D-stochastic fracture model was then generated, incorporating the two scales of fractures and this model was constrained by the shaliness of the reservoir, the S/T ratio, and the seismic facies map. In this approach, the two scales of fractures are modeled independently. The model of large-scale fractures is conditioned by the picking of lineaments on the SFA map and validated at wells, whereas small-scale fractures are modeled according to geological driven statistics on fracture density and fracture orientation. The calibration of the hydraulic properties of the fractures was achieved through the second innovation presented in this paper: the simulation of a synthetic well test using the 3D-fracture model and matched with the real data. This resulted in the calibration of effective hydraulic conductivity for each fracture type. These values were combined with the 3D-stochastic fracture model to produce 3D-fracture-properties models (porosity, permeability, and block size) for the Najmah-Sargelu.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractThe simulation of fluid flow in fractured reservoirs is mostly based on the Warren-Root formulation in which the matrix is dissected into blocks by fractures (1) . In modern simulators, the Warren-Root formulation has been extended to account for fluid flow in the matrix as well as in the fractures (2) . These two media of contrasting transmissibility and storage capacity constitute the basis of dual porosity dual permeability modeling. Hydraulic connection between the two media is managed via a transfer function. This formulation is appropriate for cases with diffuse interconnected fractures. Alternatively, fractures can be modeled explicitly by capturing their geometries, shapes and sizes in the form of numerical grid cells. But this would be prohibitively costly in terms of run time and memory. This paper presents a method that is suitable for giant fields that are dominated by clusters of sub-vertical fractures called fracture corridors. It is based on a hybrid approach of the two alternatives mentioned. The effective Warren-Root and fracture parameters are adjusted to mimic explicit fracture modeling thereby capturing the advantages of both. The method was applied to a giant carbonate field in Saudi Arabia, which has both fracture corridors and super-permeable bodies. These bodies are typically horizontal and they interconnect with the fracture corridors to form a combined high conductivity medium which is responsible for the unusual water movement observed in some parts of the field. The full field simulation model contains a homogeneous matrix and a fracture grid comprised of fracture corridors and super-permeable bodies. Fluid segregation is assumed in the fracture system in agreement with the physics inside conductive vertical fractures. The matrix-fracture corridor and the matrix-super permeable body exchanges are represented in a manner similar to matrix-fracture transfers in the Warren-Root dual-porosity system. Special attention was paid to matrix block size and to imbibition capillary pressure. This approach led to a reliable history match that captures the water arrival time and water production profiles in the reservoir.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractThis paper presents an innovative and promising, multidiscipline integrated approach that includes geology (BHI, cores, wireline logs), geophysics (seismic facies analysis), and reservoir engineering data (production data, PLT, welltest) that were combined to identify the main types of fractures, to predict their occurrence in the reservoir and to determine the hydraulic properties of the different fractures setsThe Najmah -Sargelu of West Kuwait is an oil bearing reservoir made of tight carbonates where porosity and permeability is mainly provided by the fracture network. In this paper, we will first introduce the method used to identify and predict the two main scales of fractures: joints and largescale fractures (faults and fracture swarms). The shale content (Vshale) and mechanical beds thickness were found to be the two main geological drivers on joints occurrence. Thickness of individual beds were recorded from BHI acoustic images which enabled to measure an S/T ratio (fracture spacing to bed thickness) for each fracture set and for different shalyness. Secondly, we used an innovative solution to deliver an accurate map of large-scale fractures location. This approach uses concurrently a set of selected fracture relevant attributes in a multi-variable statistical process called Seismic Facies Analysis (SFA).A 3D stochastic fracture model was then generated incorporating the two scales of fractures and constrained by the reservoir shalyness, the S/T ratio and the seismic facies map. The calibration of the hydraulic properties of the fractures was achieved through the second innovation presented in this paper: the simulation of a synthetic well test using the 3D fracture model and matched with the real data. This resulted in the calibration of the hydraulic fractures conductivity for each fracture type. These values were combined with the 3D stochastic fracture model to produce 3D fracture properties models (porosity, permeabilities and block size) for the Najmah -Sargelu of West Kuwait.