Abstract New observations of fracture corridors > 150 m tall and planes of bedding-parallel slip are integrated with sedimentological descriptions of the Asmari Formation to understand the main controls on the development of fractures in the Zagros Mountains of Iran. In the Kuh-e Pahn, fold-related fracture corridors are axis-parallel (NW-SE) and occur in the crest of the anticline. They form by neutral surface folding, but at a critical dip of the beds (c. 15°), bedding-parallel slip by flexural slip folding is the predominant mechanism. This relationship is substantiated by curvature calculations. Crestal fractures have a large vertical extent in mechanical unit B (> 150 m), primarily due to the lithological homogeneity of massive packstones within the Asmari Formation. Northerly and easterly trending fracture corridors, interpreted from satellite imagery, are spatially unrelated to the detachment folding of the cover series, but represent the distributed effect of deep-seated basement reactivation related to fault movement. These trends define high production zones in the nearby Gachsaran super giant oilfield. Observations from an adjacent eroded box fold, the Kuh-e Mish, with steeper dipping limbs (60°), revealed a contrast in the style of deformation, and we interpret these folds to represent different stages in box fold evolution.

Abstract The carbonate reservoirs in the Late Oligocene—Early Miocene Asmari Formation in the Dezful Embayment of SW Iran are characterized by low matrix permeability, and effective drainage is dependent on the occurrence of open fractures. Limited information on fracture orientation and fracture density is available from core and borehole image data, and high-quality/high-resolution three-dimensional seismic is often lacking in this area. Well and core data do not contain information on important fracture parameters like length distribution, crosscutting relationships, fracture density v. lithology and bed thickness. The understanding of fracture distribution and formation in the region and their effects on fluid flow has been greatly improved by the use of outcrop analogue data. Exposures of the Asmari Formation in the Khaviz Anticline are in close vicinity to the giant hydrocarbon fields. The Khaviz Anticline has a similar geometry and structural history to the major hydrocarbon fields in the area, and represents an excellent analogue for these. Two types of fracture features were observed: diffuse fracturing and fracture swarms. The diffuse fractures form networks and comprise structures grouped into four fractures sets, which are the typical for this type of anticline. Two orthogonal fracture sets are oriented parallel and perpendicular to the fold axis, and two conjugate fracture sets are oblique to the fold axis with their obtuse angle intersecting the trend of the fold axis. The fractures are typically stratabound, sub-perpendicular to bedding and commonly about the bounding stratigraphic surfaces. To a large extent the density and height of fractures in the Asmari Formation are controlled by the mechanical stratigraphy, which is controlled by the depositional environment and cycles. These outcrop data have been essential in the generation of discrete fracture network (DFN) models and the population of the fracture properties in the reservoir models.

Abstract It is well established that earthquake faulting can create permeability along a fault zone in high competence rocks - by mismatch of the fault walls and by secondary fracture in a surrounding damage zone - and that this permeability is created repeatedly during successive earthquake cycles. Less well proven is that such permeability is transient, being episodically reduced by precipitation of cements in the fracture porosity. The textures of carbonate dilation breccias, formed at around 1.7 km depth on the Dent Fault zone (NW England), lend support to this economically important concept of transient fracture permeability. The key observation is that many breccias reflect only a single episode of brecciation and reseal. A generally applicable explanation of such single-phase breccias is that they were resealed in the interval between major earthquakes, that this reseal made the breccia stronger that the intact rock, and that subsequent brecciation in the same rock volume was inhibited. This reseal-hardening model implies that transient permeability in fault zones may last no longer than the recurrence times of large earthquakes, and that the permeability conduits will change position in the damage zone with time, unless focused at a major fault jog or termination.

Abstract Relay ramps associated with overlapping faults are commonly regarded as efficient conduits for fluid flow across potentially sealing intra-reservoir fault zones. The current study demonstrates that structural heterogeneity in the often anomalously wide damage zone of relay ramps may represent potential baffles to intra-ramp fluid flow. A network of ramp-parallel, ramp-diagonal and curved cataclastic deformation bands causes compartmentalization of the ramp studied in Arches National Park, Utah. Harmonic average calculations demonstrate that, although single deformation bands have little or no effect on effective permeability, the presence of even a very small number of low-permeable deformation band clusters could reduce along-ramp effective permeability by more than three orders of magnitude. Thus, although relay zones may maintain large-scale geometric communication, the results of this study demonstrate that caution must be exercised when considering relay ramps as fluid conduits across sealing faults in a production situation. Although relay ramps clearly represent effective migration pathways for hydrocarbons over geological time, the extent to which they conduct fluids in a production situation is more uncertain. Quantitative approaches include adjusting the transmissibility multipliers for faults in reservoir models to allow for increased cross-fault flow. If, however, the effect of internal structural heterogeneity is not taken into consideration, this type of adjustment may lead to gross overestimation of the effect of relay ramps. Sedimentology, stratigraphy, burial history and deformation mechanisms are some of the controlling factors for the formation of such structural heterogeneities.

Abstract The seismic visibility of fractures depends on the magnitude of their normal and shear compliance and how these quantities vary with the scale of the fracture and depth of burial. Reliable estimates of fracture compliance as a function of confining pressure, in the range 10 −13 –10 −14 m Pa −1 , have been obtained from laboratory measurements on core samples. The possibility of fractal scaling of fracture parameters has been proposed, in which case fracture compliance might be expected to increase with the scale of the fracture. Laboratory and field estimates of fracture compliance are presented covering a range of fracture sizes. Compliance is shown to increase with the scale of the fractures. Results obtained are broadly consistent with the magnitudes predicted from linear slip theory, in which the displacement discontinuity across a partially sealed interface is linearly related to the traction on the interface.

Abstract This case study was undertaken for a low-porosity fractured carbonate reservoir with a complex fracture network resulting from several phases of tectonic activity. The integration of the image log and seismic-derived interpretations was problematic due to the complexity of the image log signature and the variable quality of the surface seismic data. Earlier experience indicated that VSPs may provide information on faulting and/or fracturing that may otherwise be difficult to determine with confidence from other data sources. Consequently, specialist VSP processing techniques were used to identify and map reflectors in three-dimensional space. Data acquired in two wells were reprocessed to interpret structural features and determine their geometries. The interpreted VSP reflectors were validated and integrated with the analyses of image logs and the interpretation of surface seismic data providing a constrained structural model that allowed the interpretation of seismic data away from well control and provided a starting point for seismic interpretation in areas where structural geometries were poorly imaged on surface seismic. It is shown that VSP, including vertical incidence, data can contribute to the understanding of reservoirs and enables well-derived information to be extrapolated away from the wells.

Abstract Acoustic emission (AE), ultrasonic velocity and petrophysical property measurements are increasingly being used as tools to investigate mechanical, thermal and hydraulic changes in rock masses around underground excavations and in the exploration of oil and gas. In order to develop these measurements into an effective tool it is necessary to understand how they relate to changes in the fracture and fluid content under controlled laboratory conditions. A polyaxial (true triaxial) test system has been developed to perform high-resolution measurements of AE, ultrasonic velocities and fluid permeability characteristics in a cubic specimen under true-triaxial stress. The polyaxial system also allows the measurement of AEs using an array of small-diameter, high-frequency transducers. The measurements allow the AE source locations and mechanisms to be calculated, thus providing an analysis of the distribution, orientation and type of fracture growth in the specimens. Example results are given from an experiment, showing that AE mechanisms during crack initiation describe a dominant failure mode with a significant tensile component.

Abstract Diagenesis and fracture are often linked processes in deformed rock. Empirical observations show that quartz-lined natural fractures are very common in sandstones that have been exposed to temperatures in excess of 90°C. These fractures exhibit crack-seal textures as well as cement bridges propping the fractures open and preserving fracture porosity. These diagenetic effects are examined in the context of detailed fracture characterizations generated by geomechanical modelling. Aperture, length and fracture network geometry are examined in the context of subcritical crack growth and various biaxial loading boundary conditions of varying initial anisotropy. An isotropic initial state results in more polygonal fracture patterns. A small initial anisotropy creates preferential through-going fractures that are later connected by cross-fractures. A larger initial anisotropy results in only one parallel set. The flow connectivity of isotropic and small strain anisotropic patterns appears high based on trace pattern geometry, but when the effects of diagenesis are added, preferentially filling smaller aperture fracture segments, connectivity can be significantly reduced. Finite difference, steady-state flow simulations demonstrate the permeability effects of heterogeneous fracture aperture distributions predicted by the mechanical model and permeability reduction caused by systematic diagenetic fracture sealing.

Abstract Different conceptualizations of fracture networks have been generated in order to simulate well tests in fractured reservoirs. These models have two purposes: firstly to investigate whether the behaviour of realistic fracture networks can be emulated with more simple models; and secondly to assess whether different sources of anisotropy have characteristic and recognizable influences on the pressure derivative. The anisotropy of the fracture networks was increased by decreasing the angle between two originally orthogonal fracture sets, and by increasing the permeability contrast between two orthogonal fracture sets. Results indicate that simple models can capture the first-order behaviour of more realistic examples. However because early time data are strongly influenced by the connectivity of the fracture network, the degree of anisotropy can only be assessed at later times in a test. Increasing anisotropy results in increasing heterogeneity and compartmentalization, and permeability anisotropy in an orthogonal system can only partially replicate the behaviour of a geometrically anisotropic system.

Abstract Predicting the effects of small-scale faults and fractures on reservoir behaviour requires a definition of their spatial distribution, orientation and mode. Elastic dislocation (ED) theory can predict the distribution of displacement, strain and stress in the rock volume surrounding major faults, from mapping of fault geometry and slip distribution in 3D seismic-reflection datasets. The intensity of small-scale faulting can be related to the predicted local strain, or the degree to which the shear stresses exceeded the rock failure envelope. We illustrate the methodology with three case studies: (i) a relatively-simple thrust anticline from Venezuela, where hydrocarbons are trapped in Pliocene sandstones within the faulted hanging wall anticline; (ii) the Gull-faks Field and of the North Sea; and (iii) the Miskar Field, offshore Tunisia, where large seismically mapped normal faults are forward-modelled to predict small-scale fault characteristics for comparison with detailed interpretation and seismic attribute analysis. Key requirements for the development of a robust predictive model of the small-scale fault and fracture network are a geometrically consistent framework model, judicious choice of mechanical properties, and a reasonable estimate of regional background strain.

Abstract Experiments aiming to study the conditions of fracture localization during buckle folding in brittle sedimentary layers of the upper crust were conducted using brittle paraffin wax multilayers submitted to axial and vertical load. The particular aim was to understand the mechanisms of development of subtle features often observed in apparently rounded natural folds. These consist of scarcely visible discontinuous fractured zones, more or less parallel to the fold axis, that separate domains of constant dip. These zones are of great interest for fluid dynamics in folded fractured reservoirs. It is shown that such axial kink boundaries referred to as dip-domain boundaries (DDBs) play a major role in the development of curvature in fractured buckle folds. DDBs observed in experiments consist of the localization and coalescence of tensile fractures along axial planes after the elastic instability stage. The tested physical parameters were (1) the intensity of the vertical load, (2) the bed thickness within a mechanical unit of constant thickness, (3) the stiff (multilayer)-soft (embedding material) thickness ratio and (4) the interlayer friction. In low interlayer friction conditions DDBs remain relatively localized during shortening, leading to chevron-style folds. With high interlayer friction conditions, DDBs tend to multiply laterally from an initial chevron fold. The latter phenomenon leads towards a more rounded fold, but one that still shows discontinuous curvature comparable to that observed in natural conditions. Fold models of flexural slip and tangential longitudinal strain (orthogonal flexure) are discussed under the influence of tensile failure in multilayer materials.

Abstract The Machar Oil Field is a fractured Chalk Group reservoir, of Late Cretaceous age, developed above a salt diapir in the Central Graben of the UK Continental Shelf. From 1998 to the present the field has been produced through a water injection scheme sanctioned on the basis of data acquired during a water injection pilot test carried out in 1996. After three years of strong production, performance was affected by an unexpectedly fast rise in water-cut. This prompted the field management team to review the available data and rebuild the reservoir model to serve as the basis for a revised development plan. This paper describes the data used and the lines of evidence followed by the team in order to identify, describe and model the key subsurface characteristics, and related uncertainties, controlling field behaviour and impacting reservoir management and development options. Key uncertainties are the nature and flow significance of an extensive network of faults, the presence, nature and significance of a system of sub-seismic background fractures and the role played by the matrix capillary and wettability properties in controlling recovery through imbibition levels and rates. The successful integration of a wide range of subsurface static data, such as biostratigraphy, fracture identification, stress identification, seismic fault mapping and dynamic data, such as production logs, pressure build-up tests and tracer recovery, resulted in improved reservoir characterization. The ensuing new reservoir model explained the field historic performance and suggested two redevelopment options, which were sanctioned and executed between 2003 and 2004, and resulted in improved field performance.

Abstract The Valhall Field is an Upper Cretaceous chalk reservoir located in the Central Graben area of the North Sea with production coming from the fractured Tor and Hod formations. Well tests and production history indicate that these formations are highly heterogeneous and that significant fluid flow occurs through both the matrix and fracture system. However there remained significant uncertainty about the specific controls and location of the main productivity conduits and how they would influence sweep efficiency during planned water flood. To address these uncertainties a range of possible conceptual fracture models were considered with respect to controls on major flow within the reservoir. Analysis indicated that the reservoir is dominated by a connected series of seismic scale faults acting as major flow conduits with smaller fractures providing a less significant enhancement to matrix permeability. A key input to this study was the examination of over 80 well tests. Simulation of a number of key well tests using a simple discrete fracture network model comprising a connected fault network and pseudo-matrix layer was able to reproduce the majority of the observed pressure derivative shapes. This gave some confidence to the understanding of major reservoir flow paths as well as providing calibrated fault properties for direct inclusion within the simulation model.

Abstract The Clair oilfield is a large fractured sandstone reservoir lying 75 km west of Shetland on the UK continental shelf. Fracture analysis and modelling was carried out in preparation for the phase 1 development, which started production early in 2005. Fracture clusters and discrete fluid inflows observed in wells are associated with faults and other localized deformation features tens or hundreds of metres apart. The reservoir has moderate to good matrix permeability, but well flow rates and profiles are fracture-dominated. Full-field geological models were built using conventional object modelling approaches for matrix and discrete fracture networks for fractures, and upscaled to populate a reservoir simulation grid. Dual-porosity, dual-permeability dynamic modelling (full-field and well-test) was undertaken to understand the fracture and matrix flow contributions and their interaction. Fracture models were conditioned to wells and to seismic data, including coherency and multi-azimuthal velocity information from a four-component, ocean bottom cable three-dimensional seismic survey. At this early stage in field development, there is insufficient calibration to select a single fracture model. Instead, well and depletion plans have been tested against multiple fracture models chosen to encompass a wide range of plausible outcomes.

Abstract Reservoir-scale structural heterogeneity, especially in terms of mechanical layering and natural fracture systems, is often insufficiently constrained by subsurface data alone. In North Oman, a large dataset in Cretaceous carbonates comprises data from multiple subsurface reservoirs and analogue outcrops. This provides an ideal opportunity to integrate outcrop constraints into the subsurface, and to calibrate the resulting models dynamically. For this purpose, a reservoir-scale analogue outcrop fracture template was created for the Jebel Madmar anticline in the Oman Mountains foothills. The outcrop template provides improved conceptual and quantitative constraints on (i) fracture types and dimensions (e.g. of NE-trending fracture corridors), (ii) fracture heterogeneity, both aerially and stratigraphically, (iii) fracture properties (e.g. cementation evolution, variations due to preferential fault/fracture reactivation) and (iv) structural evolution and history of reactivation. Within a regionally consistent structural framework, the outcrop template has greatly assisted in the creation of geologically realistic models for one of the fractured carbonate reservoirs, complementing the subsurface dataset. Initial dynamic calibration indicates successful application of the outcrop template in that the spatial fracture heterogeneity was succesfully captured in the reservoir models and provides a history match to production data. The reduced range of possible fracture system geometries in turn has provided better constraints on the effective fracture properties.

Abstract Borehole image logs, without accompanying flowmeter logs, are found to be of little use in identifying fluid conductive faults and fractures in a clastic reservoir in Oman. Both high permeability opening mode fluid conductive fractures and low-permeability deformation bands filled with sand and clay gouge appear conductive on image logs. Regional tectonic evolution and the nature of structural elements do not provide sufficient information to differentiate fluid conductive seismic faults or fault segments. Rapid water breakthrough may be indicative of fluid conductive faults and large fractures, but water cut rise is complex. The effect of faults and fractures may be masked by spud date, and water fingering through highly permeable layers, especially when the percentage of fluid conductive fault segments and fractures is low. Faults, deformation bands and large conductive fractures are found to cause openhole log spikes in this field. Since openhole logs are available from a large number of horizontal wells, it becomes feasible to prepare risk maps from map distribution of openhole log spikes even though it is not possible to pinpoint the exact location of fluid conductive faults and large fractures.

Abstract Well performance is extremely variable in the stacked sequence of relatively tight Devonian and Mississippian carbonates in the northern part of the Waterton Complex, Alberta, Canada. This is despite having an extensively developed fracture system present in all the wells. In order to determine why some wells penetrated more permeable fractures than others, a full reinterpretation of the geophysical, structural, stress, matrix and dynamic data sets was carried out. Flow simulations at sector scales using discrete fracture network models and fullfield continuum modelling were used to test a range of geological and dynamic scenarios. One of the most northwestern fields of the Waterton complex, the West Carbondale field, is the focus of the work presented. For this field the best-fit dynamic models consist of a major fracture zone, corresponding to either a seismic scale lineament or zone of enhanced curvature, trending through the area of most prolific wells. Outside this zone, the vast majority of the fracture system makes little contribution to the flow in the wells, other than slightly enhancing the reservoir permeability.

Abstract Fractured reservoirs contain a significant proportion of the global hydrocarbon reserves; however, they commonly exhibit unpredictable and extreme production behaviour. To develop such reservoirs it is necessary to obtain the best understanding possible of the fracture network and its impact on the fluid flow within the reservoir, using the full range of geoscience and engineering datasets available at all scales. This book presents an overview of current techniques and the latest technologies used to understand and exploit fractured reservoirs. The contributions are organized into sections on outcrop analogues, the application of geophysical techniques for fracture detection and imaging, numerical and analogue modelling and case studies from fields in the Middle East, Europe and North America. A number of the case studies also consider ways of assessing uncertainties in fracture geometry description and the implications for effective reservoir management.