Abstract The 3D geological model is still regarded as one of the newest and most innovative tools for reservoir management purposes. The computer modelling of structures, rock properties and fluid flow in hydrocarbon reservoirs has evolved from a specialist activity to part of the standard desktop toolkit. The application of these techniques has allowed all disciplines of the subsurface team to collaborate in a common workspace. In today’s asset teams, the role of the geological model in hydrocarbon development planning is key and will be for some time ahead. The challenges that face the geologists and engineers will be to provide more seamless interaction between static and dynamic models. This interaction requires the development of conventional and unconventional modelling algorithms and methodologies in order to provide more risk-assessed scenarios, thus enabling geologists and engineers to better understand and capture inherent uncertainties at each aspect of the geological model’s life.

Abstract The 3D geological model is still regarded as one of the newest and most innovative tools for reservoir management purposes. The computer modelling of structures, rock properties and fluid flow in hydrocarbon reservoirs has evolved from a specialist activity to part of the standard desktop toolkit and the application of these techniques has allowed all disciplines of the subsurface team to collaborate in a common workspace. The papers in this volume result from a two-day conference at the Geological Society that brought together modelling software practitioners from the geoscience community. The papers presented here provide an excellent illustration from across the industry and academia of where we are and give some interesting pointers to where we are heading. The authors and their parent companies and institutes are thanked for sharing this information.

Abstract In this paper, we present workflows, key relationships and results of multiple stochastic fault seal analyses conducted on geocellular geological or (static) reservoir grids. Ranges of uncertainties are computed from new and published datasets for the different input relationships (e.g. throw, VShale to VClay, fault clay prediction, fault rock clay content to permeability); these are used as input into stochastic modelling processes and the impact of each is assessed. The power of stochastic modelling to focus interpretation and risking effort is reviewed. Reducing the uncertainty distributions from the published data ranges has a massive impact on the range of predicted fault seal properties. Halving the uncertainties associated with the computation of the transmissibility multiplier, for instance, reduces this range from 7 to 1–1.5 orders of magnitude of the base-case value (no uncertainty). Importantly, when combined together, the median predictions from each individual parameter do not lead to the median value for the final prediction; average relationships combined together will not therefore produce the average final prediction. This is a powerful result for two reasons: first, current geological modelling packages use global trends to define fault properties and so are likely to predict spurious results; and secondly, reducing the uncertainty on specific relationships by around 50% is an achievable goal. Locally calibrated datasets and relationships (field-specific) based on carefully characterized samples should allow for this improvement in prediction accuracy. This paper presents a review of fault seal techniques, published data and the potential pitfalls associated with the analyses.

Abstract Geocellular modelling of diagenetically altered carbonates is challenging as geometries and pore systems often appear irregular. It has long been recognised, however, that tectonic evolution forms a framework that can influence patterns of carbonate facies, diagenesis and fracturing, the combination of which determines reservoir geometries and properties. Unravelling these processes can reveal trends that were not evident from well data alone. Such trends are useful in building geocellular models that extrapolate reservoir properties along them and can be used for economic screening of undrilled areas. This paper shows how standard reservoir modelling software can be used to model complex geology. In particular, it is shown how a carbonate reservoir model was constructed based on concepts of facies, burial diagenesis, hydrocarbon charge and fracturing. Workflows are discussed that were employed to distribute reservoir properties related to these processes.

Abstract The present study attempts to qualify the impact of tectonic parameters on hydrocarbon production in reservoir models representing four clastic depositional environments. Eleven sectors from existing 3D reservoir models, representing fluvial, tidal, shallow marine and deep marine depositional settings, were re-sampled into a fixed-volume, unfaulted model grid. Each sample was permutated into 73 different faulted model configurations by using predefined combinations of fault patterns, maximum fault-throw, shale gouge ratio and shale smear factor. The resulting 803 models were run in a fluid flow simulator and results statistically analysed to identify changes in fluid flow response caused by changing model input parameters. Finally, outcomes for each of the four depositional environments were compared. Although an inadequate database and technical limitations with the input models restrict our ability to draw quantitative conclusions, a number of qualitative interpretations can be made. The four investigated stratigraphies respond differently to identical fault parameter settings. Thus, there is a clear link between the depositional model input and the impact that faults have on production parameters. This suggests that sedimentological factors have a significant influence on which and to what extent fault parameters affect petroleum production. Varying degrees of impact can be identified for each fault parameter in each of the four depositional model types. Within the limitation of this study, a qualitative assessment of these is formulated.

Abstract Building geologically realistic reservoir models that honour well data and seismic-derived information remains a major challenge. Conventional variogram-based modelling techniques typically fail to capture complex geological structures while object-based techniques are limited by the amount of conditioning data. This paper presents new reservoir facies modelling tools that improve both model quality and efficiency relative to traditional geostatistical techniques. Geostatistical simulation using Multiple-Point Statistics (MPS) is an innovative depositional facies modelling technique that uses conceptual geological models as training images to integrate geological information into reservoir models. Replacing the two-point statistic variogram with multiple-point statistics extracted from a training image enables to model non-linear facies geobody shapes such as sinuous channels, and to capture complex spatial relationships between multiple facies. In addition, because the MPS algorithm is pixel-based, it can handle a large amount of conditioning data, including many wells, seismic data, facies proportion maps and curves, variable azimuth maps, and interpreted geobodies, thus reducing the uncertainty in facies spatial distribution. Facies Distribution Modelling (FDM) is a new technique to generate facies probability cubes from user-digitized depositional maps and cross-sections, well data, and vertical facies proportion curves. Facies probability cubes generated by FDM are used as soft constraints in MPS geostatistical modelling. They are critical, especially with sparse well data, to ensure that the spatial distribution of the simulated facies is consistent with the depositional facies interpretation of the field. A workflow combining MPS and FDM has been successfully used in Chevron to model important oilfield assets in both shallow- and deep-water depositional environments. Sedimentary environments can be characterized by a succession of deposition of elements, or rock bodies, through time. These elements are traditionally grouped into classes, commonly named ‘depositional facies’, based on their lithology, petrophysical properties, and biological structures. For example, the typical depositional facies encountered in fluvial environments are high permeability sand channels, with levées and crevasse splays, having a more variable range of permeability and net-to-gross ratio, within a background of low permeability shaley facies.

Abstract Reservoir production is highly dependent on reservoir models. A key problem faced in the development of a hydrocarbon reservoir is that of constructing a reservoir model that can generate reliable production forecasts under various development scenarios. Therefore, geological models have to be built in three dimensions (3D). Unfortunately, manual construction of 3D geological models (deterministically) is almost impossible, which explains why geologists often limit their interpretation to two dimensional (2D) correlation panels, fence-diagrams or maps. Consequently, geological conceptual models are rarely included or considerably simplified in reservoir models used for flow simulations and replaced by stochastic or geostatistic approaches. In spite of this admission of failure, sedimentological cross-sections and maps contain most of the knowledge and concepts of sedimentologists. They represent the outcome of sedimentological studies, including available well data, seismic interpretation and especially sedimentological and environmental concepts, incorporating all facies transitions and successions in a high-resolution stratigraphic framework. They allow fine temporal- and spatial-scale sedimentological heterogeneities to be identified. The integration of these fine-scale sedimentological heterogeneities is an essential step in improving the precision and accuracy of static reservoir models and volumetric calculations. This paper demonstrates the quantitative influence of introducing sedimentological information into the reservoir characterization workflow using a simple deterministic workflow. The described incorporation of sedimentological knowledge through facies 3D proportions cubes allows a direct assessment to facies distribution multi-realization scheme and associated uncertainties by applying stochastic simulations.

Abstract Rapidly developing methods of digital acquisition, visualization and analysis allow highly detailed outcrop models to be constructed, and used as analogues to provide quantitative information about sedimentological and structural architectures from reservoir to subseismic scales of observation. Terrestrial laser-scanning (lidar) and high precision Real-Time Kinematic GPS are key survey technologies for data acquisition. 3D visualization facilities are used when analysing the outcrop data. Analysis of laser-scan data involves picking of the point-cloud to derive interpolated stratigraphic and structural surfaces. The resultant data can be used as input for object-based models, or can be cellularized and upscaled for use in grid-based reservoir modelling. Outcrop data can also be used to calibrate numerical models of geological processes such as the development and growth of folds, and the initiation and propagation of fractures.

Abstract Deltaic reservoirs typically contain seaward-dipping surfaces termed clinoforms. Shale and carbonate cements covering clinoforms can frequently form a barrier or baffle to horizontal flow within reservoirs, However, clinoforms are not typically included in static or flow simulation models because they are often not identified in well data and little is known about their 3D geometry. High quality outcrops such as Cretaceous deposits of the US Western Interior Seaway provide an ideal opportunity to study clinoform geometry and shape, and to model their effects on flow. Within this study, two deltaic systems have been studied. The first is the Ferron Delta which crops out in the Wasatch Plateau, central Utah and is a highstand complex comprised of a number of small, overlapping lobes. Clinoforms are common and their 3D geometry is controlled by the position of the lobes. Large growth fault structures within the lobes add to the potential reservoir complexity. The forced regressive Panther Tongue Delta crops out in the Book Cliffs of Utah and is comprised of downstepping lobes with internal clinoforms. Data for modelling included traditional sedimentary logs, photomontages and calibrated photo logs. Models were built in IRAP RMS using a variety of modelling techniques from simple Truncated Gaussian Simulations on a regular grid to object modelling of shale barriers within a dipping grid designed to follow the clinoforms. The models were flow simulated as a means of comparing the different techniques for representing the heterogeneity results show that not modelling clinoforms explicitly in a dipping grid can lead to significant overestimates in the forecasted production; water injection in a down depositional dip position is optimum, and that there are only limited production differences between highstand and lowstand deltas.

Abstract Geological systems exhibit variability and structure at a wide range of scales. Geological modelling of subsurface petroleum reservoirs has generally focused on the larger scales, driven by the types of measurement available and by computation limitations. Implementation of explicitly multiscale models of petroleum reservoirs is now realistically achievable and has proven value. This paper reviews the main approaches involved and discusses current limitations and challenges for routine implementation of multiscale modelling of petroleum-bearing rock systems. The main questions addressed are: (a) how many scales to model and upscale; (b) which scales to focus on; (c) how to best construct model grids; and (d) which heterogeneities matter most? The main future challenges identified are the need for improved handling of variance and more automated construction of geological and simulation grids.

Abstract In the early stages of field development, there are many uncertainties and the current trend is to generate coarse-scale models so that many simulations can be run rapidly in order to determine the main sensitivities in a model. However, at a later stage, more data are available and there is less uncertainty in the model, so a more detailed modelling approach is desirable. In this paper, we discuss two modelling procedures which are suitable for different stages of the life of a field, and address the problem of upscaling at each stage. First, we consider a novel approach for generating coarse-scale models for evaluating uncertainty. When there is a large amount of uncertainty, the generation of fine-scale models is too time-consuming, and upscaling them by large factors may produce large errors. We demonstrate an alternative approach for modelling at a coarse scale, while preserving the heterogeneity of the fine-scale distribution, in such a way as to reproduce the fine-scale flow results. Secondly, we focus on more detailed geological models, generated at a later stage of field development. These models may comprise millions of grid cells, and may be highly heterogeneous. Such models require upscaling, but traditional methods may be very inaccurate. We have developed a method for upscaling using well-drive boundary conditions. Tests of this method show that it can reliably reproduce the fine-scale recovery in a range of models.

Abstract The scenario-based reservoir modelling method places a strong emphasis on the deterministic control of the model design, contrasting with strongly probabilistic approaches in which effort is focused on the ‘richness’ of a geostatistical algorithm to derive multiple stochastic realizations. Scenario-based approaches also differ from traditional ‘rationalist’ modelling, which often involves the construction of only a single, best-guess or base-case model. The advantage of scenario modelling is that there is no requirement to anchor on a preferred, base-case model, and it is argued here that selection of a base case is detrimental to achieving appropriately wide uncertainty ranges. Multiple-deterministic scenario modelling also carries the advantage of maintaining explicit dependency between model parameters and the ultimate model outcome, such as a development plan. The approach has been applied widely to new fields, where multiple deterministic reservoir simulations of a suite of static models can be easily handled. The approach has also been extended to mature fields, in which practical approaches to multiple-history matching are required. Mature field scenario modelling, in particular, illustrates the weaknesses of base-case modelling, and delivers a strong statement on the non-uniqueness of modelling in general. Current issues are the need to develop better methodologies for multiple-history matching, and for linking discrete, deterministic, scenario-based outcomes to probabilistic reporting. Experimental design methods offer a solution to the latter issue, and a simple, practical workflow for its application is described.

Abstract In this paper, we describe a reservoir-modelling case history of Mafumeira, a Chevron-operated field located in Offshore Angola. The field has only six well penetrations and lies within the closure of nearly 60 square kilometres; the purpose of the study was to capture a range of subsurface uncertainties for evaluation of the development options. We used a Depositional Facies Modelling scheme utilizing recent developments in Multiple Point Geostatistical Simulation and reservoir property uncertainty analysis to generate five static reservoir models. After scale-up, flow simulations were conducted on each model for different field development options using a Design of Experiments (DoE) methodology and a preferred development option was selected. The geology of Mafumeira field is complex. The multiple point geostatistical simulation used a training image consisting of seven depositional facies. The training image is a 3D conceptual model of the facies present and the facies associations; it captures complex spatial relationships between multiple facies, and non-linear shapes such as sinuous channels. The facies simulation was conditioned by a facies probability cube, which permitted the use of a single training image for different stratigraphic intervals of the reservoir, with different combinations and proportions of the seven facies. Multiple versions of the facies probability cube were produced to model the uncertainty in the occurrence of reservoir quality rock units. In modelling the reservoir properties, uncertainties in porosity, permeability and water saturation (‘PKS’) were taken into account. Five models were produced reflecting the combinations of high- and low-case reservoir facies, high- and low-case PKS properties and an intermediate-case. The high-, intermediate- and low-case models were then dynamically tested to ensure different flow behaviours, prior to upscaling, and the flow behaviours compared to analogue producing fields. In order to utilize DoE simulation, upscaling of the five fine-grid models was required. Flow-based simulation was chosen as the best tool to validate the behaviour of the coarse-grid models against the fine-grid models. However, this effort demonstrated that the conventional scale-up methods utilized in other reservoir models did not adequately capture the behaviour of the fine-grid models in this heterogeneous reservoir. A new method that adjusts the Dykstra–Parsons coefficient was investigated and successfully employed to tune the coarse-scale models. Twelve development alternatives for the field were defined and deterministic economics, based on results from the mid-case simulation model, were run in order to narrow down the number of alternatives to be carried forward into probabilistic analysis to five. The DoE approach allowed us to undertake a thorough evaluation of the key subsurface uncertainties and design an overall development plan. The probabilistic simulation results along with full Decision Analysis (DA) allowed us to identify a phased development, which would mitigate potential downside risks while preserving the ability to capture upside potential.

Abstract The Urdaneta West Field is located on the western margin of Lake Maracaibo in northern Venezuela. Biodegraded oil (12–15° API) is reservoired in Tertiary sandstones and produced through a series of lateral wells. The productive sandstones of the Icotea and Misoa Formations are thin, calculated as 0.5–4.5 m (1.5–15 ft) vertical thickness, and of limited lateral extent. This heterogeneity, coupled with heavy oil, results in poor fluid communication and low recovery efficiency. During 2004, a full field review was undertaken to address future development. Despite production from the Icotea and Misoa Formations, subsurface uncertainty was identified as a major issue due to the clustered nature of field development, complex depositional and structural environments, and reservoir fluid characteristics. In order to rank and mitigate reservoir uncertainties, a series of static models were built. The first phase of static modelling used a simple structural framework and reservoir interval averages to generate minimum, mid- and maximum volume cases. Dynamic simulation of these identified two major areas of uncertainty impacting oil-in-place and productivity – net sandbody connectivity and hydrocarbon contact. Two further phases of static and dynamic modelling concentrated on evaluating the full range of these uncertainties within a detailed structural framework.

Abstract Evaluating the static volume potential of a field from a single geological reservoir model can be a risky business. Each piece of input data used to build the model carries an uncertainty that is not expressed in a single deterministic realization. In evaluating the technical and economic feasibility of drilling a new production well on the StatoilHydro operated Glitne Field, a quantified assessment of the range in expected volumes was undertaken. A geological uncertainty study was initiated to identify and quantify the input parameters of greatest impact on static volumetric uncertainty in the reservoir model and to identify potential upsides or downsides that would strongly affect the economics of the potential well target areas. For each geological input parameter, a high-case and low-case scenario was established to capture the end members (approximating to P90–P10) in that parameter uncertainty. IRAP RMS was used in combination with an in-house Microsoft Excel macro together with @Risk to produce a quantitative analysis of the uncertainty range in STOIIP and a ranking of the parameters most affecting the uncertainty in this range. This study has contributed to making a better-informed decision for drilling a new production well on the Glitne Field and thus increasing ultimate recovery and field life further. The workflow used has its limitations but this study shows that a geological uncertainty study can be performed relatively simply using only a limited number of software applications. This study also hopes to highlight the importance of having these studies undertaken by company asset teams as part of their reservoir characterization routines.

Abstract The need for a new Schiehallion full field reservoir simulation model was driven by the requirement to re-evaluate the reserves in the field: the existing model indicated that the modelled volumes were potentially too conservative. This, coupled with a 50% increase in the wells database through ongoing development drilling, was the main reason for building the new model. An integrated multidisciplinary team consisting of BP and Shell staff was set up to build a new full-field reservoir simulation model for reserves re-evaluation. The paper outlines the workflow employed in building the new model, FFM2003, and describes elements of this workflow in more detail, concentrating on lessons learned during the process.