Next Generation Parallel Computing for Large-Scale Reservoir Simulation

Optimal decisions about well placement and control in geologically complex reservoirs can reduce subsurface risks associated with field development plans. Especially, in deepwater developments, economic constraints impose stringent limits on the number of wells that can be drilled without compromising economic profitability. Thus, optimal placement and operation of wells often have a significant impact on project rewards. Reservoir management decisions are made based on evaluations of subsurface models that integrate all available data by use of geomodeling and reservoir simulation. It is important to recognize that uncertainty is inherent to reservoir models due to limited data/knowledge about subsurface. A robust reservoir management strategy calls for rigorous techniques that account for stratigraphic and structural uncertainty in geologically complex deepwater reservoirs. A robust well-location optimization method is developed particularly suitable for such reservoirs. A steepest-ascent algorithm that takes advantage of distributed parallel computation constitutes the computational core of the optimization method. The method operates in parallel on a number of probable reservoir models and maximizes the expected value of ultimate recovery factor while delivering minimum variance over the model uncertainty range. A realistic synthetic test case is evaluated for demonstrating the concept. The case uses alternative geologic models of a lobate turbidite reservoir with complex stratigraphic architecture. The results validate the auspicious features of the robust well location optimization method.

Distinguished Author Series articles are general, descriptive representations that summarize the state of the art in an area of technology by describing recent developments for readers who are not specialists in thetopics discussed. Written by individuals recognized as experts in the area, these articles provide key references to more definitive work and presentspecific details only to illustrate the technology. Purpose: to informthe general readership of recent advances in various areas of petroleumengineering. "Today, desktop computers and a wide range of reservoir. simulation systemsprovide the engineer with an economic means to solve complex reservoir.engineering problems with efficiency." Introduction This paper describes the state of the art in reservoir simulation for thenonexpert, focusing primarily on advances that have occurred since a similarpaper was published by Coats1 in 1982. For completeness, a historyof simulation is given first, followed by a brief and simple review of thesimulation process. Advances in simulation technology are then described, highlighting those in reservoir description and simulation of naturallyfractured reservoirs, hydraulic fracturing, and horizontal wells. The future ofreservoir simulation is discussed last. Many papers have been presented since Coats' review in 1982. Rather thanrepeat the exhaustive list provided in the SPE indices, a reduced set ofpertinent references is given. I offer my apologies in advance to the authorsof the many excellent papers contributed during the past decade that are notmentioned here. History of Simulation In the broadest sense, reservoir simulation has been practiced since thebeginning of petroleum engineering. By the late 1940's, the potential ofreservoir simulation was recognized and several companies embarked on thedevelopment of both analog and numerical models to enhance the existinganalytical solutions, material balances, and ID Buckley-Leverett displacementcalculations. By the late 1950's, major commitments to fundamental research on thenumerical solution of flow equations began to bear fruit. The results werecrude, but certainly useful, computer programs for simulating reservoirs. Theseprograms represented a major advance and used the solution of sets offinite-difference equations to describe 2D and 3D, transient, multiphase flowin heterogeneous-porous media. For the first time, reservoir engineers couldsolve complex problems. The driving force for development of numerical-simulation systems was theability to predict the future performance of a reservoir under various "whatif" conditions so that better reservoir-management decisions could be made. This force still drives the use of reservoir simulation today. Current businessincentives for virtually all reservoir-simulation applications fall into one ormore of the following categories: economics; credibility and reliability;decision-making; arbitration and unitization; performance monitoring;nonelective studies resulting from safety, environmental, or regulatoryconcerns; and communications and personnel utilization.2 Throughout the 1960's, reservoir-simulation development was devoted largelyto three-phase, black-oil reservoir problems. Recovery methods that weresimulated included pressure depletion and various forms of pressuremaintenance. Many computer programs were developed to address most of thegeneral reservoir problems encountered. These programs operated on mainframecomputers and used decks of cards for input. During the 1970's, the picture changed remarkably because of theproliferation of research in EOR processes, advances in numerical-solutiontechniques, and the decreased size and increased speed of computers. Mathematical simulators were developed that extended beyond conventionalprimary and secondary recovery into such areas as chemical flooding, steamflooding, and in-situ combustion. Research during this period resulted insignificant advances in characterizing the physics of hydrocarbon displacementsunder the influences of temperature, chemical agents, and complexmulticomponent phase behavior. Simulators had to reflect chemical adsorptionand degradation, reaction kinetics, interfacial-tension-reduction effects, andmore complex equilibrium phase behavior. The range of reservoir-simulation application continued to expand during the1980's. Reservoir description, finally becoming a major topic on its own, advanced to the use of geostatistics to describe heterogeneities and to providereservoir definition not expressed in reservoir models. The technology to modelnaturally fractured reservoirs, including compositional effects, was alsodeveloped, with extensions into the simulation of hydraulic fracturing andhorizontal wells. and application to such complex processes as reservoirmonitoring.

Reservoir simulation is used in several important tasks in the petroleum industry. Most of such tasks require many simulation runs either because there are uncertainties in reservoir characterization and the model has to be calibrated or because many options have to be tested in manual or optimizing procedures.In the past, computer time was one of the most limiting variables but nowadays computers are becoming less expensive. On the other side, people time is becoming more valuable so they must have tools to make the best use of their minds. Despite of these facts, many important decisions in reservoir engineering are still strongly based on manual procedures even when computers, parallel algorithms and automation tools are available. Furthermore, many companies do not use their network adequately.In order to contribute to a change in this scenario, this work shows (1) some advantages of building tools using reservoir simulation, parallel computing and optimizing techniques to history matching and reservoir management decisions, and (2) how to make the best use of existing networks and PVM (parallel virtual machine) to accelerate such processes.A case study using an offshore field is presented to illustrate (1) the advantages of having a computer program to automate some steps of production history matching and (2) the time savings obtained by parallel computing. Some procedures are discussed to obtain the best balance between people and computer time.

This paper discusses the role reservoir simulators play in formulating initial development plans, history matching and optimising future production and in planning and designing enhanced oil recovery projects. The Hibernia Field in Canada and the Hassi R'Mel in Algeria illustrate how simulation can be used to assist in initial reservoir development. The Lookout Butte Rundle (Alberta) and others are cited to exemplify optimisation of future production plans with the aid of simulation. Finally, applications to several reported EOR projects are briefly discussed with major emphasis concentrating on the Bati Raman Field in Turkey.

Technology Focus As the development of shale oil and gas becomes increasingly significant, so does the need for modeling tools for their accurate and timely forecasting. The basic question then arises: Are the simulation tools that we have suitable for the job? Can they explain, and ultimately forecast, the contribution to flow from natural or hydraulically induced fractures? Ultimately, we need to be confident that the results from such models are sufficiently reliable for decision-making purposes. As Alexandre Emerick stated in JPT (April 2014), “The goal is to generate models for production forecasting aiding the decision-making process involved in the development and management of petroleum reservoirs.” The ability to model this increasingly important asset class is one of the primary issues facing reservoir and simulation engineers today. Our existing arsenal of simulation tools is certainly time-tested for conventional assets, but can they deliver reliable forecasts for this emergent class? Recent literature is rich with relevant studies ranging from fundamental laboratory experiments to entirely empirical, data-driven, approaches for predictive reservoir management. These articles raise numerous questions, including the possible oversimplification of a complex problem, the very nature of the production mechanisms, and the role played by any pre-existing discrete fracture networks. Additional concerns relating to accuracy and reliability of the underlying data used to populate such models have also been raised. Keeping an open mind on this topic is, I feel, appropriate. One needs to balance the very real need for such predictive tools today with the recognition that future evidence may disrupt certain preconceptions. To this end, four articles have been selected that span a spectrum of ideas. One considers practical applications and solutions, while another covers the emerging technique of molecular simulation. Another eloquently probes various notions in our current state of understanding of shale assets and examines issues in interpreting the forecasts furnished by our simulation models. The fourth article demonstrates some creative thinking by representing a planar fracture by use of a proxy comprising a coupled flowing network model, a concept that may prove flexible and warrants further investigation. Conventional modeling tools used to predict flow in tight, fractured shales (e.g., dual-porosity models, local grid refinements, tartan grids) may, indeed, provide reliable forecasting. Analytic and data-driven approaches are also acceptable in some circumstances. However, as our knowledge in exploiting these assets deepens with new theory, laboratory experiments, and field and operational experience, we should be prepared to reconsider some of today’s axioms in the future.

Modelling the impact of Alkaline-surfactant and Alkaline-surfactant-polymer flooding processes on scale precipitation and management

Drilling cost is one of the most critical aspects in the reservoir management plan. Costs may exceed a million dollars; thus, optimal designing of the well trajectory in the reservoir and completion are essential. The implementation of a multilateral well (MLW) in reservoir management has great potential to optimize oil production. The object of this study is to develop an integrated workflow of end-point multilateral well placement optimization integrated with the reservoir simulator supported by artificial intelligence (AI) methods. The paper covers various types of MLW construction, including: horizontal, bi-, tri-, and quad-lateral wells. For quad-lateral wells, the capital expenditure is highest; nevertheless, acceleration of oil production affects the project’s NPV (net present value), indicating the type of well that is most promising to implement in the reservoir. Tri- and quad-lateral wells can operate for 12.1 and 9.8 years with a constant production rate. The decreasing hydrocarbon production rate is affected by reservoir pressure and the reservoir water production rate. Other well design patterns can accelerate water production. The well’s optimal trajectory was evaluated by the genetic algorithm (GA) and particle swarm optimization (PSO). The major difference between the GA and PSO optimization runs is visible with respect to water production and is related to the modification of one well branch trajectory in a reservoir. The proposed methodology has the advantage of easy implementation in a closed-loop optimization system coupled with numerical reservoir simulation. The paper covers the solution background, an implementation example, and the model limitations.

Traditional reservoir simulators cannot capture the complicated interactions between fluid production and reservoir rock deformation during hydrocarbon recovery. In particular, in a recovery process when both phase behavior and deformation of reservoir rock play critical roles, a coupled geomechanics and compositional reservoir model can rigorously capture the above physical relations between solid and fluids and thereby perform more precise history matchings and predictions for better well planning and reservoir management decisions. This paper considers a fully coupled geomechanics and compositional modeling process. A novel solution procedure for the associated fully coupled Jacobian system is presented using the Schur complement technique. The proposed approach has the following features: (1) it has similar convergence properties as the fully implicit method, (2) combined with Krylov subspace iterative solvers, it can deal with symmetric and unsymmetric problems, and (3) it can be easily parallelized to solve large-scale reservoir simulation problems. In this work, a parallel, fully implicit, equation of state compositional reservoir simulator was fully coupled with a geomechanics FEM module. For each Newton iteration, an original two by two block Jacobian system is converted to a Schur complement system, and then the associated Schur complement system is solved using a BiCGstab(l) iterative method; subsequently all the unknowns for the reservoir simulator and the geomechanics module are obtained at the same time. The above partition solution procedure is equivalent to a block Gaussian elimination; hence its convergence is similar to a fully implicit coupled procedure. The proposed approach is implemented to solve two- and three-dimensional compositional reservoir simulation case studies considering geomechanics effects. We also compare the results between the fully coupled and the iteratively coupled approaches. It is shown that this approach presents a viable alternative in solution methods for solving general coupled physical problems.

Reservoir simulation is becoming increasingly complex because of more advanced wells in the fields, including intelligent and multilateral wells. Advanced completions are also evolving to increase recovery efficiencies. This increasing complexity presents two difficulties, which include the design of advanced completions within reservoir simulators and increased simulation runtime. To describe a well in a reservoir model, a reservoir engineer typically defines a network of hundreds of nodes using keywords and specifies properties for each node. This is a cumbersome and error-prone process. Additionally, detailed well models can slow down reservoir simulation and often cause poor convergence. A new iterative round-trip approach has been implemented, in which an engineer imports an initial reservoir model into a nodal analysis simulator that models flow from the reservoir through complex completions to the wellhead. The simulator accurately models well production in steady state and designs completion strings in detail. After the design is complete, the nodal analysis simulator converts the well model into reservoir simulator keywords that are imported into full-scale simulations for transient analysis. Using this method, reservoir simulators can also model multiple annuli, which was not feasible previously. Highly detailed well models of several thousand nodes that accurately describe completion strings can be generated automatically. Reservoir engineers typically do not possess complex knowledge of well design because this is usually performed by completion/production engineers, who seldom have access to a reservoir simulator. Consequently, they have a limited ability to experiment with different well designs. This paper presents an approach that helps facilitate reservoir and production engineer collaboration, thus helping enable fine-tuning of final completion designs to maximize production, prevent early water/gas breakthrough, and increase overall recovery. This paper describes the application of the new process in an openhole well and presents various completion designs of the same well.

Hydrocarbon saturation below present-day free water levels (FWL) is observed in many fields and can be attributed to either past geologic events or more recent production-related pressure changes and fluid movement. However, modeling residual oil saturation (Sorw) below a present-day FWL is a topic of limited understanding within reservoir characterization, particularly in evaluating the impact on water injection efficiency and reserves uncertainty in dynamic reservoir modeling. The presence of significant volumes of residual oil below the FWL can have profound implications for reserves estimations and development scenarios in field development plans. This study, inspired by a recent oil discovery in clastic sediments on the Norwegian Continental Shelf, explores this phenomenon through an integrated approach that spans from core and log scale analyses to field-scale assessment. This paper will present a case-study showing the recognition of a significant paleo-oil column based on petrophysical logs and core data, and both the definition of a present-day FWL and prediction of a paleo FWL. This paper will further describe a versatile technique for modeling residual oil saturation below the current FWL by integrating geological and engineering data from routine (RCA) and special core analyses (SCAL), well logs, and 3D reservoir models. The methodology begins with a rigorous quality check of core analysis data and other relevant information to ensure accuracy and suitability of the data used. We then applied a novel rock typing technique based on similarities in pore geometry and structure (PGS), incorporating RCA and SCAL data. These were integrated into a newly developed Corey SCAL modeling method to establish static and dynamic saturation functions for each rock type, which were then used in 3D reservoir modeling and simulation. Subsequently, we estimated the SCAL uncertainty range for each rock type and integrated it into the reference 3D static and dynamic models. After assimilating all input data into the 3D reservoir models, production forecasts were generated, including sensitivity and uncertainty workflows. Additionally, we developed new water saturation model workflows for static modeling by implementing saturation depth functions (SDF) with established Pc-imbibition curves. Furthermore, a fishbone completion model for water injection wells was constructed, and production was forecasted under the selected scenario. A key result of this work is effectively modelling residual oil saturation below a current FWL by implementing PGS rock typing and a saturation depth function based on Pc- imbibition in both static and dynamic models. The presence of oil saturation below the current FWL matched the observation from core and logs and increased Initial Oil in Place (IOIP) by approximately 25% compared to the original case without residual oil modeling. Meanwhile, this also reduced the recoverable oil volume by approximately 11% compared to the reference case primarily due to decreased injectivity in water injection following relative permeability corrections based on SCAL modeling techniques. To mitigate this negative impact, we evaluated methods for improving injectivity using a fishbone completion model in water injector wells during reservoir simulation, which resulted in an incremental oil recovery of up to 9% compared to the base case scenario of this study. As a novelty, this study proposes a new solution for estimating oil saturation below the current FWL based on saturation depth functions that are consistent with rock typing and wettability physics derived from imbibition capillary pressure curves. Additionally, the value of SCAL modeling, particularly in capturing the uncertainty in reserves evaluation through relative permeability and PGS rock types, proves suitable for full-field reservoir simulations, uncertainty analyses, and field performance assessments. This workflow also facilitates permeability prediction as a function of porosity and irreducible water saturation while generating systematic saturation functions based on capillary pressure data.

Quick evaluation of reservoir performance is one of the main concern in decision making. Time consuming data preparation and processing, and data uncertainty (geological, petrophysical, reservoir engineering) limit the use of numerical simulators in addition to long term response for reservoir management. Effective reservoir management needs quick action regarding injected fluid distribution to improve areal and vertical sweep efficiency during secondary and tertiary recovery processes. Therefore, simpler methods that provide quick results to complement or substitute reservoir simulation are important for reservoir monitoring and management. Capacitance Resistance Model (CRM) is one proven method to address the above challenges. The CRM model is based on the hypothesis that reservoir performance can be inferred from analyzing production and injection data and a simplified analytic model structure. CRM is an input-output and material balance based model that use continuity equation to quantify the connectivity among injector and producer wells, time constant and productivity indices. In this study, a CRM model was generated to build a forecasting waterflood model by to matching historic production and injection data. After history matching, the model can be used to: (1) effectively forecast oil and water production, (2) evaluate reservoir and production-injection data uncertainties, and (3) optimize injection settings to improve oil recovery factor. CRM methodology is used as a cooperative approach to reservoir simulator generated streamlines. Different case studies were designed to investigate the robustness of CRM by providing different levels of complexities as compared with numeric simulation results. In the first case, a constant flowing bottomhole pressure and injection rate was considered, and in the second case, changes to flowing bottomhole pressure, injection rate and productivity index per each time step were considered. The CRM was able to replicate the synthetic historic data generated using numerical simulator within a satisfactory error, i.e. less than 7% in the worst case. The CRM injector producer allocation factors were also in line to those computed using streamlines. In both cases, the CRM model showed a higher allocation factor for the injector producer pairs which had better connectivity allowing the evaluation of waterflood performance by pattern while identifying the poor sweep efficiency areas. In a later optimization exercise, a dry oil production objective function was maximized subject to the constraints in the injection system. Thus, strategies derived from CRM model increased production in the range of 12% with associated water cut reduction just by reallocating injection rates.

Summary Grid selection is one of the most difficult and time-consuming tasks in the simulation of geologically complex reservoirs. A grid is imposed on a reservoir so that we can solve the nonlinear flow equations that predict the reservoir's response to changes at wells or other boundaries. Therefore, I first discuss the role of grid geometry in the evaluation of each term in the flow equations. Then I describe several conventional and some new gridding techniques, as well as their advantages and disadvantages. Next I make recommendations based on my own experience and discussions with colleagues. My assessment is that grids should be as close to orthogonal as possible. Also, local grid refinement should be used with care; its inappropriate use can be counterproductive. Considerable progress has been made recently on the use of flexible grids, but efficient use of unstructured, locally orthogonal grids requires further research. Techniques that automatically align the grid with major reservoir features (e.g., faults) are being developed and are expected to simplify grid generation. Introduction A reservoir simulator predicts reservoir performance by solving flow equations on a discrete grid chosen by the simulation engineer to represent the reservoir. The grid normally is selected with one or more of the following considerations in mind.Reservoir geology and size and the data available for reservoir description.Type of fluid displacement or depletion process to be modeled.Past and anticipated field development (location and type of wells).Numerical accuracy desired.Available software options.Objectives of the simulation study.Competence of the simulation engineer or team.Available computer resources, time constraints, or project budget. In the early days of reservoir simulation, the last consideration often determined the number of gridblocks, and the available software limited the choice of grid type, usually to block-centered Cartesian or cylindrical grid. Research in reservoir simulation and hardware developments, especially over the past 10 years, has greatly extended the gridding options available.

Three-dimensional, multiphase flow problems are solved on static, composite grids. The composite grid consists of a global, rectangular coarse grid and a set of locally refined patches around active wells or other areas with important local flow properties. Curvilinear (locally orthogonal) refinements can be utilized to approximate radial or near-radial flow around wells if such flow is assumed to be appropriate. More general localized refinement can be utilized otherwise. Finite differencing is presented which connects the local and global grid cells in an accurate manner. Other methods for composite grid applications which have appeared in the SPE literature generate linearizations which have lost their banded structure and must utilize corresponding solution algorithms that are extremely difficult to vectorize for efficient solution. Preconditioned iterative techniques will be discussed which allow easy implementation of the composite grid techniques in existing simulators with full vectorization capabilities. Pressure calculations in a single-phase flow problem with a distribution of injection and production wells exhibit excellent agreement between numerical calculations on locally radial grids and analytical solutions. Water and gas coning results for a single well also show very good agreement between computations on a hybrid grid and a circular cylindrical grid. This indicates the potential for incorporating coning models around any well in a full field-scale application without destroying the efficiency of the original simulator.

Integrated asset modelling allows coupling fluid behavior in the full production system, from the reservoir to the facilities. This approach is crucial when different field development scenarios need to be evaluated during concept selection. Different situations must be systematically examined allowing for the optimal operational constraints (rate and pressure limits) and the dimensions of individual elements to be defined, whilst ensuring that material balance considerations are honoured. Traditionally, reservoir, production and facilities groups separately conduct field modelling and simulations for each element in the system, as part of an iterative process. Reservoir engineers receive facilities constraints and lift performance curves from production engineers as an input. In return, production and facilities engineers receive production profiles from reservoir engineers and design the appropriate facilities capable of safely processing and transporting the produced fluids throughout the lifetime of the project. Any changes in production configuration affect all the models and leads to a need for each discipline to revaluate their simulation models. This process often results in approximations being adopted and frequently raises challenges in the audit trail, as assumptions may be inadequately documented. Integrated simulation generates representative field production forecasts and the ability to quantify cumulative production and key performance indicators for each scenario. As an example, rates, erosional velocities, pressures and temperatures profiles, compressor and pump requirements and many other variables may be explicitly studied throughout the entire life of the field. This approach reduces the risk of facilities bottlenecking and flow assurance becoming an HSE issue. This paper describes some of the integrated asset management studies conducted to evaluate different strategies for the development of an ultra-deep water exploratory block, and the key technical factors considered. For each concept studied, multi-disciplinary models were created to simulate reservoir, tubing, flowline and surface network behaviour for gas condensate, light oil and water production streams. Subsea separation and surface process facilities options were modelled to coordinate the work streams of the different engineering domains, in a coupled solution. Some complex development scenarios were included. A gas-reinjection case was modelled, where the separated gas from process simulation model is reinjected into the reservoir. Another scenario considered an ultra-deep subsea separation and boosting unit, which was modelled by process simulation software, connected to both the surface network and the reservoir simulation. All were linked with the FPSO process simulation. A range between 5 to 17% of oil gain was obtained in the different scenarios. As a result of these studies, the decision making process for future investments in the field development plan was optimized.

Evaluation of the economic benefit of field development projects in the oil and gas industry is naturally subject to subsurface uncertainties and risks. While high economic returns from produced volumes compensated for the lack of robust uncertainty estimates in the past, lower margins in tight market conditions and more complex project designs require increased efforts in delivering reliable uncertainty estimates for improving reservoir management decision support in times to come. An increasing number of field development projects include rigorous uncertainty quantification workflows based on parameterized subsurface uncertainties. Estimation and reduction of prediction uncertainty has no value unless it can potentially influence a decision process. This requires an integrated workflow design across disciplines and a proper risk mitigation plan. This work presents the implementation of an uncertainty workflow in reservoir management with a focus on reservoir modeling and simulation. Practical workflow design steps and intermediate milestones are discussed for generating alternative reservoir model realizations including historical production data. Calibrated model realizations define the basis for prediction estimates of a predefined field development plan. Practical steps are discussed with application to a complex oil rim development for an existing reservoir with long production history. This study aimed at an uncertainty assessment of a field development plan and a proposal for optimized relocations of infill wells considering multiple geological realizations. In conclusion, this work describes examples, experiences, and recommendations for an industry implementation for a risk assessment of a field development plan under subsurface uncertainties.