Sun, Zhuang (The University of Texas at Austin) | Tang, Hewei (Texas A&M University) | Espinoza, D. Nicolas (The University of Texas at Austin) | Balhoff, Matthew T. (The University of Texas at Austin) | Killough, John E. (Texas A&M University)
The reduction of pore pressure caused by depletion can induce significant reservoir compaction, especially in unconsolidated reservoirs. Experiments using unconsolidated core samples are often sparse and costly. We develop a numerical approach based on computer-based simulations of rock samples and mechanical tests. The numerical sample consists of crushable grains simulated with the discrete element method (DEM) and the bonded-particle model (BPM). Model parameters are calibrated through numerical single-grain-crushing tests which reproduce the experimentally-measured sand strength. Grain crushing induced by the uniaxial strain stress path results in a pronounced reduction of porosity and permeability, which manifests more readily for samples with large grain size. The change of particle size distribution indicates that the high effective stress causes grain crushing and produces a significant amount of fines. We perform numerical uniaxial strain tests on numerical samples comprising stiff and soft mineral grains. Simulation results indicate that the presence of soft grains and inclusions (e.g. shale fragments) facilitates the grain crushing. Reservoir simulations, incorporating the change of porosity and permeability as a compaction table, show that the upscaled compaction can enhance production due to compaction drive but also reduces production rate by impairing the reservoir permeability. This multiscale numerical workflow bridges particle-scale compaction behavior and field-scale reservoir production. In this paper, (a) DEM simulations provide a useful tool to investigate compaction effects and complement laboratory experiments; (b) the multi-scale numerical approach can predict the depletion-induced evolution of reservoir production.
Models for steam or hot-water injection into a fractured diatomite or shale reservoir are developed from existing analytic models of energy transport and countercurrent imbibition.
Radial convective heat flow through a horizontal fracture system is modeled with conductive heat flow into the low-permeability matrix. The flow geometry approximates hot-fluid injection into a five-spot pattern. Recovery mechanisms accounted for in the models include capillary imbibition and thermal expansion. Temperature dependence of viscosity and interfacial tension (IFT) are included in the imbibition estimate. Laboratory data are needed to quantify the magnitude of the imbibition mechanism, which is usually the primary contributor to oil recovery. Reservoir properties representative of either the Belridge Diatomite or the Antelope Shale, two giant fractured oil reservoirs, are used for the model forecasts. Currently, however, only temperature-dependent imbibition data for diatomite reservoirs are available.
The steamflood model has been partially validated against a large-scale project in the Belridge Diatomite. By use of public-domain information, a reasonable comparison was obtained between the model and the field project during a 4-year injection period. Comparison with conventional thermal simulation was also performed, and it indicated reasonable agreement with the steamflood analytical model.
The models have been used to determine the key factors determining the success of thermal recovery in fractured, low-permeability reservoirs. Steam injection is shown to be superior to hot-water injection in heating the matrix. Key factors enhancing recovery include reduced fracture spacing, increased matrix permeability, and increased injection temperature. Model results indicate that steamflood recoveries of more than 40% of the original oil in place (OOIP) may be achieved by injection in diatomite containing light oil. Application to diatomites containing heavy oil also shows good performance. Successful application in diatomite reservoirs is forecast to be possible in the current low oil-price environment. Economic application in fractured shales, assuming similar imbibition behavior as in diatomites, would require a higher oil price because of the higher well costs and lower oil content relative to diatomite projects.
Because of the significant volumes of remaining oil in place (OIP) in both the diatomite and shale reservoirs, the application of thermal enhanced oil recovery (EOR) to these resources represents the logical next step in steamflood development.
Continuous variable pressure steam injection is a new approach to manage heat flow within a reservoir and reduce the temperature of fluids arriving at producers thereby avoiding shut-in caused by high temperature. An optimal injection scenario with varied pressure and constant steam quality is shown to have the largest oil recovery given the same amount of energy injected under various reservoir settings. The economic effects were evaluated to select the optimal variable pressure injection schedule with the greatest Net Present Value (NPV).
Various scenarios with different injection pressures and steam qualities were investigated and compared based upon aspects of oil recovery, energy consumption, and periods of producer shut in. Sensitivity analysis was performed regarding permeability, well distances, heat losses, as well as geomechanical model parameters including Young's Modulus, Poisson's Ratio, and initial stress. Grid refinement experiments dictated the optimal unit grid dimension for accurate 3D modeling including geological heterogeneity. Reservoir subsidence was also investigated. The variable-pressure injection cycles were designed and evaluated using an economics model.
Results teach that the proposed steam injection schedules provide continuous oil production (i.e., no producer shuts in) with the largest energy efficiency and economic benefits. The hybrid injection strategy for a 3-D heterogenous case showed an increase in 15% more oil recovered compared to the constant pressure injection scenario after 10 years of production. We conclude that steam injection design effectively controls the reservoir temperature and avoids producer shut-in. The sensitivity analysis further supports the feasibility of this optimal scenario under various reservoir conditions. Inclusion of geomechanics provides important information about reservoir deformation, but geomechanics has little effect on the oil produced or the optimal injection strategy. Importantly, the subsidence results do not indicate well failure under the conditions studied. By performing a systematic investigation, the combination of continuous high and low constant-pressure steam injection for fixed time intervals yields the largest NPV.
Thermohydromechanical effects can have significant impacts on the operations of heavy oil and geothermal fields. This motivates the use of accurate reservoir simulators in field development planning. Previously, simulation using continuous Galerkin methods was examined. However, these methods can experience difficulty in convection-dominated flow problems. Consequently, the discontinuous Galerkin method was considered as an alternative. The discontinuous Galerkin method's literature suggested the need to state a generalized, single-phase thermohydromechanical formulation to clarify previous work. This paper states this generalized formulation and performs a poromechanical benchmark of uniaxial compaction in FEniCS using the discontinuous Galerkin method. Additional work on extension to multiphase flow and the incorporation of thermal effects is needed.
Key components of well integrity include the well casing's condition and the low-permeability, cemented annulus between the casing and the borehole wall. Properly cemented casing provides zonal isolation between the reservoir, overlying strata, drinking-water aquifers, and the surface. Although proper casing and cementing is a challenge in itself, subsequent thermal, mechanical, and chemical alteration can damage this zonal-isolation system and potentially result in leakage toward the surface (Bellabarba et al. 2008; Wotjanowicz 2008; Bois et al. 2011; Carey 2013). Fluid injection is expected to cause an increase in reservoir pore pressure, which can induce reservoir uplift (Vasco et al. 2010) as the reservoir volume increases and effective stress in the reservoir decreases.
Along with the introduction of the Bang‐Bang theory, it is known that on‐off control of injection and production in waterflooding of hydrocarbon reservoirs improves recovery by means of simply switching the wells or valves on and off where optimal control theory is applied that employs an optimization procedure, which is gradient‐based with adjoint formulation. It has also been shown under what conditions reservoirs might have Bang‐Bang improvements. In this study, we extend the concept to include the geomechanics factor in softer formations where subsidence/uplift are observed along with injection or withdrawal of fluids. In this study, a robust commercial optimization and uncertainty/sensitivity software is coupled with a full‐physics commercial simulator that models the phenomenon applying it on major types of reservoirs with different geomechanical characteristics in order to illustrate the significance and applicability of Bang‐Bang control. A comprehensive theoretical background is provided to illustrate the approach to clarify the differences not only for the reservoir characteristics but also the simulation results. The optimization and uncertainty results prove useful in illustrating the significance of each decision and uncertainty parameter enabling the decision makers to make robust reservoir management decisions under economically‐tight environments providing not only the optimum values of amount of injection but also the right combination of injection‐production strategy to increase net present value or recoveries. The additional knowledge of temporal and spatial geomechanical behavior and response of reservoirs proves useful in application of Bang‐Bang control with more realistic expectations, thus leading to better reservoir management decisions. Although Bang‐Bang control and application of optimal control theory are not new in the industry, there has not been a comprehensive study taking into account the geomechanics. Thus, this study outlines the behavior and performance of reservoirs that experience subsidence and uplift at different levels in the application of this methodology in waterfloods.
Popa, Andrei S. (Chevron Corporations) | O'Toole, Conor (University of Southern California) | Munoz, Juan (Chevron Corporation) | Cassidy, Steve (Chevron Corporation) | Tubbs, Dallas (Chevron Corporation) | Ershaghi, Iraj (University of Southern California)
The successful waterflooding field development depends not only of the understanding of reservoir characterization, sub-surface injection displacement and sweep-efficiency, but also an accurate and effective design and operation of the surface network’s water injection distribution. In certain cases, the latter is critical for the successful waterflood operations in large fields with thousands of wells and high volume of new development activity.
For the purpose of this study, we are presenting a new data-driven approach to accurately estimate the injection rate in all non-instrumented wells in a large waterflooding operation. A collection of data driven tools, including statistics, clustering, simulation and an artificial neural network model were employed to prime and model the data. As a final point, the neural network leverages instrumented wells’ data and serves as an accurate real-time proxy to estimate missing injection rate measurements in non-instrumented wells. The system’s accuracy was validated by comparing the estimated rates for different wells on a different branch with the ones measured at physical wells. The neural network model trained on the cleansed data set revealed a high performance system with a >0.93 R2 values for both training and validation sets.
The paper outlines both the methodology and procedures used to analyze a branch of the water network system, and the modeling of accurate estimation of injection rates. The model performance is remarkable having used only field and wellhead measured data and considering the natural uncertainty inherited in these values. Finally, this system provides the capability to estimate the flow rate for every non-instrumented well in the field and respond to exceptions in relevant time.
Suhag, Anuj (University of Southern California) | Ranjith, Rahul (University of Southern California) | Balaji, Karthik (University of Southern California) | Peksaglam, Zumra (University of Southern California) | Malik, Vidhi (University of Southern California) | Zhang, Ming (Rafflesia Energy) | Biopharm, Frontida (Rafflesia Energy) | Putra, Dike (Rafflesia Energy) | Wijaya, Zein (HESS) | Dhannoon, Diyar (Texas A&M University) | Temizel, Cenk (Aera Energy LLC)
Conformance improvement is the key to success in most enhanced oil recovery (EOR) processes including CO2 flooding and steamflooding. In spite of technical and economic limitations, foam has been used as dispersions of microgas bubbles in the reservoir to enhance mobility. Steam-foam has numerous applications in the industry, including heavy oil reservoirs, which are a significant part of the future energy supply. Steam-foam applications have been used to prevent steam channeling and steam override, thus improving overall sweep efficiency, in both continuous steam and cyclic steam injection processes. The objective of this study is to investigate the key components of this complex process, where relatively high temperatures are recorded, in order to have a robust understanding of chemistry and the thermal stability of surfactants.
The efficiency and therefore economics of the steam-foam process are strongly reliant on surfactant adsorption and retention. This requires a good understanding of the process for effective sizing of the foam injected. In this study, a commercial reservoir simulator is used where surfactant transport is modeled with surfactant availability and is determined by a combination of surfactant adsorption, surfactant thermal decomposition, and oil partitioning due to temperature. The degree of mobility decrease is interpolated as a result of factors that contain aqueous surfactant kind and concentration, the presence of an oil phase, and the capillary number. An empirical foam modeling method is employed with foam mobility decrease treated by means of modified gas relative permeability curves.
The simulation results outline the sensitivity of these parameters and controlling agents, providing a better understanding of the influence of surfactant adsorption and thus, a number of chemicals to be used in an efficient manner. Optimum values for decision parameters that we have control on have been determined by coupling a commercial optimization software with the reservoir simulator. Uncertainty parameters such as surfactant adsorption have been analyzed in terms of significance on the recovery process.
Even though steamflooding is thoroughly studied in the literature, there is no recent in-depth study that not only investigates the decision parameters but also uncertainty variables via a robust coupling of a reservoir simulator and an optimization/uncertainty software that model use of foam in steamflooding. This study aims to fill this gap by outlining the optimization workflow, the comparison of parameters with tornado charts and providing useful information for the industry.
Balaji, Karthik (University of Southern California) | Suhag, Anuj (University of Southern California) | Ranjith, Rahul (University of Southern California) | Yegin, Cengiz (Texas A&M University) | Saracoglu, Onder (Consultant) | Hendroyono, Arief (OXY) | Dhannoon, Diyar (Texas A&M University) | Temizel, Cenk (Aera Energy LLC)
Along with the introduction of the Bang-Bang theory, it is known that the on-off controls of injection and production in waterflooding of hydrocarbon reservoirs improves recovery using only switching wells or valves on and off where optimal control theory is applied that employs an optimization procedure is gradientbased with the adjoint formulation. The conditions under which the reservoirs might have improvements in the Bang-Bang controls are also shown. In this study, we extend this concept to include the geomechanical factor in softer formations where subsidence or uplift are observed with injection or withdrawal of fluids. In this study, the significance of bang-bang control theory is demonstrated on the main reservoir types consisting of different geomechanical characteristics, using a robust commercial optimization and uncertainty/sensitivity software which is coupled with a full-physics commercial simulator. A comprehensive theoretical background is provided to illustrate this approach so that the differences not only in the reservoir characteristics but also the simulation results are shown. The optimization and uncertainty results prove useful in demonstrating the significance of each decision and risk parameter enabling the decision makers to make sound reservoir management decisions under economically-tight environments, providing not only the optimum values for the amount of injection but also the right combination of injection-production strategy to increase net present value or recoveries. The additional knowledge of temporal and spatial geomechanic behavior and the response of reservoirs proves useful in the application of Bang-Bang control with more realistic expectations, thus, leading to better reservoir management decisions. Although Bang-Bang control and application of optimal control theory are not new in the industry, there has not been a comprehensive study taking into account the geomechanics. Thus, this study outlines the behavior and performance of reservoirs that experience subsidence and uplift at different levels in the application of this methodology to waterfloods. 2 SPE-185727-MS
InSAR (Interferometric Synthetic Aperture Radar) is a technology used to measure changes in surface elevation between successive passes of orbiting satellites. These changes can be used to understand imbalances in the subsurface between fluid withdrawal and injection, as well as near-surface ruptures caused by failure of well integrity.
Satellites have recorded SAR data since the 1990s, and the data have become increasingly higher resolution and more frequently acquired. Combined with faster algorithms and processing chains for interferometry, this has enabled detection of smaller and faster changes at the surface. This in turn has caused a step-change in the usefulness of the data and the interpretations. The result is the ability to depend on the data to monitor the effects of production and injection processes almost continuously.
We review several cases to demonstrate the value of rapid revisit, high resolution InSAR. The first is the giant Belridge field in the San Joaquin valley, California, historically the poster child for this application. The diatomite reservoir rock has 60% porosity and is fluid supported. When equilibrium between injection to production is not maintained, the volume changes in the reservoir cause the ground surface to move up or down by amounts detectable with InSAR enabling a feedback loop for injection optimization. The field also has many wells with compromised wellbore integrity that can provide a pathway for reservoir fluids to move upwards towards the ground surface. When water, oil, or steam move out of the reservoir and into the overburden, a potential precursor can be detected provided InSAR is configured carefully. In a second case, InSAR also provides visualizations of ground level changes over gas fields and gas storage fields. At the Groningen gas field in the Netherlands, long term InSAR time series measurements of elevation changes are used to constrain models about compaction and reactivation of buried faults. Parts of the field that are used for seasonal gas storage and charging/discharging cycles can also be effectively monitored.
Measurement of surface deformation by high resolution, fast revisit, optimized InSAR provides an insight into the reservoir and the efficiency of its management. It also provides an early warning of potential problems that, if not corrected, may result in harm to the environment. These step changes in quantity and quality of available InSAR data mean that the remaining barrier to being used for actionable insights is in the widespread inverse modeling of the surface data to sub-surface mass flows.