Petroleum Geoscience - Volume 20, Issue 1, 2014

It is well known that heterogeneities in carbonate reservoirs impact fluid flow during production. However, few studies have examined the impact of the same heterogeneities on flow behaviour with different fluid properties and production scenarios. We use integrated flow simulation and experimental design techniques to investigate the relative, first-order impact of stratigraphic and sedimentological heterogeneities on simulated recovery in carbonate ramp reservoirs. Two production strategies are compared, which promote dominance of either horizontal or vertical flow.

We find that the modelled geology is more important than the simulated fluid properties and production scenarios over the ranges tested. Of the heterogeneities modelled here, rock properties and stratigraphic heterogeneities that control reservoir architecture and the spatial distribution of environment of deposition (EOD) belts are important controls on recovery regardless of the production strategy. The presence of cemented hardground surfaces becomes the key control on oil recovery in displacements dominated by vertical flow. Permeability anisotropy is of low importance for all production strategies. The impacts of stratigraphic heterogeneities on recovery factor and water breakthrough are more strongly influenced by fluid properties and well spacing in displacements dominated by vertical flow. These results help to streamline the reservoir modelling process, by identifying key heterogeneities, and to optimize production strategies.

Whilst sophisticated multiphase fluid flow models are routinely employed to understand behaviour of oil and gas reservoirs, high-resolution data describing the three-dimensional (3D) distribution of rock characteristics is rarely available to populate models. We present a new approach to developing a quantitative understanding of the effect of individual controls on the distribution of petrophysical properties and their impact on fluid flow. This involves simulating flow through high-detail permeability architectures generated by forward modelling of the coupled depositional–diagenetic evolution of isolated platforms using CARB3D+. This workflow is exemplified by an investigation of interactions between subsidence and climate, and their expression in spatial variations in reservoir quality in an isolated carbonate platform of similar size and subsidence history to the Triassic Latemar Platform.

Dissolutional lowering during subaerial exposure controls platform-top graininess via platform top hydrodynamics during the subsequent transgression. Dissolved carbonate is reprecipitated as cements by percolating meteoric waters. However, associated subsurface meteoric dissolution generates significant secondary porosity under a more humid climate. Slower subsidence enhances diagenetic overprinting during repeated exposure events. Single-phase streamline simulations show how early diagenesis develops more permeable fairways within the finer-grained condensed units that can act as thief zones for flow from the grainier but less diagenetically altered cyclic units.

Partially dolomitized carbonate rocks of the Middle East and North America host large hydrocarbon reserves. The origin of some of these dolomites has been attributed to a hydrothermal mechanism. The Benicàssim area (Maestrat Basin, eastern Spain) constitutes an excellent field analogue for fault-controlled stratabound hydrothermal dolomitization: dolostone geobodies are well exposed and extend over several kilometres away from seismic-scale faults. This work investigates the main controls on the formation of stratabound versus massive dolomitization in carbonate sequences by means of two-dimensional (2D) reactive transport models applied to the Benicàssim case study. Simulation results suggest that the dolomitization capacity of Mg-rich fluids reaches a maximum at temperatures around 100 °C and a minimum at 25 °C (studied temperature range: 25–150 °C). It takes of the order of hundreds of thousands to millions of years to completely dolomitize kilometre-long limestone sections, with solutions flowing laterally through strata at velocities of metres per year (m/a). Permeability differences of two orders of magnitude between layers are required to form stratabound dolomitization. The kilometre-long stratabound dolostone geobodies of Benicàssim must have formed under a regime of lateral flux greater than metres per year over about a million years. As long-term dolomitization tends to produce massive dolostone bodies not seen at Benicàssim, the dolomitizing process there must have been limited by the availability of fluid volume or the flow-driving mechanism. Reactive transport simulations have proven a useful tool to quantify aspects of the Benicàssim genetic model of hydrothermal dolomitization.

A workflow is presented for constructing an outcrop-based static geological model for Miocene heterozoan, photozoan, oolitic and microbial carbonates in SE Spain. Fieldwork and LiDAR (Light Detection And Ranging) data were integrated to create a photorealistic virtual outcrop. Field-based data are depicted on the virtual outcrop, and used to identify and correlate horizons. Multivariate regression is used to extend stratigraphic horizons away from the outcrop and to create realistic time-equivalent model layers. Mapping and measured stratigraphic sections are used to assign depofacies within model layers. Pinning points are used to reconstruct sea-level history, and palaeotopography is used to define palaeobathymetry. Geospatial analysis of depofacies occurrence is used to extrapolate facies while also enforcing palaeobathymetric controls on depofacies distribution. Dolomitization and meteoric calcite cementation are dominant diagenetic products affecting porosity and permeability. Their distribution was mapped in the field and amounts were quantified in the laboratory, and these were used to populate diagenetic products into geomodel cells. Six scenario models were built to represent different stages and combinations of diagenetic effects on porosity and permeability. Construction of the static geological models required the development of new methods. These include: (1) a new workflow to extend clinoform surfaces across the model area by fitting equations to horizon picks made on the outcrop; (2) a new workflow relating relative sea level to model layers to allow calculation of palaeo-water depth and relate that to facies probability; and; (3) an experimental application to predict porosity and permeability from objective visual descriptions of carbonate samples.

Fractures in subsurface reservoirs are known to have significant impacts on reservoir productivity. Quantifying their importance, however, is challenged by limited subsurface observations, and intense computations for modelling and upscaling. In this paper, we present a workflow to construct and upscale fracture models based on outcrop studies in the Latemar carbonate platform (northern Italy). Fractures were first analysed to investigate their distinct characteristics throughout different sedimentological domains; that is, slope, margin and platform interior. Several fracture models were then built to represent different domains and were used to upscale fracture-network properties. Small-scale models were preferred to the reservoir-scale ones to enable multiple realizations and various sensitivity analyses in a time-efficient manner.

The fracture characteristics of the Latemar Platform vary across different sedimentological domains. This variation results in a non-homogeneous permeability field that will influence flow behaviours. Fractures in the slope domain are typically tall but low in intensity, resulting in relatively low effective permeabilities. In the platform interior, smaller sizes of fractures combined with higher intensities give rise to higher effective permeabilities. In general, fracture intensities, aperture and their intrinsic permeability would have a significant impact on the permeability field. Fracture shape and orientations are more important in affecting the connectivity.

Field observations from several outcrops in the Eastern High Atlas Mountains, near Amellago (Morocco), are used to determine fracture-network model parameters, such as the aspect ratio of the fractures represented as rectangles whose longer side is horizontal, the volumetric area of fracture surfaces, the fracture mean size and the fracture density. The fracture orientations can be roughly approximated by Fisher distributions, where the parameters are determined by outcrop measurements. The permeability of the fracture networks can be calculated by application of the Snow equation for infinite fractures or by numerical resolution of the flow equation for fracture networks generated with the parameters deduced from the outcrop measurements. These two permeability estimations are shown to be in good agreement, which suggests that theoretical or semi-empirical solutions may provide reasonable approximations of fracture-network permeability in some carbonate reservoirs when conditioned to appropriate outcrop and subsurface data.

Carbonate reservoirs pose significant challenges for reservoir modelling and flow prediction due to heterogeneities in rock properties, limits to seismic resolution and limited constraints on subsurface data. Hence, a systematic and streamlined approach is needed to construct geological models and to quickly evaluate key sensitivities in the flow models. This paper discusses results from a reservoir analogue study of a Middle Jurassic carbonate ramp in the High Atlas Mountains of Morocco that has stratigraphic and structural similarities to selected Middle East reservoirs. For this purpose, high-resolution geological models were constructed from the integration of sedimentological, diagenetic and structural studies in the area. The models are approximately 1200×1250 m in size, and only faults (no fractures) with offsets greater than 1 m are included. Novel methods have been applied to test the response of flow simulations to the presence or absence of specific geological features, including proxies for hardgrounds, stylolites, patch reefs, and mollusc banks, as a way to guide the level of detail that is suitable for modelling objectives. Our general conclusion from the study is that the continuity of any geological feature with extreme permeability (high or low) has the most significant impact on flow.

A high-resolution three-dimensional (3D) outcrop model of a Jurassic carbonate ramp was used in order to perform a series of detailed and systematic flow simulations. The aim of this study was to test the impact of small- and large-scale geological features on reservoir performance and oil recovery. The digital outcrop model contains a wide range of sedimentological, diagenetic and structural features, including discontinuity surfaces, shoal bodies, mud mounds, oyster bioherms and fractures. Flow simulations are performed for numerical well testing and secondary oil recovery. Numerical well testing enables synthetic but systematic pressure responses to be generated for different geological features observed in the outcrops. This allows us to assess and rank the relative impact of specific geological features on reservoir performance. The outcome documents that, owing to the realistic representation of matrix heterogeneity, most diagenetic and structural features cannot be linked to a unique pressure signature. Instead, reservoir performance is controlled by subseismic faults and oyster bioherms acting as thief zones. Numerical simulations of secondary recovery processes reveal strong channelling of fluid flow into high-permeability layers as the primary control for oil recovery. However, appropriate reservoir-engineering solutions, such as optimizing well placement and injection fluid, can reduce channelling and increase oil recovery.

Relative permeability is one of the key parameters governing fluid flow through porous media. Determination of relative permeability is traditionally conducted in the laboratory using either recombined reservoir oil or laboratory oil at simulated reservoir conditions, or simply at laboratory conditions. This is because it is expensive to sample representative uncontaminated reservoir fluids and extremely difficult to cut reservoir cores without altering their surface properties. Restoring rock properties to their original reservoir conditions has been a technical challenge to the industry. Upscaling laboratory special core-analysis data to reservoir scale is also a concern. Consequently, the industry has been researching new methods to extract relative permeability in situ, including the utilization of specially designed permanent downhole electric resistivity array, pressure and flow rate measurements. In this study, a different approach was taken. A semi-analytical model, developed to infer relative permeability from resistivity, was verified using experimental and field data. Relative permeability and resistivity were measured simultaneously in the laboratory. The results demonstrated that relative permeability derived from measured resistivity was close to the measured relative permeability. Relative permeability calculated from resistivity logs in two wells was compared to measured relative permeability with encouraging results.