Publisher Wiley RightsYou are free to: Share -copy and redistribute the material in any medium or format The licensor cannot revoke these freedoms as long as you follow the license terms. Under the following terms: Attribution -You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use. NonCommercial -You may not use the material for commercial purposes. NoDerivatives -If you remix, transform, or build upon the material, you may not distribute the modified material. No additional restrictions -You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
Polymer flooding is an effective enhanced oil recovery technology used to reduce the mobility ratio and improve sweep efficiency. A new polymer injection scheme is investigated that relies on the cyclical injection of low-salinity, low-concentration polymer slugs chased by high-salinity, high-concentration polymer slugs. The effectiveness of the process is a function of several reservoir and design parameters related to polymer type, concentration, salinity, and reservoir heterogeneity. We use reservoir simulations and design-of-experiments (DoE) to investigate the effectiveness of the proposed polymer injection scheme. We show how key objective functions, such as recovery factor and injectivity, are impacted by the reservoir and design parameters. In this study, simulations showed that the new slug-based process was always superior to the reference polymer injection scheme using the traditional continuous injection scheme. Our results show that the process is most effective when the polymer weight is high, corresponding to large inaccessible pore-volumes, which enhances polymer acceleration. High vertical heterogeneity typically reduces the process performance because of increased mixing in the reservoir. The significance of this process is that it allows for increased polymer solution viscosity in the reservoir without increasing the total mass of polymer, and without impairing polymer injectivity at the well.
Modeling fluid flow in fractured media is of importance in many disciplines, including subsurface water management and petroleum reservoir engineering. Detailed geological characterization of a fractured reservoir is commonly described by a discrete-fracture model (DFM), in which the fractures and rock-matrix are explicitly represented by unstructured grid elements. Traditional static-based and flow-based upscaling methods used to generate equivalent-continuum models from DFM suffer from low accuracy and high computational cost, respectively. This work introduces a new deep-learning technique based on neural networks to accelerate upscaling of discrete-fracture models. The objective of this work is to automate the process of permeability upscaling from detailed discrete-fracture characterizations. We build an "image-to-value" model to map the nonlinear relationship between high-resolution DFM images, provided as input, and the equivalent-continuum model (output), which comprises the predicted equivalent permeability of each grid-block. The proposed upscaling workflow incorporates the generation of the training datasets, design of the neural network architecture, and the validation process. The implemented deep neural networks consist of 20 layers including 18 hidden layers. Good quality input data and suitable network structure are crucial to obtaining a successful trained model. Therefore, high-resolution simulations were used to generate the training set. A training-validating process is conducted to update of training data and to optimize the network architecture until the trained model reaches an accuracy exceeding 95%. We first verify the deep learning-based (DL-based) approach by applying it to a set of single rotating fractures with known analytical solutions. We then demonstrate it on a synthetic DFM including connected and disconnected fractures, and on a DFM from an actual outcrop. We compare the method performance to reference fully-resolved solutions and to an advanced flow-based upscaling method, referred to as a multi-boundary fracture upscaling method. In the tested cases, simulation results show that the DL-based method is computational more efficiency than the flow-based method by two orders of magnitude without significant loss of accuracy. This work demonstrates the potential of a physics-based DL approach for the upscaling of high-resolution images of fractured media. The proposed DL approach is more accurate than the static-based upscaling methods and more efficient than the flow-based upscaling methods, and therefore, has the potential to be used to improve the computational performance of CPU-intensive upscaling methods.
History matching is critical in subsurface flow modeling. It is to align the reservoir model with the measured data. However, it remains challenging since the solution is not unique and the implementation is expensive. The traditional approach relies on trial and error, which are exhaustive and labor-intensive. In this study, we propose a new workflow utilizing Bayesian Markov Chain Monte Carlo (MCMC) to automatically and accurately perform history matching. We deliver four novelties within the workflow: 1) the use of multi-resolution low-fidelity models to guarantee high-quality matching, 2) updating the ranges of priors to assure convergence, 3) the use of Long-Short Term Memory (LSTM) network as a low-fidelity model to produce continuous time-response, and 4) the use of Bayesian optimization to obtain the optimum low-fidelity model for Bayesian MCMC runs. We utilize the first SPE comparative model as the physical and high-fidelity model. It is a gas injection into an oil reservoir case, which is the gravity-dominated process. The coarse low-fidelity model manages to provide updated priors that increase the precision of Bayesian MCMC. The Bayesian-optimized LSTM has successfully captured the physics in the high-fidelity model. The Bayesian-LSTM MCMC produces an accurate prediction with narrow uncertainties. The posterior prediction through the high-fidelity model ensures the robustness and precision of the workflow. This approach provides an efficient and high-quality history matching for subsurface flow modeling.
The tectonic setting of Saudi Arabia enriches the country with significant geothermal resources, such as those in Al-Lith and Jizan in the southwestern area. Recently, there has been interest to explore the geothermal potential to diversify the country's energy-mix, which is driven by the Kingdom's Vision 2030. One key challenge in geothermal systems is in their low efficiency compared to traditional hydrocarbon-fired plants. This inefficiency is related to the thermal flow behavior in the subsurface and to the energy conversion technology at the surface. In this study, we provide a workflow for feasibility assessment of geothermal reservoir development with potential application in Saudi Arabia. The proposed workflow is within the Design of Experiment (DoE) framework, which allows conducting numerous simulations with low computational cost. Computations are performed using a proxy modeling approach, which reflects a multidimensional response-surface emerging from the optimization problem. Two steps in the workflow were found to be critical. First, identify and select the most significant uncertainty parameters to focus the design. Second, address the nonlinearity of the problem by filling up any potential gaps within the response space. In this work, two-level folded Plackett-Burman design is used to identify and select the most significant parameters relative to the energy recovery and enthalpy production factors. Three-level Taguchi design is then applied to create a more rigorous proxy model. We used a space-filling technique to address lack of sampling and nonlinearity in the response surface. Monte Carlo simulations are performed, at the final stage, to generate probabilistic forecasts under uncertainties. The energy recovery factor and the enthalpy production behavior are found to be influenced by the volume of the reservoir, rock permeability and porosity, heterogeneity, well spacing, and fluid production rate. Our Monte Carlo simulations show that, at the Jizan's geothermal conditions, the energy recovery factor is within 12% to 24%, which is encouraging as they are above the typical recovery factor of 10%-17% worldwide.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.