Static and Dynamic Characterization of Fracture Pattern in the Upper Jurassic Reservoirs of an Offshore Abu Dhabi Field: From Well Data to Full Field Modeling

Abstract

In the oil bearing reservoirs of a mature field in the offshore Abu Dhabi, the understanding of the fracture network is essential with growing gas and water productions. Fracture characterization and modeling at the full field scale is the key to match the production profile and to optimize infill drilling. The interpretation of fracture data from image logs and cores allows defining accurately the fracture pattern in terms of orientation, typology, density and relation with lithology and faulting. Two main fracture sets are chronologically observed:NW-SE fractures are generally mineralized and clustered around faults of similar orientations;NE-SW fractures are generally open, homogeneously distributed with their density controlled by the lithology and a few portion located in swarms. The calibration of fracture data with PLT's in horizontal drains suggests that:the main flow comes from the matrix in porous layers;in areas without mineralized fractures, wells can be connected to water, either through matrix (permeable layers) or through open NE-SW fractures, particularly in zones of large curvature where a higher fracture density is expected;in areas with mineralized fractures, water in permeable layers is channelized and tight layers can be vertically connected to more permeable layers by open fractures whereas the lateral flow can be hampered. Full field fracture modeling is done through an innovative geostatistical approach. First, discriminant analysis is used to derive, from geological attributes that are known to be physically related to fracturing, a (geological) component best correlated to well fracture data. Second, sequential indicator simulation generates equiprobable fracture density maps. Third, optimization of infill drilling is made through risk analysis. The pertinence of this approach for the optimization of well planning is validated against recently drilled wells. Matching the production profile is made more difficult however by the different opening times of production wells and by their complex completions. Introduction The oil bearing field under study was discovered in 1969 and put on stream in 1973. In 1979 a water injection scheme was launched followed by a gas injection development; first pilot in 1991 then full field starting 1997. To date, more than 120 wells were drilled including several horizontal drains. Although fractures were identified in this field at an early stage, their contribution to field performance was not initially considered as important. However, with the decline of the field and the water and gas injection programs, well behaviors showed that this early statement should be revisited. Indeed, the impact of fracture systems on such mature carbonate reservoirs can range from very restricted to conduit for fluid flow and reservoir impairment. This impact does not depend only on the type of fracture filling but also on fracture geometry, connectivity and density. According to the full field distribution of the fracture pattern, the three above impacts are possible. Consequently, only the characterization of all fracture parameters can lead to a comprehensive fracture model. This fracture model needs to be evaluated against dynamic data and the spatial distribution of fractures at the entire field scale is required to improve field development plans. The model can be also evaluated against new data and revised if necessary. This is the purpose of this paper.

Fast and Efficient Modeling and Conditioning of Naturally Fractured Reservoir Models Using Static and Dynamic Data

“…An illustration on a case-study involving different seismic, structural and geological attributes can be found in Gauthier et al (2002c). Other case-studies are also presented in Gauthier et al (2002a and2002b).…”

“…The spatial distribution of fracture densities being necessarily uncertain, a probabilistic approach should also be preferred, that relies on multivariate statistical analysis, for relating fracture density to explicative variables, and on geostatistics for addressing the spatial variability issue. Such an approach has been proposed by Gauthier et al (2002aGauthier et al ( , 2002bGauthier et al ( and 2002c where it proved efficient on different case-studies. It is the one presented later in this article.…”

“…Without getting into details, spatial and nonspatial model parameters are generally distinguished, as well as short-scale and large-scale dynamic data, thus leading to multiple-step and iterative approaches. Calibration of model parameters is often hand operated (Araujo et al 2004, Gauthier et al 2002band 2002c, Heffer et al 1999, seldom automatic (Suzuki et al 2005).…”

Characterization of Fractures and Faults from High-Resolution Image Logs to Optimize the Geological Model of a Fractured Carbonate Reservoir, Offshore Abu Dhabi

“…They have a very long vertical extent. The south-eastern flank and the north-western flank clearly show different faulting, the former being more faulted than the later [2] . The present day insitu stress field in the entire area is still under the influence of Zagros orogeny [3] .…”

Calibrating discrete fracture-network models with a carbonate three-dimensional outcrop fracture network: Implications for naturally fractured reservoir modeling

“…To model the full field fracture distribution in 3D, methods have been developed that are based on multivariate correlations between discrete fracture properties (e.g., fracture frequency) and other largescale continuous variables available from seismic data, geomechanical modeling, or other reservoir model properties (Gauthier et al, 2000;Gauthier et al, 2002a;Maerten and Maerten, 2006). These methods do, however, implicitly assume that smallscale well fracture data are hard data that should be honored and that they are implicitly geologically related to large-scale fracture or deformation systems, which is not always the case (Lamarche et al, 2012).…”