Polymer flooding is a widely used chemical enhanced oil recovery technique in sandstones, but it has limited applications in carbonates. This is due to the harsh conditions of carbonates including high temperature, high salinity, low permeability, and high heterogeneity. Polymer flooding improves volumetric sweep efficiency through improving the adverse mobility ratio by increasing water viscosity. Therefore, the effect of trapping number on oil recovery is usually neglected. This paper numerically investigates the effect of trapping number on oil recovery by polymer flooding in carbonates under harsh conditions. The trapping number effect was predicted by running several 1D simulations using measured reservoir rock and fluid data. A previously history matched biopolymer rheological behavior was considered. This study aims to validate the flow behavior of this biopolymer through history matching recently conducted coreflooding experiments. Sensitivity studies were performed on different uncertain parameters to history match the oil recovery observed in the lab. These parameters include inaccessible pore volume (IPV), polymer adsorption, permeability reduction, shear rate coefficient, hardness of makeup water, and trapping number. The results showed that IPV, polymer adsorption, permeability reduction, shear rate coefficient, and hardness of makeup water have negligible effect on oil recovery by biopolymer flooding. Also, history matching of oil recovery data was not possible when these parameters were varied within their typical range of values. Interestingly, when trapping number effect was applied through capillary desaturation curve (CDC), residual oil saturation was reduced to 10% as opposed to 20% without considering its effect. The pronounced effect of trapping number in this case is due to the increase in water viscosity post-polymer flooding compared to pre-polymer flooding, which is about two orders of magnitude. Therefore, the common practice of neglecting trapping number effect during polymer flooding might be misleading, particularly in carbonates that are characterized by a shallow CDC. Trapping number effect on oil recovery by polymer flooding depends on rock type and the targeted polymer viscosity, so the results might vary. The results indicate as well the effect of polymer flooding in improving both displacement and volumetric sweep efficiencies. The effect incorporated in this study is the elastic effect without considering polymer viscoelasticity, which might require further investigation. The study at the laboratory-scale is considered as a basis for field-scale predictions.

Fast and reliable reservoir simulation is a key for the successful decision making in integrated reservoir studies. Large and complex multiphase reservoir models usually require expensive computational infrastructure. Physics-based model order reduction (MOR) methods have been introduced and applied (POD-DEIM, POD-TPWL), especially for mitigating the computational cost of black oil models in workflows that require multiple calls of the reservoir simulator. However, only a limited number of methods have looked deeper at the effectiveness of these techniques to multiphase and compositional simulation where expensive phase equilibrium calculations are added to the level of complexities associated with obtaining robust solutions. In this work, we develop the physics-based MOR techniques for rapid compositional simulations that accelerate calibrating of system of equations and phase equilibrium. The combination of proper orthogonal decomposition (POD) and discrete empirical interpolation method (DEIM) has been used extensively in two-phase flow systems. POD reduces the size of the system to be solved, and DEIM contributes to approximate the nonlinear terms for faster computation. These snapshot-based methods work in a two-step process. In the training case one can obtain the snapshots, which are the solutions at each time step, to derive the POD basis and DEIM basis. Then the test case is utilized to validate if the reduced model works for the different well control schedule cases, and to compare the speed of the simulation run time. Results show that the POD-DEIM technique enables us to approximate the conventional model with high levels of accuracy up to more than 99%. And it also enables a faster simulation owing to the reduced order system. In this study, we show the robustness of POD-DEIM method to reduce the computational cost for multi-phase, multi-component 3D reservoir model.

Large data volumes and complex physical processes coupling multiphase flow, numerics and production assessment, as in the case of reservoir simulation, are difficult to analyze rapidly but are needed to guide operators for designing production strategies and subsequent reservoir management. The traditional reservoir simulation process is time consuming and alternatives such as data-driven proxy modeling can overcome the computation complexity drawbacks. A machine learning technique called recurrent neural network (RNN) has been proved useful for reservoir modeling with sequence data. In this paper, we develop a novel end-to-end production prediction workflow that can be used to fast guide reservoir development when production begins. In this work, we apply RNN on analyzing control parameter data and synthetic historical production data for better reservoir characterization and production prediction. More specifically, we would like to build a model to directly link the control parameters (flow rate and bottom hole pressure) with desired production output, e.g. production rate and water cut. One immediate benefit of the model is to avoid the state variable calculation (pressure/saturation). In addition, as this is a data-driven end-to-end production prediction model, it will not require the numerical iteration and gradient calculation once the training is completed. We explore two types of RNN based structure: cascaded LSTM and Ensemble Kalman filter (EnKF) enhanced LSTM. The LSTM (long short-term memory) is used to compensate the weakness of standard RNN for preserving long time information dependencies. The structure of LSTM takes into account the memory of previous calculation when modeling the current response. Our cascaded LSTM is an improvement to regular LSTM as it incorporates physical quantities of interest such as water breakthrough. The model is conceived with two consecutive networks: one network for breakthrough time estimation with output being fed into a second network that reconciles other features important for oil production prediction. The EnKF enhanced LSTM has the capability of performing data assimilation based on real time production data, thus providing a way to update our model constantly. In this work, we first show the methodology applied to the two-phase water flooding reservoir with five spot production scenarios. Then we conduct the comparison of Bayesian optimization tuned cascaded LSTM vs. standard LSTM. Finally, we showcase the usefulness of Ensemble Kalman filter in improving and updating current model. The method presented in this paper uses RNN (specifically cascaded LSTM) to learn the pattern from sequence data and identify the reservoir simulation proxy model, which can accurately predict surface production rate and water cut without the state variable calculation. The study also shows improved accuracy (over standard methods) for EnKF trained RNN and its capability of updating flow rate prediction based on new observation data.

We have developed a computational platform that allows the simulation of coupled flow-geomechanics-fracturing in a seamless manner. This allows consistent simulation of water flow back, proppant placement and production. The simulation process starts with the hydraulic fracturing process which is implemented by solving the fully coupled flow-geomechanics problem. Next, a critical stress intensity factor is computed which rules the fracture propagation process. For the results presented in this paper, we consider the simultaneous propagation of all the fractures belonging to a single stage. After the fractures have been successfully created, injection is stopped for a given number of days after which oil production starts. We showed a set of synthetic results that illustrate the stress and pressure changes that occur during fracturing and production of unconventional wells.

The "Oficina" formation is the most important in the Eastern Venezuela Basin and represents the main target for this study, which is to optimize the oil production and increase well productivity by the means of reducing the oil viscosity through the application of downhole electrical heating cable. The main objective of this study is to investigate and quantify the effect of downhole electrical heating cable in the performance of the extraheavy oil reservoir and horizontal wells interaction by lowering the oil viscosity applying a coupled reservoir-wellbore simulation approach. A fully integrated 3D reservoir and discretized mechanistic wellbore simulation approach is carried out in order to estimate downhole heating effect and interaction between reservoir inflow and wellbore performance during the heating inside the wellbore. Bottom hole pressure and temperature profile inside wellbore, productivity index, production fluid rates along horizontal well section are estimated under several sensitivities and optimization process including: heating rate, length and location of the cable inside wellbore. Numerical simulation is performed with commercially available thermal simulation software that has the feature of a wellbore modeling tool that can couple to the reservoir model. Additionally, optimization software is used to perform the sensitivity analysis studies and the automatic optimization of the heating rate, cable length and position. The heating cable simulation approach allowed understanding the importance of considering wellbore heating parameters and the interaction of the highly viscous extraheavy oil in the reservoir by increasing the temperature inside the wellbore. By applying wellbore calculations coupled to the reservoir simulation, it was possible to understand critical aspects of the oil and gas flow in the reservoir due to different drawdown along horizontal well section, pressure lossess along the wellbore due tor friction and the general efficiency of the heating cable and consequences in oil production. Bottom hole pressure, productivity index and production fluid rates along horizontal well section, were estimated under several sensitivities and optimization workflow including the heating cable parameters and location inside wellbore in order to estimate the additional cumulative oil production and increase of the productivity index and bottom hole pressure. Therefore, the dynamic model assessment permitted to define several criterias for well selection candidates and founded the basis for wellbore completion design and monitoring protocol program for pilot test in the Field. Coupling wellbore reservoir simulation is more accurate when compared to a conventional numerical simulation approach since it was possible to conclude that location of the heating in the horizontal section and gas breakthrough could affects negatively the oil production of the wells due to gas expansion inside wellbore with temperature increase. Therefore, it is very important that heating cable assessment consider the fully integrated 3D reservoir and wellbore modeling.

For unconventional reservoirs, the role of accurate reservoir characterization is becoming increasingly important for optimization of field development. The Fast Marching Method (FMM) has provided a novel approach for rapid flow simulation for unconventional fields, as it reduces the depletion in a 3D heterogeneous reservoir model to that of an equivalent 1D model. In this paper, a practical approach to handle highly heterogeneous matrix properties for dual porosity FMM simulation is established. Various averaging methods for the heterogeneous matrix properties are examined through integrated case studies. Current implementation of FMM simulation can directly account for fracture heterogeneity but requires averaging of the matrix property variations. In this study, we examine the error in the fluid production rate between the 3D heterogeneous reservoir model and the equivalent 1D averaged reservoir model. For some amount of matrix heterogeneity given by the standard deviation of the natural logarithm of permeability, 100 different realizations are created with a commercial geostatistical simulator. This paper evaluates the averaging approach in terms of the mean error and the distribution of errors for each value of the standard deviation. The error shows significant dependence on averaging methods especially for the reservoirs with high heterogeneity. Geometric based averaging shows the smallest error from the viewpoint of the mean error for high heterogeneity. Also, its distribution of the errors is narrowest, which indicates that the geometric based averaging method demonstrates the most stable performance. In contrast, the harmonic based averaging method represents the biggest error in terms of the mean value and has the least stable error from the viewpoint of its distribution. The work establishes an approach to consider the matrix heterogeneity in unconventional reservoirs with FMM. Through the integrated case studies, the most appropriate method for the different heterogeneities is evaluated. This study makes a significant contribution on the development of Fast Marching based reservoir characterization for more robust application for unconventional reservoirs.

In this work the potential for gas production from two selected methane hydrate deposits which are situated offshore from Uruguay is assessed along with the validity of numerical simulations as a tool for analysis in this environment. Gas hydrates are crystalline solids formed by gas and water, in which gas molecules are accommodated within a solid water lattice in a cage-like structure. They form in thermobaric conditions of relatively high pressure and low temperature which in nature occur in permafrost and deep water sediment environments. Marine methane hydrates represent a huge potential as an unconventional gas resource and production tests have already been perfomed offshore Japan and China confirming the validity of depressuration as a method of production. Available 3D seismic data was utilized for the identification of interesting areas for gas hydrate studies focused on resource exploitation allowing the acquisition of the corresponding architectural parameters. Due to the lack of well data at the selected locations, geological models and reservoir properties were defined based on published data from studies on analogue situations including data from the first production test performed offshore of Japan. Reservoir simulations were carried out to assess the response of selected prospects to depressurization induced dissociation. Two prospects, interpreted as turbidite type deposits and located at 1850 m and 788 m of water depth, were selected for the modelling studies. The simulation of short term production tests of 60 days indicates average gas release rate values from 34 100 std m 3 /d to 6700 std m 3 /d for the deeper and shallower prospect respectively. The simulations were greatly affected by geometrical non-geological parameters like the proximity of model boundaries as well as type and level of discretization. We found that for finer discretization cases, the use of logarithmically distributed radial grid cells led to the existence of artifacts at early time on the gas release rate curves while the use of uniformly distributed radial cells results in more stable behaviour of the gas release rate. Several realizations of the geological models were used and sensitivity analysis was carried out regarding permeability and hydrate saturation. A longer term production regime (10 years) for a heterogeneous layered case was also simulated for the deeper prospect resulting in very useful average gas release rates of approximately 70 000 std m 3 /d, essential to satisfy gas requirements of Uruguay. We predict that only a few wells would be needed. For the first time, reservoir simulation was applied for prospects in Uruguay and the gas release potential for marine methane hydrate deposits in the Southern Atlantic margin was assessed. Simulation results are encouraging. Additionally the results of this work at the identified prospects may be useful for site selection for any future campaign for gas hydrate exploration offshore Uruguay.

La Cira Infantas (LCI) is the oldest oilfield in Colombia, with 100 years of oil production history spanning three periods: primary depletion, the first waterflood (WF) period, and the second WF period. After the first WF period, the field had neared its economic limit and an abandonment program was underway from 2003 to 2005. Partners Ecopetrol and Occidental made the decision to implement a waterflood redevelopment project in 2005 that included reconfiguration of the old WF areas and WF expansion into new areas. The partners formed integrated reservoir development and management teams (RDT and RMT) that have combined workovers, infill drilling, WF optimization, and other IOR/EOR methods. The oil production has increased from approximately 5,000 to over 43,000 bopd, which is above the predicted 40,000 bopd. One of the key components was WF conformance controls to improve sweep efficiency and to improve injection profiles across multiple stacked-sand layers. The selective injection method was implemented using multiple packers and side-pocket mandrels/valves. The design of selective injection strings involved complex engineering calculations and left some room for improvement. This paper describes a new workflow used to optimize the selective injection design using multiphase flow simulation based on actual WF patterns. A field pilot test used an inverted 5-spot WF pattern to demonstrate this new workflow. The simulation results predicted a 45% incremental oil production above the base case. The simulation optimization process reduced the number of mandrels/valves from nine to six, which saves about 30% of the associated completion costs compared with the initial completion proposal for this case study.

This paper presents a technical evaluation of the results obtained in a field pilot to optimize the cyclic steam stimulation (CSS) by evaluating the performance of the indicator of oil steam ratio (OSR). The pilot was based on numerical simulation, where several alternatives or strategies were identified to improve the current technique applied for wells with more of 15 cycles as a method of improved or enhanced recovery in a madure heavy oil field. In a heat transfer process, in this case used by cycle steam injection, once it has maximum use, the next step in most of the termal process is to apply or move to steam flooding tecnic, but the implementation of it requires high investments that currently are not viable. However, it is necessary to provide solutions that contribute to improve the CSS efficiency of the field. The following workflow was developed: (1) research of literarature and field experiences of different alternatives to increase the actual CSS efficiency by only changing the steam injection strategy, without including additional costs related to additives or technologies. (2) selection of the pilot area based on the current petrophysical, production, steam injection conditions and supported by a numerical reservoir simulation model with different sensitivities to define the base case to be developed in the field and (3), the optimal case was implemented in the field during one year. Results are discussed in this paper. An area of the field under study was selected, considering the following conditions: low pressure, high water cut, number of CSS per well greater than 15, inefficient CSS response. (4) alternatives for the optimization of the CSS were identified; these were evaluated by numerical simulation and it was found the optimal development strategy the intermittent injection. With these results, a field pilot was implemented which consisted of 3 injection wells in 3 different patterns, for a period of 12 months of evaluation. During this period, it was possible to reduce the strong decline in oil production presented in the area from 23% to increase to −25% A.e., saving 20,000 mmbtu/ of steam in a year, generating an increase in the OSR of 0.12 BBl / MMBTU; and a decreasing the number of well services jobs from 51 to 27. This result was achieved because areas with oil saturation that had not been previously stimulated in the conventional process of CSS were contacted under this new scheme. The current conditions of the field, low pressure and high saturation of unconnected remaining oil during the conventional cyclic steam injection until now applied in the field, led to propose new cyclic stimulation alternatives such as the one presented in this paper; highlighting the technical results of the pilot developed without additional cost, just changing the injection strategy to optimize the current state of the field.

An integral aspect of field development planning is the process of identifying remaining, feasible, and actionable field development opportunities (FDOs) such as recompletions and new drills. In mature brown fields, this process is extremely difficult to undertake in a reasonable timeframe due to the sheer size of the available datasets (e.g., thousands of wells, decades of production history), geologic complexity (e.g., dozens of layers, faulting, folding, fractures, pinchouts), and engineering challenges (e.g., commingled production and injection, field compartmentalization) unique to each reservoir. The automated framework is a fully streamlined framework that aims at building a comprehensive opportunity inventory consisting of behind-pipe recompletion opportunities, vertical new drill locations, sidetrack opportunities, and optimal deviated/horizontal targets. In addition to opportunity identification, it can generate additional deliverables for vetting purposes including a well book report and a standardized Petrel project. The main objective of this process includes: 1) fast, data-driven, and consistent approach for evaluating and identifying field development opportunities, 2) fast execution time, accelerating processes from months to days, 3) integration of multi-disciplinary data, 4) consistency and repeatability, minimizing subjectivity, and 5) streamlined visuals for opportunity vetting. This framework has been successfully applied to a mature field in the Middle East with more than 1,000 producers and 55 years of production history. Hundreds of recompletion opportunities were initially identified for the field. A final list of opportunities is then generated by running the opportunities through a series of user-provided attribute filters such as minimum initial rate, minimum oil thickness, maximum acceptable uncertainty, and so on. The process is then typically finalized with manual vetting of the opportunities by subject matter experts including geologists, reservoir engineers, and production engineers that ensure the opportunities are geologically sound, mechanically feasible, and meet various validation criteria that are discussed in this document. A final list identified 116 recompletion opportunities in 116 wells with a total potential production upside of around 63,000 B/D.

History matching is a critical part of the simulation workflow. History matching in shale is unique because it involves data from both hydraulic fracturing and from production. This paper presents a step by step workflow to facilitate history matching numerical simulation models of hydraulically fractured shale wells. We perform sensitivity analysis simulations using a coupled hydraulic fracturing, geomechanics, and reservoir simulator. The results are used to develop ‘motifs’ that inform the history matching process. Using intuition from these simulations, we show how to expedite history matching by changing matrix permeability, fracture conductivity, matrix pressure dependent permeability, boundary effects, and relative permeability. Finally, a worked example, using two Eagle Ford hydraulically fractured wells, is presented to demonstrate the workflow.

Reservoir simulation is a popular tool to understand unconventional reservoirs dynamics. Applications include estimating long-term production behavior, enhancing well spacing and pad modeling efficiency, optimizing completion and stimulation of horizontal wells, and understanding production drivers that cause differences in productivity between wells. The objective of this work is to revisit fundamental concepts of reservoir simulation in unconventional reservoirs and to give several real examples that form part of an archive of lessons learnt. Our work includes several reservoir simulation models in unconventional plays worldwide. These models are a function of the specific objective (from the 4 aforementioned applications) and reservoir type. They include structured and unstructured grid models, high- and low-resolution gridding, single porosity and dual porosity, compositional and black oil PVT, variations in the definition of complex hydraulic fractures in shale reservoirs (accounting for variations in properties that occur in time and space), and different protocols for incorporating initial water in-place as a result of hydraulic fracturing fluid. We have encountered several challenges during this work. Some of these include: simulation model non-uniqueness, accounting for variations of reservoir properties with time and space and implications of production forecasting in volatile oils or gas condensates. Our work has brought to light important aspects of modeling unconventional reservoirs, such as changes in apparent well productivity after well shut-in or choke changes, cluster spacing, grid size and complex fracture thickness effects in horizontal wells, supercharging effects resulting from hydraulic fracture treatments, and simulation grid cell size impact on reservoir simulation cases where modeling transient flow behavior between perforation clusters. In most cases, to reduce computational time we have taken advantage of modeling a portion of the lateral and scaling up the results. The lessons learnt continue the forum for further discussions regarding shale reservoir well and production modeling in our industry. They will provide a useful reference especially for those with little experience in unconventional reservoir simulation to better understand and develop both new and existing unconventional reservoirs.

Simulation of drilling processes involves the seamless integration of strongly coupled physics, numeric, and economic models. Operators desire advanced drilling simulation software that can mimic reality and yet be practical for daily operations. At the same time, academic researchers are devoting time and effort to develop robust and accurate models of every module of an integrated drilling environment but are challenged with translating their original contributions into the operational setting. We seek to solve both of these issues through a distinctive collaboration consortium based on an open-source paradigm. We describe and propose an advanced drilling simulation developed through a joint industry consortium at Texas A&M University. The simulator provides free access to sophisticated and reliable drilling dynamics solvers. This includes drill string dynamics, torque and drag analysis, geologic modeling, hydraulics, and wellbore pressure modeling, visualization, and operator interfaces. The simulator is written in C++ and is built on top of high-performance open source libraries that leverages parallel computing, GPU acceleration, multibody, and computational fluid dynamics solvers as well as other tools (see figure for further details). We will assist industry and academic users to employ our open-source tool. In addition, we will help said parties further develop our framework into proprietary and protected business software via the addition of custom functionality. The simulator core was successfully developed and has been used to model horizontal drilling, while providing near real-time performance. The framework has been used in several applications successfully, including modeling shallow horizontal jetting, rock drilling, and steering; as well as reconstructing spatial data from high fidelity downhole signals for borehole evaluation, and the development of improved downhole navigation algorithms. We believe that a collaborative development effort between academia and industry will bring solutions to complex modeling needs faster and more economically than any individual endeavor. To this end, the authors will also show how one can partner with Texas A&M University to obtain access to the open-source enabled drilling simulation framework and platform to collaborate, improve and create solutions for new and existing problems.

In the case of this pilot test using polymer flood, evaluating the performance of injector wells was one of the main priorities. Losing injectivity is recognized as one possible cause of poor performance in the case of enhanced oil recovery processes, so estimation of the length of the well that contributes with injection, the identification of by passed zones, the diagnostics of the well behavior, and recommending solutions for a future project, were part of the objectives stablished at design of the project. For those reasons, we chose to complete the wells using a distributed temperature sensor (DTS). After installation of fiber optic sensors system in the wells associated with the project, the acquired data was analyzed using different techniques: Comparison of the effective length obtained from the static model with the one obtained from fiber optic sensors, and also pressure test analysis, later the distribution of the injection rate in each case was obtained. Initially it was necessary to establish a base behavior: Real-time measurements of pressure and temperature, characteristics of the injected fluid, and petrophysical evaluation of the completed intervals, which was used as departure point for studies performed. Maintaining a good injectivity in polymer injection projects normally is a major concern, in this case we present a real case of a polymer injection project in an extra heavy oil reservoir, which is producing oil with an average viscosity of 4150 cP at reservoir conditions. Polymer solution with a concentration of 1600 ppm was injected since 2017 and the injection is maintained so far without important reduction in the injectivity index. An estimated of the injection interval was established, and also estimation of the injection distribution in the horizontal section. The estimation was then validated with the results of the measured surface rate, and pressure transient test analysis. In average 30 % of the horizontal section of the well were receiving polymer solution, and then thanks to this new information zones not receiving polymer solution were identified. The results were used to improve the reservoir simulation models, and as diagnostic tool for pilot test. This paper describes how a better understanding of the connection between the well and the reservoir, can improve the reservoir simulation models, decision making process, and well performance evaluation. Also the importance of determine the possible not swept areas in the reservoir, which could result in low volumetric swept efficiency, and as a consequence low recovery factor. Finally the results presented here could be useful to the general interest of engineers dedicated to polymer flooding projects.

Camisea asset is possibly the most important mega gas asset in South America, and constitutes for more than 90% of Peru’s gas production. The asset consists of several fields with varying degrees of subsurface complexities, multiple fluid types ranging from very lean to very rich gas condensate. The combined production is routed to gas processing plant via a network of pipelines. The entire development is in an environmentally sensitive area and adds to the complexity of the surface network design process. For these reasons, performing integrated production systems simulations is essential for designing the surface network with appropriate fidelities for each component of the IPSM. In current work, a complex integrated production system was developed for Block A of the asset. It is a 2-field system with complex fluid behavior. The various fluids blend in the well tubings, as well in the combined surface network. The produced gas is treated in the plant for liquid extraction, and dry gas is routed for sales and re-injection. For all this work, a new multi-fidelity IPSM tool was used in both implicit as well as explicit coupling mode. This was achieved by doing both implicit as well as explicit coupling of the reservoir simulation models with well models and surface network model. As a result of this work, a reliable long-term production forecast of the field could be performed. Additionally, the pipe design and network constraints could be evaluated and compression needs variation with time could be assessed. Lastly, with the use of a fit-for-purpose fidelity level for any IPSM component, more efficient and reliable forecasts could be readily generated. A comparison between implicit and explicit coupling of IPSM components is generally missing from the available literature. This study fills in this gap by providing the option to do the two couplings using a single simulation tool, thus providing a consistent comparison. In addition to the various long-term production forecasts, the benefits and drawbacks of the two coupling approaches has been presented in this paper.

Applying simpler and more powerful waterflood performance analytical modeling tools to history match and forecast fluid (oil and water) production rates is always a subject of interest. Increasing improvements of these no grid-based tools trigger their use as predictive and trustworthy precursors of grid-based modeling, providing significant insights and allowing a previous assessment of historical and future waterflood performance without a significant time- and investment-consuming modeling. Setting up a unified fractional-flow model (UFFM), that considers both the traditional Buckley-Leverett-based stable fractional-flow model (BLBFFM) and instabilities due to oil viscous-fingering effect, to accurately predict the oil recovery from traditional waterflood performance analytical modeling tools is the objective of this paper. The unification is based on using appropriate k ro/ k rw vs. water saturation expression that considers viscous-oil effect based on the effective finger model (EFM), allowing the substitution of a classic semilog linear relationship of k ro/ k rw vs. S w (constant coefficients A and B ) by a unified nonlinear oil-water relative permeability ratio. This UFFM approach aims to boost the analysis from well-known waterflood performance analytical methods: water-oil ratio (WOR), X-plot, Y-function and the capacitance-resistance model (CRM). Such waterflood performance diagnosis methods, traditionally based on the BLBFFM, were redefined by using the new UFFM. Conceptual and field cases were tested using new definitions to evaluate impact on ultimate oil recovery and various reservoir parameters. Results demonstrated that unified approach improves reliability when estimating oil recovery for stable and unstable waterfloods. As expected, they quantitatively yield, for example, that the greater the viscous-fingering effect, the lower the viscous-oil recovery. The universalization of fractional-flow functions ( f w and f o ) by incorporating UFFM upon all traditional waterflood performance methods based on BLBFFM can allow insights of important physical phenomenon that affect relative permeability, such as the oil-water viscosity ratio, injection rate, and known physical parameters affecting ultimate oil recovery.

This work aims to optimize the development plan involving Low Salinity Water Injection – LSWI for the Namorado Field (located in the Campos basin in Brazil). Multiple mechanisms have been proposed in the literature to represent LSWI, mechanisms that we are going to discuss in this paper. Additionally, it will be studied the impact to evaluate multiple geostatistical realizations at the same time with multiple development scenarios (Robust Optimization) to select the optimal Net Present Value (NPV) of the project considering the reservoir uncertainties. This methodology combines compositional simulation and a mathematical optimization tool that uses artificial intelligence to maximize the NPV of the project under evaluation. Considering that LSWI alters the initial chemical equilibrium in the reservoir and induces changes in the system, a geochemical model is included with aqueous reactions, dissolution / precipitation mineral reaction and ion exchange. The optimization project was carried out in three stages, first a multiple assisted history match, followed by a probabilistic forecast and finally, a robust optimization. As a result, an optimum operational strategy of LSWI was obtained in which the impact of changing the type, location and number of producer and injector wells as well as the volume of injected water (High Salinity Water Injection (HSWI) vs. LSWI) are considered. On average, the optimal scenario for LSWI obtained a recovery factor of 41%, compared to 34% if only conventional water is injected. The selection and optimization of the recovery strategy considered multiple geological realizations and multiple history matched models in the field development plan, this considers the uncertainty associated with the porosity, permeability and net to gross distribution, with the objective of increasing the accuracy and reliability of the results obtained for the recovery factor and the NPV. For the optimal LSWI case and considering five different selected geological scenarios (P10 to P90), the recovery factor obtained was between 37.89% (P10) and 42.21% (P90) and the project NPV between 807 MMUSD (P10) and 3315 MMUSD (P90). This adapted methodology allowed to obtain the optimal development plan considering the uncertainty associated in the reservoir, using multiple geostatistical realizations and multiple history matched models at the same time (usually just one history match is considered).

The process of identifying and performance forecasting of the remaining, feasible, and actionable field development opportunities (FDOs) is the core component of optimum field development planning and management. In the present work, we introduce a multiscale data-driven forecasting framework that applies a series of novel technologies to provide short- and long-term production forecast and optimization for both field and well level performance. Our workflow can be applied to a comprehensive FDO inventory including behind-pipe recompletion, infill drilling, and sidetrack opportunities. Using smart spatio-temporal clustering, we automatically divide the reservoir into a specific number of compartments with distinct static and dynamic properties, in which the direction-dependent multiphase flow communication is a function of nonlocal phase potential differences. The reservoir connectivity structure is encoded in an adjacency matrix describing the neighbor and non-neighbor connections of comprising compartments. We then apply a recently proposed robust Ensemble Smoother Levenberg-Marquardt (rES-LM) method to generate plausible model realizations which replicate the reservoir energy by adjusting first-order model parameters such as pore volumes, fault transmissibilities, aquifer strength, and matrix-fracture split. These calibrated upscaled network models serve as pre-conditioner for a detailed model calibration step. We carry out a second round of full-scale reservoir simulation model calibration, anchoring updates on large-scale model parameters estimated from the network model. Representative models are further improved in a sensitivity-based local inversion step to match multiphase production data at the well-level. This structured, multiscale approach offers improved stability in reservoir model calibration. Finally, calibrated models are directly passed to the forecasting and optimization engine to assess and optimize field opportunities and development scenarios. The proposed workflow is applied to a major fractured offshore field in South America. Leveraging the fast forward model, an efficient ensemble-based history matching framework was applied to reduce the uncertainty of the global reservoir parameters, such as inter-blocks and aquifer-reservoir communications, and fault transmissibilities. The ensemble of history-matched models was then used to provide a probabilistic forecast and optimization for different field development scenarios. A novel hybrid approach is presented in which we couple a physics-based nonlocal modeling framework with data-driven clustering techniques to provide a fast and accurate multiscale modeling of compartmentalized reservoirs. Our approach facilitates a flexible framework to rapidly generate reliable forecasts and quantify associated uncertainties in a robust manner. This advantage in flexibility and robustness is tied to our fast and automated two-stage model calibration workflow that leads to substantial saving in computational time. This research also adds to the literature by presenting a comprehensive work on spatio-temporal clustering for reservoir studies’ applications that consider the clustering complexities, the intrinsic sparse and noisy nature of the data, and the interpretability of the outcome.

This case study describes the successful outcomes and learnings from a brownfield tie-back project in offshore Trinidad. The greater Angostura Asset consists of multiple Oligocene-age turbidite sandstone reservoirs with complex geology and compartmentalization. The Asset has produced oil since 2005 from thin oil rims overlaid by large gas caps. This has been supported by significant gas reinjection and constrained by gas coning. Commencement of gas sales added value but also increased gas cap blowdown in the oilfields. The Angostura Phase 3 project was conceptualized to enable gas production from an adjacent gas discovery and extend the overall productive life of the existing oil fields. Key subsurface uncertainties and operational constraints were identified early in the project. Historically, drilling in the Angostura field has been challenging due to the complex geology and compartmentalization. Poor seismic data in the Phase 3 area added complexity to optimal well placement. Ensuring high deliverability was another critical project requirement. The main reservoir management challenge in the existing fields was optimizing both oil and gas sales while managing voidage and coning. Limited topside facilities and space constrained addition of new gas at flowing at higher pressures. To address these challenges, a multi-disciplinary project team was formed to plan and execute the project. Material balance, detailed geological modeling and numerical simulation were used to understand the impact of key uncertainties on recovery. The integrated evaluation resulted in an optimized development plan with three subsea gas wells tied back to existing facilities. Production performance of existing wells helped highlight critical performance drivers for Phase 3 wells. Modified completion designs along with a flow back and stimulation program during completion helped to maximize productivity. Phase 3 wells have performed at the high levels expected during project design and sanction. Detailed surveillance (including pressure transient analysis (PTA) and interference testing during completion, start-up and production phases) has continuously helped to optimize production. Oil production in the existing fields has also been improved by better balancing gas sales and injection. The project added significant value from the increased gas sales and oil production. This paper discusses the planning, successful execution and impact on the overall value of the Angostura Asset from the Phase 3 project. Optimized completions along with an improved approach for evaluating stimulation options during completion (detailed near well-bore modelling, real-time PTA and simulation) enabled strong well performance. We highlight some of the critical factors for achieving success in brownfield tie-back projects including having (right from project inception): multi-disciplinary project teams, tight coordination with existing operations and a focus on key uncertainties and constraints. Building optionality and flexibility into development plans and being prepared for contingencies have been the other critical learnings from this project

The objective of this paper is to develop predictive models to optimize the (1) characterization of the stimulated reservoir volume (SRV), (2) discretization of the fracture network, and (3) hydraulic fracturing modeling, by combining machine learning (ML) algorithms and reservoir engineering in low permeability reservoirs. An unsupervised learning algorithm is implemented to characterize the fracture network developed by micro-seismic observations during hydraulic fracturing. A Self Organizing Map (SOM) and Multi-Attribute Analysis are performed on the available seismic data to map the extension of the hydraulic fracturing stages and the fracture network complexity in a low permeability reservoir. To correlate the mapped fracture network and discretized SRV, a 3D Finite Element Model (FEM) is developed to estimate fracture behavior, stress response, and hydraulic fracture propagation, on the predicted and forecasted multi-attribute map of the reservoir. A 3D hydraulic fracture propagation model (HFPM) is introduced, to delimit the fracture geometry and remove data outliers in the SOM algorithm. Unsupervised algorithms rely on data quality. The efficiency of hydraulic fracturing modeling is improved with a machine learning approach by refining the certainty and quality of the data. An Artificial Neural Network (ANN) model helps to select the most significant parameters related to fracture modeling and simulation in the field. This approach allows us to recreate and forecast complex fracture networks in low permeability reservoirs, based on the learned geostatistical maps and hydraulic fracturing parameters, particularly where the microseismicity is limited or unavailable. To validate the implementation of the 3D-HFPM in the field, an earthquake model is compared with statistically significant microseismic events obtained by the unsupervised iso-cluster algorithm. The relationship showed a good agreement, which suggests the HFPM agrees with seismic observations in the field. The machine learning application to fracture network modeling provides the capability to identify susceptible areas to well interference and possible frac hits with higher certainty. This is so because the approach improves the selection of seismic data and hydraulic fracturing parameters, employed to develop the complex fracture network in numerical commercial reservoir simulators. This helps to determinate the reservoir interconnectivity and flow patterns in the fracture network. This approach presents a robust manner for characterizing the SRV using a relative fast methodology, based on the combination of geostatistical and unsupervised learning modeling. The seismicity and hydraulic fracturing are connected using a multi-attribute and multi-disciplinary interpretation. It is a powerful tool for characterizing problematic fracture networks in unconventional reservoirs.