There are two big challenges which restrict the extensive application of fully coupled geomechanics-reservoir modeling. The first challenge is computational effort. Consider a 3-D simulation combining pressure and heat diffusion, elastoplastic mechanical response, and saturation changes; each node has at least 5 degrees of freedom, each leading to a separate equation. Furthermore, regions of large p, T and σ′ gradients require small-scale discretization for accurate solutions, greatly increasing the number of equations. When the rock mass surrounding the reservoir region is included, it is represented by many elements or nodes. These factors mean that accurate analysis of realistic 3-D problems is challenging, and will so remain as we seek to solve larger and larger coupled problems involving nonlinear responses. To overcome the first challenge, the displacement discontinuity method is introduced wherein a large-scale 3-D case is divided into a reservoir region where Δp, ΔT and non-linear effects are critical and analyzed using FEM, and an outside region in which the reservoir is encased where Δp and ΔT effects are inconsequential and the rock may be treated as elastic, analyzed with a 3D displacement discontinuity formulation. This scheme leads to a tremendous reduction in the degrees of freedom, yet allows for reasonably rigorous incorporation of the reactions of the surrounding rock. The second challenge arises from some forms of numerical instability. There are actually two types of sharp gradients implied in the transient advection-diffusion problem: one is caused by the high Peclet numbers, the other by the sharp gradient which appears during the small time steps due to the transient solution. The way to eliminate the spurious oscillations is different when the sharp gradients are induced by the transient evolution than when they are produced by the advective terms, and existing literature focuses mainly on eliminating the spurious spatial temperature oscillations caused by advection-dominated flow. To overcome the second challenge, numerical instability sources are addressed by introducing a new stabilized finite element method, the subgrid scale/gradient subgrid scale (SGS/GSGS) method.

Modern reservoir management typically involves simulations of geological models to predict future recovery estimates, providing the economic assessment of different field development strategies. Integrating reservoir data is a vital step in developing reliable reservoir performance models. Currently, most effective strategies for traditional manual history matching commonly follow a structured approach with a sequence of adjustments from global to regional parameters, followed by local changes in model properties. In contrast, many of the recent automatic history matching methods utilize parameter sensitivities or gradients to directly update the fine-scale reservoir properties, often ignoring geological inconsistency. Therefore, there is need for combining elements of all of these scales in a seamless manner. We present a hierarchical streamline-assisted history matching, with a framework of global-local updates. A probabilistic approach, consisting of design of experiments, response surface methodology and the genetic algorithm, is used to understand the uncertainty in the large-scale static and dynamic parameters. This global update step is followed by a streamline-based model calibration for high resolution reservoir heterogeneity. This local update step assimilates dynamic production data. We apply the genetic global calibration to unconventional shale gas reservoir specifically we include stimulated reservoir volume as a constraint term in the data integration to improve history matching and reduce prediction uncertainty. We introduce a novel approach for efficiently computing well drainage volumes for shale gas wells with multistage fractures and fracture clusters, and we will filter stochastic shale gas reservoir models by comparing the computed drainage volume with the measured SRV within specified confidence limits. Finally, we demonstrate the value of integrating downhole temperature measurements as coarse-scale constraint during streamline-based history matching of dynamic production data. We first derive coarse-scale permeability trends in the reservoir from temperature data. The coarse information are then downscaled into fine scale permeability by sequential Gaussian simulation with block kriging, and updated by local-scale streamline-based history matching. he power and utility of our approaches have been demonstrated using both synthetic and field examples.

This thesis investigates methane emissions and methane production potentials from abandoned longwall panels produced or emitted due to mining activities either from coal seam or any underlying or overlying formations. These emissions can increase greenhouse gas concentrations and also pose a danger to the underground working environment and to miners. In addition to the safety issues, recovery and utilization of this gas is an additional source of energy. In this study, methane concentrations measured from ventilation air ways in Yeni &Ccedil / eltek Coal Mine, which is located in Suluova basin, Amasya, and contains thick, laterally extensive Lower Eocene coal seams, were integrated within a numerical vi reservoir model. Key reservoir parameters for history matching are cleat permeabilities, cleat porosity, diffusion time and Langmuir volume and Langmuir pressure. Thirteen cases were studied. According to the results, Case-10 determined as the best fitted case for both of the production wells. Cleat permeabilities and Langmuir pressure were the most effective parameters. Reservoir parameters matched are cleat permeabilities of 5, 4 and 1 md and fracture dimensions of 0.8, 0.4, and 0.1 m in x, y and z direction respectively, 2 % cleat porosity, 0.3 % water saturation. Diffusion time was determined as 400 days and 2000 kPa Langmuir volume and 6.24279 m3 /tone gas content estimated. According to these results it can be said that methane production will not be economically feasible, however / to remedy underground working conditions and safety of workers methane management should be taken into consideration.

We connect the Mimetic Finite Difference method (MFD) with the finite-volume two-point flux scheme (TPFA) for Voronoi meshes. The main effect is reducing the saddle-point system to a much smaller symmetric-positive definite matrix. In addition, the generalization allows MFD to seamlessly integrate with existing porous media modeling technology. The generalization also imparts the monotonicity property of the TPFA method on MFD. The connection is achieved by altering the consistency condition of the velocity bilinear operator. First-order convergence theory is presented as well as numerical results that support the claims. We demonstrate a methodology for using MFD in modeling fluid flow in fractures coupled with a reservoir. The method can be used for nonplanar fractures. We use the method to demonstrate the effects of fracture curvature on single-phase and multi-phase flows. Standard benchmarks are used to demonstrate the accuracy of the method. The approach is coupled with existing reservoir simulation technology. / text

A review of the developments within the field of structural reliability theory shows that some gaps still exist in the reliability prediction process and hence there is an urgent desire for improvements such that the estimated structural reliability will be capable of expressing a physical property of the given structure. The current reliability prediction process involves the continuous estimation and use of reliability index as a way of estimating the safety of any given structure. The reliability index β depends on the Probability Density Function (PDF) distribution for the wave force and the corresponding PDF of resistance from respective structural members of the given structure. The PDF for the applied wave force will depend on the PDF of water depth, wave angular velocity and wave direction hence the reliability index as currently practiced is a statistical way of managing uncertainties based on a general probabilistic model. This research on Smart Offshore Structure for Reliability Prediction has proposed the design of a measurement based reliability prediction process as a way of closing the gap on structural reliability prediction process. Structural deflection and damping are some of the measurable properties of an offshore structure and this study aims at suggesting the use of these measurable properties for improvements in structural reliability prediction process. A design case study has shown that a typical offshore structure can deflect to a range of only a few fractions of a millimetre. This implies that if we have a way of monitoring this level of deflection, we could use the results from such measurement for the detection of a structural member failure. This advocated concept is based on the hypothesis that if the original dynamic characteristics of a structure is known, that measurement based modified dynamic properties can be used to determine the onset of failure or failure propagation of the given structure. This technology could reveal the location and magnitude of internal cracks or corrosion effects on any given structure which currently is outside the current probability based approach. A simple economic analysis shows that the recommended process shows a positive net present value and that some $74mln is the Value of Information for any life extension technology that could reveal the possibility of extending the life of a given 10,000bopd production platform from 2025 to 2028.

Miscible gas flooding is a common method for enhanced oil recovery. Reliable design of miscible gas flooding requires compositional reservoir simulation that can accurately predict the fluid properties resulting from mass transfer between reservoir oil and injection gas. Drawbacks of compositional simulation are the efficiency and robustness of phase equilibrium calculations consisting of flash calculations and phase stability analysis. Simulation of multicontact miscible gas flooding involves a large number of phase equilibrium calculations in a near-critical region, where the calculations are time-consuming and difficult. Also, mixtures of reservoir oil and solvent such as CO₂ and rich gas can exhibit complex phase behavior at temperatures typically below 120°F, where three hydrocarbon-phases can coexist. However, most compositional simulators do not attempt to solve for three hydrocarbon-phases because three-phase equilibrium calculations are more complicated, difficult, and time-consuming than traditional two-phase equilibrium calculations. Due to the lack of robust algorithms for three-phase equilibrium calculations, the effect of a third hydrocarbon-phase on low-temperature oil displacement is little known. We develop robust and efficient algorithms for phase equilibrium calculations for two and three phases. The algorithms are implemented in a compositional reservoir simulator. Simulation case studies show that our algorithms can significantly decrease the computational time without loss of accuracy. Speed-up of 40% is achieved for a reservoir simulation using 20 components, compared to standard algorithms. Speed-up occurs not only because of improved computational efficiency but also because of increased robustness resulting in longer time-step sizes. We demonstrate the importance of three-phase equilibrium calculations, where simulations with two-phase equilibrium approximations proposed in the literature can result in complete failure or erroneous simulation results. Using the robust phase equilibrium algorithms developed, the mechanism is investigated for high efficiency of low-temperature oil displacements by CO₂ involving three hydrocarbon-phases. Results show that high displacement efficiency can be achieved when the composition path goes near the critical endpoint where the gaseous and CO₂-rich liquid phases merge in the presence of the oleic phase. Complete miscibility may not be developed for three-phase flow without considering the existence of a tricritical point. / text

Made available in DSpace on 2014-12-17T14:08:32Z (GMT). No. of bitstreams: 1 JoseTMAP.pdf: 3429155 bytes, checksum: 9d06d453f156a78546476b9e1a22c995 (MD5) Previous issue date: 2008-12-15 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / Until the early 90s, the simulation of fluid flow in oil reservoir basically used the numerical technique of finite differences. Since then, there was a big development in simulation technology based on streamlines, so that nowadays it is being used in several cases and it can represent the physical mechanisms that influence the fluid flow, such as compressibility, capillarity and gravitational segregation. Streamline-based flow simulation is a tool that can help enough in waterflood project management, because it provides important information not available through traditional simulation of finite differences and shows, in a direct way, the influence between injector well and producer well. This work presents the application of a methodology published in literature for optimizing water injection projects in modeling of a Brazilian Potiguar Basin reservoir that has a large number of wells. This methodology considers changes of injection well rates over time, based on information available through streamline simulation. This methodology reduces injection rates in wells of lower efficiency and increases injection rates in more efficient wells. In the proposed model, the methodology was effective. The optimized alternatives presented higher oil recovery associated with a lower water injection volume. This shows better efficiency and, consequently, reduction in costs. Considering the wide use of the water injection in oil fields, the positive outcome of the modeling is important, because it shows a case study of increasing of oil recovery achieved simply through better distribution of water injection rates / At? o in?cio da d?cada de 90, a simula??o de fluxo em reservat?rio de petr?leo basicamente utilizava a t?cnica num?rica de diferen?as finitas. Desde ent?o, houve um grande desenvolvimento da tecnologia de simula??o com base em linhas de fluxo, de modo que hoje ela tem sido usada em v?rios casos e se encontra adequada para representar os diversos mecanismos f?sicos que influenciam o fluxo de fluidos no reservat?rio, tais como compressibilidade, capilaridade e segrega??o gravitacional. A simula??o baseada em linhas de fluxo ? uma ferramenta que pode ajudar bastante no gerenciamento de projetos de inje??o de ?gua, pois fornece informa??es importantes n?o disponibilizadas atrav?s da simula??o tradicional de diferen?as finitas e mostra, de forma direta, a influ?ncia entre po?o injetor e po?o produtor. Este trabalho apresenta a aplica??o de uma metodologia existente na literatura para otimizar projeto de inje??o de ?gua em um reservat?rio pertencente a um campo de petr?leo da Bacia Potiguar que possui elevado n?mero de po?os. Esta metodologia considera modifica??es das cotas de inje??o dos po?os ao longo do tempo, com base nas informa??es dispon?veis atrav?s da simula??o por linhas de fluxo. Esta metodologia reduz as vaz?es de inje??o dos po?os de menor efici?ncia e aumenta as vaz?es de inje??o dos po?os mais eficientes. No modelo proposto, a metodologia se mostrou eficaz. As alternativas de otimiza??o apresentaram uma maior recupera??o de ?leo associada a um menor volume de inje??o de ?gua. Isto demonstra uma melhor efici?ncia e, consequentemente, uma redu??o de custos. Considerando a ampla utiliza??o da inje??o de ?gua em campos de petr?leo, o resultado positivo da modelagem ? importante, porque apresenta um estudo de caso de aumento da recupera??o de petr?leo, alcan?ado simplesmente atrav?s de uma melhor distribui??o das vaz?es de inje??o de ?gua

Made available in DSpace on 2014-12-17T14:08:32Z (GMT). No. of bitstreams: 1 TommyAP_capa_ate_cap4.pdf: 809913 bytes, checksum: 0090aa039ef326f260cb40945133e6d1 (MD5) Previous issue date: 2009-04-27 / Petr?leo Brasileiro SA - PETROBRAS / The gas injection has become the most important IOR process in the United States. Furthermore, the year 2006 marks the first time the gas injection IOR production has surpassed that of steam injection. In Brazil, the installation of a petrochemical complex in the Northeast of Brazil (Bahia State) offers opportunities for the injection of gases in the fields located in the Rec?ncavo Basin. Field-scale gas injection applications have almost always been associated with design and operational difficulties. The mobility ratio, which controls the volumetric sweep, between the injected gas and displaced oil bank in gas processes, is typically unfavorable due to the relatively low viscosity of the injected gas. Furthermore, the difference between their densities results in severe gravity segregation of fluids in the reservoirs, consequently leading to poor control in the volumetric sweep. Nowadays, from the above applications of gas injection, the WAG process is most popular. However, in attempting to solve the mobility problems, the WAG process gives rise to other problems associated with increased water saturation in the reservoir including diminished gas injectivity and increased competition to the flow of oil. The low field performance of WAG floods with oil recoveries in the range of 5-10% is a clear indication of these problems. In order to find na effective alternative to WAG, the Gas Assisted Gravity Drainage (GAGD) was developed. This process is designed to take advantage of gravity force to allow vertical segregation between the injected CO2 and reservoir crude oil due to their density difference. This process consists of placing horizontal producers near the bottom of the pay zone and injecting g?s through existing vertical wells in field. Homogeneous models were used in this work which can be extrapolated to commercial application for fields located in the Northeast of Brazil. The simulations were performed in a CMG simulator, the STARS 2007.11, where some parameters and their interactions were analyzed. The results have shown that the CO2 injection in GAGD process increased significantly the rate and the final recovery of oil / A inje??o de g?s tem sido o m?todo de recupera??o mais utilizado nos Estados Unidos e o ano de 2006 marca a primeira vez em que a produ??o proveniente dos processos de inje??o de g?s superou os de inje??o de vapor. No Brasil, a instala??o do complexo petroqu?mico, no estado da Bahia, oferece oportunidades para inje??o de g?s na Bacia do Rec?ncavo. As aplica??es de inje??o de g?s em campo foram quase sempre associadas com dificuldades operacionais e de projetos. A raz?o de mobilidade, que controla a varredura volum?trica, entre o g?s injetado e o banco de ?leo deslocado, ? tipicamente desfavor?vel devido ? baixa viscosidade do g?s. Al?m de que a diferen?a entre as densidades conduz a segrega??o gravitacional severa dos fluidos no reservat?rio e consequentemente, levando a um pobre controle na varredura volum?trica. Atualmente, grande parte dos projetos de inje??o de g?s no mundo ? do tipo WAG. Entretanto, na tentativa de resolver os problemas de mobilidade, o processo WAG causa outros problemas como o aumento na satura??o de ?gua no reservat?rio, incluindo a diminui??o na injetividade do g?s e aumento da competi??o no fluxo do ?leo. O baixo desempenho na recupera??o de ?leo deste processo no campo, na ordem de 5 a 10%, ? uma clara indica??o destes problemas. A fim de encontrar uma alternativa eficaz para o WAG, o processo de drenagem gravitacional assistida por inje??o de g?s (GAGD) foi desenvolvido. O processo GAGD ? projetado para tomar vantagem da for?a gravitacional causando assim, a segrega??o vertical entre o CO2 injetado e a reserva de ?leo devido ? diferen?a de densidades. Este processo consiste na coloca??o de po?os produtores horizontais pr?ximos a parte inferior da zona de ?leo e inje??o de g?s atrav?s de po?os verticais existentes em campo. Modelos homog?neos foram utilizados neste trabalho os quais podem ser extrapolados para aplica??o comercial em campos localizados no nordeste brasileiro. As simula??es foram realizadas no simulador da CMG STARS 2007.11, sendo analisados alguns par?metros e suas intera??es. Os resultados mostraram que a inje??o de CO2 no processo GAGD aumentou significativamente a vaz?o e a recupera??o final de ?leo

Made available in DSpace on 2014-12-17T14:08:38Z (GMT). No. of bitstreams: 1 Janaina Medeiros Dantas Barbosa_DISSERT_01_35.pdf: 3284535 bytes, checksum: 7cd819680331f233db6c8299e7522b7b (MD5) Previous issue date: 2009-06-12 / Petr?leo Brasileiro SA - PETROBRAS / Continuous steam injection is one of heavy oil thermal recovery methods used in the Brazilian Northeast because of high occurrence of heavy oil reservoir. In this process, the oil into the reservoir is heated while reduces, substantially, its viscosity and improves the production. This work analyzed how the shaly sand layers influenced in the recovery. The studied models were synthetics, but the used reservoir data can be extrapolated to real situations of Potiguar Basin. The modeling was executed using the STARS - Steam Thermal and Advanced Process Reservoir Simulator - whose version was 2007.10. STARS is a tool of CMG Computer Modeling Group. The study was conducted in two stages, the first we analyzed the influence of reservoir parameters in the thermal process, so some of these were studied, including: horizontal permeability of the reservoir and the layer of shaly sand, ratio of horizontal permeability to vertical permeability, the influence of capillary pressure layer of shaly sand and as the location and dimensions of this heterogeneity can affect the productivity of oil. Among the parameters studied the horizontal permeability of the reservoir showed the most significant influence on the process followed by diversity. In the second stage three models were selected and studied some operational parameters such as injection rate, distance between wells, production time and completion intervals. Among the operating parameters studied the low rate and intermediate distances between wells showed the best recoveries / A inje??o cont?nua de vapor ? um dos m?todos t?rmicos de recupera??o de ?leos pesados utilizados no Nordeste Brasileiro, devido ? grande ocorr?ncia de reservat?rios contendo este tipo de ?leo. Este processo consiste em aquecer o ?leo existente no reservat?rio diminuindo substancialmente a sua viscosidade melhorando assim, a sua produ??o. Neste trabalho foi estudada qual a influ?ncia da presen?a de camadas de areia argilosas no reservat?rio para este tipo de recupera??o. Os modelos s?o sint?ticos, por?m com dados de reservat?rios que podem ser extrapolados para situa??es de aplica??es pr?ticas na Bacia Potiguar que apresenta reservat?rios com essa heterogeneidade. A modelagem foi implementada no simulador de recupera??o t?rmica STARS (Steam, Thermal, and Advanced Processes Reservoir Simulator) vers?o 2007.10 da CMG (Computer Modelling Group). O estudo foi realizado em duas etapas, na primeira foi analisada a influ?ncia dos par?metros de reservat?rio neste processo t?rmico, sendo assim, alguns desses foram estudados, dentre eles: permeabilidade horizontal do reservat?rio e da camada de areia argilosa, rela??o da permeabilidade horizontal com a permeabilidade vertical, influ?ncia da press?o capilar da camada de areia argilosa e como a localiza??o e as dimens?es dessa heterogeneidade podem afetar a produtividade de ?leo. Dentre os par?metros estudados a permeabilidade horizontal do reservat?rio foi o que apresentou maior influ?ncia significativa ao processo seguido da heterogeneidade. Na segunda etapa foram selecionados tr?s modelos e estudado alguns par?metros operacionais, tais como: vaz?o de inje??o, dist?ncia entre po?os, tempo de produ??o e intervalos de completa??o. Dentre os par?metros operacionais estudados as baixas vaz?es e as dist?ncias entre po?os intermedi?rias foram as que apresentaram as melhores recupera??es

[pt] Os modelos de simulação de reservatórios estão sujeitos à incerteza presente em uma grande variedade de seus parâmetros de entrada. Esta incerteza é o resultado da heterogeneidade das formações geológicas, erros nas medições dos dados e da modelagem petrofísica, estrutural e do transporte dos fluidos no meio poroso. Uma quantificação precisa da incerteza requer, na maioria dos casos, uma quantidade elevada de simulações, o que é usualmente inviável se considerarmos o tempo consumido para simular modelos de grande escala. Por outro lado, uma avaliação adequada da incerteza aumenta a qualidade e robustez das decisões tomadas para o gerenciamento dos campos de petróleo. Com esta motivação, foi investigado o método das Expansões por Caos Polinomial (PCE, por suas siglas em inglês). PCE é uma técnica de convergência rápida utilizada para analisar como se propaga, na saída de um modelo, a incerteza presente nos parâmetros de entrada. Mediante PCE, pode-se representar a resposta aleatória de um modelo de simulação de reservatórios de petróleo como um polinômio, construído a partir de uma base de funções que dependem da distribuição de probabilidade das variáveis incertas de entrada. Por outro lado, quando a relação entre os parâmetros de entrada e a saída do modelo têm um componente não polinomial, o algoritmo de Programação Genética (PG) pode ser utilizado para representar esta dependência utilizando funções ou operadores mais complexos. PG é um algoritmo de regressão simbólica capaz de encontrar uma expressão aleatória explícita, que aproxime a saída de um modelo de simulação de reservatórios de petróleo, conhecendo-se a priori a distribuição de probabilidade dos parâmetros de entrada. Neste trabalho foram aplicadas as duas técnicas, antes mencionadas, num modelo de simulação de reservatórios baseado no campo PUNQ-S3, considerando até vinte e três parâmetros incertos durante um período de produção de 13 anos. Foi feita uma análise de incerteza, calculando-se a distribuição de probabilidade completa da saída do simulador. Os resultados foram comparados com o método de Monte Carlo, indicando um alto desempenho em termos de custo computacional e acurácia. Ambas as técnicas conseguem níveis de ajuste superiores a 80 porcento com uma quantidade de simulações consideravelmente baixa. / [en] Reservoir simulation models are subject to uncertainty in a wide variety of its inputs. This uncertainty is a result of the heterogeneity of the geological formations, data measurement errors, and petrophysical, structural, and fluid transport in porous media modelling. An accurate uncertainty quantification requires, in most cases, a large number of simulations, which is unviable considering the time it takes to simulate large scale models. On the other hand, a proper uncertainty assessment, increases the robustness of the decision making process for the oil field management. To this end, the method of Polynomial Chaos Expansions (PCE) was studied. PCE is a fast paced convergence technique, used to analyze the uncertainty propagation of the input parameters all the way to the output of the model. Through PCE is possible to represent the response of an oil reservoir simulation model as a polynomial, built from a function basis, that depend on the probability distribution of the uncertain input variables. Furthermore, when the relationship between the input and output parameters of the model has a non-polynomial component, the algorithm of Genetic Programming (GP) can be used to represent this dependency by more elaborate functions or operators. GP is a symbolic regression algorithm, capable of finding an explicit expression that approximates the output of a reservoir simulation model, with prior knowledge of the probability distribution of the input parameters. In this work, the two previously mentioned techniques were applied in a reservoir simulation model, based on the oil field PUNQ-S3, considering up to twenty three uncertain parameters during a simulation period of 13 years. An uncertainty analysis of the output of the simulator was conducted, calculating the entire probability distribution. The results were compared to the Monte Carlo simulation method, presenting a satisfactory performance in terms of accuracy and computational cost. Both techniques show adjustment levels higher than 80 percent, with a considerable small amount simulations.