First EAGE/SBGf Workshop 2013, Rio de Janeiro - Fractures in Conventional and Unconventional Reservoirs

Open fractures are notoriously difficult to locate and characterise in the subsurface but are very easy to anticipate in a general way with only a moderate understanding of their structural geology setting and rock types and very little understanding of their mechanical evolution. But a general interpretation is not adequate. How can we become more educated in our prediction of fracture location, character and effect on reservoir performance? Part of this problem involves a number of individual beliefs that may be incomplete or outdated; part is because even apparently supported beliefs, when taken together, are inconsistent. What do we actually know about fracture formation and development and how does it deviate from what we think we know? If we are clearer in our understanding of what we really do know then our thinking about fractures and their character will be clearer, and hopefully more useful. We outline some common beliefs about fractures, primarily from outcrop observations and theory, and assess them individually. We then look at them together assessing how they support or negate each other and place them in a geomechanical context. Finally we consider how experimental results can help consolidate and progress our understanding.

We investigated the Quaternary epigenic karst system in the Jandaíra Formation, a Turonian–Campanian carbonate platform in the Potiguar basin, northeastern Brazil. We concentrated our investigation in the vadose zone of the present-day karst, but also used borehole data. The leaching zones are preferentially concentrated along pre-solution openings composed of faults, joints, and bedding planes. These dissolved and enlarged faults and beddings form a system of caves and sinkholes, which must be included in the architecture of karst systems. This controlled-karst architecture presents a predictable geometry. This process occurs when the carbonate platform is exposed for a long period and can be thus affected by surface processes such as dissolution of erosion. The Quaternary epigenic karst features we describe may correlate with paleokarst systems in other carbonate platforms.

The characterisation of carbonate fractured reservoir is often a difficult challenge due to the amount of data and the scale issues related to the fault and fracture network. The correct characterisation of the structural framework at all scales has a large impact on the correct understanding of the potentials of carbonate fractured reservoirs (deliverability), on the circulation of fluids (hydrocarbons and water) into these units and on the appraisal and development strategies to maximise the recovery factors in this type of fields. Mechanical Stratigraphy can be an important tool when trying to characterise the fault and fracture network in carbonate fractured sequences if an integrated approach of the available subsurface and outcrop analogues information is considered.

Outcropping analogs can provide key information on the 3D organization of fracture networks affecting carbonate reservoirs. Such information, however, needs to be integrated in a consistent work flow which includes i) 3D geometric model of the reservoir architecture, ii) mechanic modeling to determine areas of fracturing, iii) characterization of fracture networks, iv) determination of fracture-related permeability in small representative volumes and v) fluid flow simulations. Fracture networks have been recently analysed from km-scale pavements of the Jandaira Formation in NE Brazil. Combining images from a drone with outcrop analysis and thin section we present the first multi-scale characterization of the fracture field affecting the shallow water carbonates.

The porosity-permeability ratio on carbonate reservoirs is a complex subject. The objective of this study is to contribute for the investigation of the role of fractures on carbonate rocks´ permeability. For that we studied, from a geomechanical perspective, the process of development of the second biggest cave on the carbonate rocks of the Jandaíra Formation (Potiguar Basin, NE Brazil) . This cave (the Furna Feia) presents evidences of structural control and the numerical models performed on this study indicated that the fractures on the cave´s area are in dilation as a function of the present-day stress regime on the Basin. This process probably is a main factor for the conduction of fresh water flow along the natural fracture network on that area.

The considered oilfield was recently discovered part of Petrobras effort in developing subsalt assets. As with many carbonate fractured reservoirs it has proved difficult to describe and model the fracture distribution and their consequence on the permeability especially in its current exploration and delineation phase where the wells are limited to three or four at most. Description of Paper: This integrated study used geophysical, geologic, and engineering data simultaneously to improve the reservoir description. The successful characterization of the fractures was mostly due to the extensive use of narrow azimuth 3D post and pre-stack seismic data. Using unique seismic technologies to enhance seismic resolution and using the derived enhanced seismic in post and pre-stack stochastic inversions, key seismic attributes were generated and were used to dramatically improve the characterization of the reservoir key rock properties including the fractures in the considered field. These key seismic attributes were used as input in the integrated fracture modeling process. High resolution seismic attributes were also used in a neural network algorithm as a constraint to derive geologic drivers such as lithology, porosity, water saturation, permeability and other key reservoir properties. Many geologic and seismic drivers were correlated, using artificial intelligence tools, against a fracture density computed from FMI interpretations available at very limited wells. The fracture models were built using the Continuous Fracture Modeling (CFM) technology and then calibrated to well test data to create an effective permeability model that is able to account simultaneously for the effects of the fractures and matrix. Results, Observations, and Conclusions: The derived 3D seismic attributes and 3D geologic and fracture models were used to better understand the factors affecting the distribution of the key rock properties including the fractures and their effect on the permeability. The resulting porosity, water saturation and permeability models were input into a reservoir simulator and allowed a good match of the individual well performances without the need for any extensive history matching thus validating the input reservoir models. This achievement was made possible by the use of key high resolution post and pre-stack seismic attributes as drivers in the CFM technology that provided the resulting rock properties. Applications: This study demonstrates the successful use of all available GG&E reservoir data in an integrated approach despite the very limited number of available wells. The use of high resolution post and pre-stack seismic attributes in the CFM technology is illustrated with a carbonate field as a viable solution for the characterization and simulation of such complex reservoirs during the exploration and delineation phases where the number of wells could be very limited. Equipped with these seismically driven reservoir models, validated with dynamic models, optimal field development strategies could be designed and implemented with a reduced risk.

Fractures are often bounded by layer interfaces. The usual explanation for this observation can be improved if we consider that fracture assemblies represent strain. Geomechanical simulations of deformations show that layer interfaces are often where strain gradients are high, or even discontinuous, due to both material-property contrasts, as well as interface properties. Using deformation simulation outcomes, it is possible to create models in which the fracture distributions are correctly linked to the necessary strain, providing a prior model against which stochastic methods can be conditioned. Such trap-deformation models lead to fracture predictions that are in good agreement with outcrop observations, providing a process explanation for the observed heterogeneities in terms of orientations, sizes and local intensities of fractures. Such distribution models make good inputs for the real-time understanding of fractured reservoir responses during production, where geomechanical interactions play a potentially strong role.

In the case of low permeability reservoirs, hydraulic fracturing is a completion strategy that increases the contact between the well and the reservoir that finally will increase the well production. Fracture propagation is a result of the interaction of different physical processes like creation of new surface (fracture toughness), fluid flow in the fracture and porous media and their interaction, fracture geometry, rock elasto-plastic behavior, pre-existing fracture sets and in situ stress.

Seismic fracture prediction is becoming a more important exploration and development problem with the growing focus on unconventional reservoirs. A non-linear inversion technique is presented to estimate layer-based fracture parameters on azimuthal reflectivity data. The earth model assumes a single set of vertical fractures per layer parameterized in terms of linear slip parameters - the normal and tangential fracture weaknesses - and fracture strike. A linearized form of the Zoeppritz equations expressed in terms of Fourier Coefficients is used in a convolutional modelling scheme to estimate seismic amplitude data. The inverse problem is solved in a nonlinear fashion using simulated annealing. The new technique has several advantages over performing amplitude versus azimuth analysis (AVAz).

Thermally induced fractures, TIFs, occur in almost every water injection well. They are the result of the cooling of the reservoir rock as water reaches the sand face and near wellbore area. This cooling generates a contraction in the rock that results in a change in reservoir mechanical properties and the near wellbore stress conditions. The importance of understanding and engineering the process of TIFs is quite often underestimated, as TIFs are assumed to happen as under normal injection conditions. This paper presents the results of technical and field efforts carried out to overcome a variety of problems in water injection wells in two (2) particular types of reservoir: a high permeability (> 0.8 Darcy) and a very low permeability (< 0.03 Darcy); both sandstones located in the North Sea. Work was carried out in four(4) wells to utilize TIFs as a tool to maximize water injection and minimize well and reservoir integrity issues. This paper presents in detail the design process, its operational implementation and the results obtained for each reservoir. In both types of reservoirs, TIFs has allowed us to restore long term injectivity by overcoming a large number of constraints such as high levels of formation damage.