Understanding Naturally Fractured Reservoirs

A recent Applied Technology Workshop (ATW) held in Houston, “Naturally Fractured Reservoirs,” explored reservoir characterization, fracture-network modeling and reservoir simulation, and experience with improved-oil-recovery (IOR) and enhanced-oil-recovery (EOR) processes in naturally fractured reservoirs (NFRs). The ATW featured a technical agenda divided into four sessions and included more than 20 presentations. The four main sessions were:

NFR Characterization (cochaired by El Paso Energy Petroleum Engineer Enrique Peralta and Texas A&M U. Assoc. Professor David Schechter).
Constructing NFR Geologic and Flow Models (cochaired by Ecopetrol Chief Engineer and Vice President–Exploration Andreas Mantilla and Repsol Reservoir Engineer Hugo Araujo).
Predicting NFR Performance (cochaired by Texas A&M U. Professor Thomas Blasingame and Colorado School of Mines Research Professor Hossein Kazemi).
Improved and Enhanced Oil and Gas Recovery in NFRs (cochaired by Kazemi and Schechter).

NFR Characterization

Heterogeneity: Cause and Effect—The presentation by Chevron Energy Technology Center Structural Geologist Wayne Narr, titled “Guilt and Absolution in NFR Characterization: The Cause and Devilish Effect of Heterogeneity,” examined the concept of representative elemental volume, illustration of fracture-network intersection with the well, and evaluation interval of the fracture density. Outcrop analogs necessary for qualifying heterogeneity and modeling of fractured-formation systems also were addressed. The presentation concluded with discussion among workshop participants on fracture allocations for simulation, the combination of primary and secondary porosity in simulation, and numerical modeling.

Dissimilar Fracturing—Sandia Natl. Laboratory Geologist John Lorenz presented “Difference Between Fractures in Limestones and Sandstones,” in which observations regarding lithology were made that supported the conclusion that production-induced pressure (stress) changes can destroy permeability. Regarding fracture creation, discussion centered around two schools of thought: the idea of more fractures creating less matrix permeability and whether extension fractures are conductive.

Seismic for Reservoir Characterization—Prism Seismic Director of Consulting Abdel Zellou, in “Seismically Driven Fractured Reservoir Characterization,” illustrated high-resolution inversion (elastic and post-stack) and other processes that can provide a basis for modeling, the use of seismic attributes (e.g., impedance) to derive geologic models (e.g., porosity) constrained by the seismic data, and fracture drivers (seismic, geological, and geomechanical) in 3D. One topic of discussion was drilling recommendations based on seismic attributes.

La Concepción—Denis Marchal, Petrobras Technical Advisor in Structural Geology, delivered a presentation on “Reservoir Characterization and Tailored Technology Application in the Mature Cretaceous NFR of La Concepción Field, Venezuela.” Presenting the key operational strategies used in redevelopment of the mature Cretaceous NFR in this field, Marchal emphasized the successful application of selective technologies and methodologies and the need to fully integrate the structural model with the dynamic one.

Geologic and Flow Models

Permeability Estimation—U. of Texas at Austin Assistant Professor Jon Olson, in “Permeability Estimation in Fractured Reservoirs: A Fracture Mechanics Approach Incorporating Diagenetic Effects,” emphasized interwell fracture-pattern attributes, industry-standard models, and constrained modeling. Participants discussed the origin of strain rates, whether to include pore pressure as a nonlinearity, and the creation of extension fractures.

Role of Outcrop Analogs—Juan-Mauricio Florez, iReservoir Geoscientist, offered an overview of creating equally probable geologic models and the reasoning behind the use of outcrops, and he presented illustrative outcrop cases that indicate hierarchical shearing and progressive deformation, spacing and dimension of fracture hierarchies linked to stratigraphy, and antithetic conjugate faults creating fracture swarms.

Integration of Conventional, Image, and Production Logs—Mantilla presented a case study in which reproduction of dynamic behavior (identification of fracture patterns and principal stress) was reviewed. Discussion centered around fracture indicators other than logs, fractures at the pore scale, and whether fracture damage during drilling is permanent.

Discrete Fracture-Network Models—Araujo presented the rationale of dynamic behavior of discrete fracture-network (DFN) models based on geological reservoir characterization as the first step in constructing a credible reservoir flow model, classical deterministic methods, and/or quantitative stochastic approaches. Considerable emphasis was given to well-test analysis (fracture network and non-Darcy-flow effects). Conclusions reached included NFRs being adequately represented using a matrix system and a DFN model with a fine-grid dual-porosity model, fracture properties needing to be calibrated with transient well-test data, and dynamic upscaling as the best approach to upscale fracture properties.

Predicting Performance

Fracture Network Modeling and Reservoir Simulation—Golder Assocs. Principal Tom Doe explored fracture models (stochastic, mixed stochastic/deterministic); future uses of DFN models (whether we are getting enough data in a producing reservoir, rigorous testing of methodologies and hypotheses); improvements to look for in DFN modeling (better hydraulic data through improved flow logging to determine conductor spacing and transmissivity distributions and better monitoring data); and issues of scaling (diffusivity values from interference responses provide correct scaling; matching is nonunique with respect to permeability and storage for linear flow).

Simulating Multiphase-Fluid Flow and Reservoir Production—U. of Utah Professor Milind Deo reflected on the current state of reservoir simulation, offering case studies and observations. According to Deo, finite-difference simulators dominate the market. Pointing out that complex domains (with complex well patterns/arbitrary fracture patterns) are difficult to simulate, it was noted that finite-element models with body-fitting meshes are more appropriate for these systems. Following observations regarding capillarity, ATW participants were introduced to a new 3D/three-phase black-oil simulator based on the control-volume finite-element formulation. Case studies with tilted, partially penetrating faults were presented to demonstrate the applicability of the simulator.

Development of the West Rozel Fractured-Basalt Heavy-Oil Reservoir—In addition to giving a comprehensive overview of deposition description, a description and validation of the reservoir model, a layered faulted-section model, and a full-field model, MK Tech Solutions Principal Myron Kuhlman presented the formula for optimum oil recovery in this reservoir. Findings from this case study regarding optimum operating conditions included very little drawdown with horizontal wells, limiting liquid production to a few barrels per foot of well, optimizing cumulative production with a low gas-production rate, and maintaining reservoir pressure within 100 psia of the original pressure and reservoir oil viscosity within a few percent of the original.

Early Diagnosis of Natural-Fracture Contribution to Permeability at the Field-Appraisal Stage—Presenting a case history from eastern Siberia, Beicip Reservoir Modeling Manager Rolf Ackerman shared fracture characterization and conclusions for geological analysis of this reservoir in addition to discussing how calibrated hydraulic properties of the fractures were obtained by integrating the dynamic data on one well. Conclusions from this study included that integration of different data sets could be used to characterize the fracture network on one well, that diffuse fractures were identified from core description and borehole image data; that bed thickness and shalyness were factors that controlled diffuse fracture occurrence, and that fracture occurrence was based on analysis of the fold geometry (strain-partitioning model).

Determination and Sensitivity of Fracture Characterization Parameter for History Matching—Nitec Director Chet Ozgen addressed issues of fracture connectivity, sensitivity analyses, and the data sources for fracture parameters. Discussion took place on the use of finite-difference models in integrated studies, the simulator parameters used in finite-difference models, and how to correlate data sources to simulator parameters.

Improved and Enhanced Recovery

Enhanced Gravity Drainage Through Immiscible CO2 Injection—Kinder Morgan Reservoir Engineering Director John Dobitz shared experiences in the Yates field in Texas. In addition to an overview of field history and current activities in this field, possible EOR processes were examined, including thermal operations, methane injection, natural-gas-to-liquids (NGL) injection, and immiscible CO2 injection. According to Dobitz, thermal methods would have yielded a doubtful response, methane would be expensive for voidage replacement, and NGL injection would prove expensive and challenging. Immiscible CO2 injection was selected on the basis of reasonable cost and positive compositional effects. Gassed-out horizontal producers were converted into CO2 injectors.

Dilute-Surfactant-Aided Gravity Drainage—U. of Houston Professor Kishore Mohanty reported on recent experiments of traditional vs. fractured carbonates. Experiments included wettability, adsorption, oil-recovery range for anionic-surfactant solution, and brine imbibition. Discussion centered on the effects of formation water and lithology on the proposed chemicals.

Wettability Alteration in Carbonate Formations With Surfactants—Rice U. Professor George Hirasaki addressed wettability alteration, phase behavior and low interfacial tension, anionic-surfactant adsorption in calcite, spontaneous imbibition, and force displacement. Among conclusions reached were: dependence of optimal salinity on surfactant concentration and water-to-oil ratio can be correlated with natural soap/surfactant mole ratio; anionic-surfactant/Na2CO3/NaCl system is capable of inducing wide ultralow interfacial tension and wettability alteration; nitrogen blend allows the surfactant solution to be injected at optimal salinity as a single phase; and sodium carbonate reduces the adsorption of anionic surfactants on calcite as long as the surfactant solution remains clear.

Experience With IOR and EOR Processes—Kazemi introduced IOR and EOR processes, noting how gravity and capillary forces affect the ultimate result. According to him, EOR processes in single-porosity systems have a strong viscous displacement component; for NFRs, viscous displacement plays a small role while gravity drainage plays a major role. A comparison was made between IOR and EOR for steam injection (heating and gravity drainage), CO2 injection (compositional effects), and wettability alteration. It was concluded that NFRs can be produced efficiently, and use of appropriate IOR and EOR methods should improve ultimate recovery in fracture-/karst-dominated oil reservoirs. Viscous displacement in the matrix is minimal.

Information provided by Thomas Blasingame, Texas A&M U.