Summary. The Frigg field is a major North Sea gas reservoir composed of turbiditic sediments. This paper describes the detailed geologic modeling and three-dimensional (3D) reservoir simulation of the field.A geologic model containing sand lobes and intercalating shales has been defined from seismic and well data. Special attention was paid to the realistic modeling of the shales. The more continuous (deterministic) shales between the turbiditic sand lobes of the reservoir were directly implemented in the model as horizontal flow barriers. The more discontinuous shales within the lobes were modeled with the method of Haldorsen and Lake and Begg and King. This method uses statistical geologic information and well data to calculate effective vertical permeability. The simulator was built as a 3D, two-phase (gas/water) model. Field data for 8 years of production were matched. A good match of both fluid levels and pressures was obtained. This reservoir study demonstrates that the impact of the shales on the reservoir behavior in general and the movement of the gas/liquid contact (GLC) in particular is essential. Introduction The Frigg field straddles the Norway/U.K. boundary in the northern North Sea (Figs. 1 and 2). Discovered in 1971 and brought on stream in 1977, the field is unitized and jointly owned by the Frig-U.K. Assn. (Elf U.K. and Total Oil Marine and the Frigg-Norwegian Assn. [Elf-Aquitaine Norge (operator), Norsk Hydro, Total Oil Marine Norsk, and Statoil]. Top reservoir is at about 1790 m [5.875 ft] mean sea level (MSL). The gas has a maximum column of 160 m [525 ft] overlying an ∼ 2- to 10-m [∼ 6.6- to 32.8-ft] -thick oil rim. Gas initially in place was illustrated at 265 × 10(9) std m3 [9,360 × 10(9)scf] before unitization in 1976. Initial model studies assumed a homogeneous sand reservoir with local occurrences of shale and limestone (Fig. 3) A tuff and shale layer separating the Frigg from the Cod formation, well below the GLC, was considered the only barrier for flow in the reservoir model. That this barrier had to contain permeability windows could be deduced from the active aquifer response during the production phase.GLC movements in the first observation well, Well 25/1-A22, did not give rise initially to drastic changes in the basic concept of the reservoir. However, results of the second observation well, Well 10/1-A25, which was deepened in Aug. 1984. demonstrated an ∼40-m [∼130-ft] -higher rise in GLC than observed in Well 25/1-A22. Furthermore, repeat formation tester (RFT) data of this well showed a pressure step over shales above the assumed tuff and shale barrier and no pressure step across the "tuff zone" itself (Fig. 4). This indicated a more complex, dynamic behavior of the reservoir than originally anticipated. An intensive appraisal involving the drilling of three remote appraisal wells (Wells 1011–5, 25/1–7, and 25/1–8) and the deepening of two platform wells was consequently undertaken. The planning of a 3D seismic survey was also initiated. The presence of shales in the Frigg formation that varied in lateral extent, forming horizontal flow barriers, appeared to be the main cause for the complex behavior of the reservoir. At the end of 1984, Norsk Hydro initiated the independent study described here to model and to simulate the effect of these shales in the Frigg reservoir. This 3D simulation study was concluded in Sept. 1985 and incorporates the results of all recently drilled remote appraisals and the deepening of Well 25/1A 14. A combination of direct and statistical methods was selected for the modeling to reflect the impact of the shales. Note that the geologic complexity combined with relatively sparse well spacing gives room for many uncertainties. For this reason, the interpretation presented here should also be seen as Norsk Hydro's own. among others. Note also that the field operator and the other Frigg partners are still conducting substantial studies. Theoretical Concept The basic philosophy behind the theoretical concept of this study is defined as follows. The complete incorporation of the effects of contrasting lithologic units is essential for reservoir simulation. Consequently, realistic estimations of occurrence and geometry of these units have to be made when data are not sufficient to define all units separately. Application of this philosophy will improve consistent incorporation of reservoir heterogeneity, which is often neglected when only the correlatable events/units are represented in the model. It will consequently yield a more realistic production forecast on a field scale. The situation in the Frigg field was typical for the application of the stated philosophy because insufficient data were available to define all contrasting lithologic units (sands and shales) separately at the time the study was initiated.A combination of direct and statistical methods has been selected for the modeling. The more continuous shales at the boundaries of the reservoir units (lobes) were modeled directly as vertical transmissibility barriers. They are classified here as deterministion shales. Their mapping, involved much postulation, however, because of a lack of data. The less continuous and uncorrelatable shales within the reservoir units (lobes) were handled statistically and are classified as stochastic shales (Fig. 5). Concepts based on the ideas of Haldorsen and Laker and Begg and King' were used for the statistical handling of the stochastic shales. In fact, a modified version of the Begg-King streamline method was applied. SPEFE P. 493^

No abstract

A numerical simulator is often used as a reservoir management tool. One of its main purposes is to aid in the evaluation of number of wells, well locations and start time for wells. Traditionally, the optimization of a field development is done by a manual trial and error process. In this paper, an example of an automated technique is given. The core in the automization process is the reservoir simulator Frontline. Frontline is based on front tracking techniques, which makes it fast and accurate compared to traditional finite difference simulators. Due to its CPU-efficiency the simulator has been coupled with an optimization module, which enables automatic optimization of location of wells, number of wells and start-up times. The simulator was used as an alternative method in the evaluation of waterflooding in a North Sea fractured chalk reservoir. Since Frontline, in principle, is 2D, Buckley-Leverett pseudo functions were used to represent the 3rd dimension. The areal full field simulation model was run with up to 25 wells for 20 years in less than one minute of Vax 9000 CPU-time. The automatic Frontline evaluation indicated that a peripheral waterflood could double incremental recovery compared to a central pattern drive. Introduction In the evaluation of North Sea reservoirs numerical simulators play a major role. The simulators are vital tools in the field development phase when it comes to the planning of number of wells, their locations and starting schedules. Also, the simulators are key elements in the process of managing and optimizing the reservoirs in the production phase. Fractured chalk reservoirs in the North Sea that originally were developed by primary depletion, have shown substantial decline in production due to loss of reservoir pressure. A major field has responded successfully to water injection as a secondary recovery process, and more fields are currently being reviewed for waterflooding. The planning of the waterfloods involves simulation studies for the optimization of number of wells and well patterns. Some new wells must be drilled and others are converted from producers to injectors. The aim is in principle, to maximize recovery of oil with the lowest possible number of wells. For this purpose, full field models are required in some context. P. 473^

A new method for generating pseudo relative permeability curves and its applications in a high accuracy front tracking simulator is presented.The method is applied on 3 different cases, which all gave good results. The different cases consisted of homogenous and inhomogeneous horizontal reservoir models.The results using a front tracking simulator were more accurate than the results presented by Nyrl/lnning [1], using a finite difference simulator.

This paper describes the application of a front tracking simulator for studying flow in a 2D cross-section of a braided river system. The internal sand body geometries were based on a braided river outcrop from the USA. Rock properties were derived from a North-Sea oil field. Front tracking represents a direct, accurate and fast method to simulate oil and water flow in highly hetergeneous reservoirs. In the front tracking simulator the saturation solution is decoupled from the grid and its accuracy is to a large extent independent of grid block sizes end geometry. This makes the reservoir simulator well suited for modelling of complex geological architectures. The simulator has been compared with analytical methods and traditional finite difference simulators. Results from a water-oil displacement process in a braided river system are presented, showing fingering of the saturation fronts due to par-ability contrasts. For coarse grid systems. the front tracking simulator proved superior to the finite difference simulator with respect to the numerical solution. An automatic grid fitting option that utilizes triangular grid cells enabled an easy and fast construction of the complex geological model. Introduction The increasing use of statistical geological models implies a more detailed geological description and necessitates an increased number of geological realizations. This in turn calls for simulators which are able to correctly describe detailed deterministic models without an excessive CPU-consumption. It is also required that the construction of new realisations can be carried out in an easy and time efficient manner. The use of standard five-point finite difference simulators has practical limitations. A detailed geological description demands in itself a large number of grid blocks and to avoid inaccuracies in the numerical solution, a dense grid system is required. The front tracking simulator as presented in this paper fully meets the above requirements, both in terms of CPU- and man-time efficiency, as well as accuracy in the numerical solution. To demonstrate the capabilities of the front tracking simulator, water flooding in a cross-section from a braided river system has been simulated. THEORETICAL ASPECTS The reservoir simulators most commonly used today are based on finite difference methods to approximate the partial differential equations that describe fluid flow in porous media. Since both the pressure and the saturation equation are solved iteratively, and stable solutions require restricted time step times. such models can be CPU-demanding, particularly when simulating larger grid systems. Usually. a five-point difference scheme is used and unwanted grid effects will occur in case of skewed grid cells. Hence, such simulators have their limitations when modelling irregular reservoir geometries. Also. numerical dispersion can be significant when large grid blocks are used. This creates a need for larger grid systems to obtain reliable simulation results. The front tracking simulator is based on a different numerical method compared with traditional finite difference simulator, but the mathematical foundation is the same.