Accurate representation of near-well flow is an important subject matter in reservoir simulation. In today’s field-scale reservoir simulation, cell-centered structured grids remain the predominant practice. Typically, well inflow performance of the perforated cells is connected to the finite-volume solution via well indices which may not be well defined when the wellbore intersects the finite-volume cells in a complex trajectory. Fine gridding is required to resolve the flow dynamics in the near-well regions. Strong grid orientation sensitivities can also contribute to the numerical errors and may require significant local grid refinement to alleviate. There are on-going R&D efforts on applying unstructured grids to better represent the near-well flow in reservoir simulation, but their applications are mainly in single-well study or sector modeling with a few wells. Some of the reasons cited for this include: (1) the lack of an effective, easy-to-use full-field complex well gridding tool; (2) the lack of supporting unstructured workflow for full-cycle reservoir simulation; (3) the cost of unstructured grid simulation; (4) the availability of post analysis and visualization tools for unstructured grid simulation. The paper describes a method to generate unstructured grids which conform to complex well paths in field-scale simulation. The method uses a multi-level approach to place cells optimally within the solution domain based on the "regions of interests". The wellbore geometry is honored via the construction of a near-well grid that is complemented with multi-level quad-tree refinements to achieve the desired resolution in grid transition zones. The method includes an algorithm to remove small cells and pinching cells based on local grid quality measures and cell prioritization to honor well paths. The gridding process forms a component of a production level reservoir simulation workflow. The use of unstructured grid results in computational saving by placing cells where the resolution is needed. An in-house massively parallel simulator is used to run the unstructured grid models. Simulation examples for full-field applications with hundreds of complex wells using both structured grids and unstructured grids will be used to compare results, accuracy, and performance of the gridding method for reservoir simulation.

Summary Accurate representation of near-well flow is an important subject matter in reservoir simulation. In today's field-scale reservoir simulation, cell-centered structured grids remain the predominant practice. Typically, well-inflow performance of the perforated cells is connected to the finite-volume solution by means of well indices that may not be well-defined when the wellbore intersects the finite-volume cells in a complex trajectory. Fine gridding is also required to resolve the flow dynamics in the near-well regions. Strong grid-orientation sensitivities can also contribute to the numerical errors and may require significant local grid refinement to alleviate. There are ongoing resesarch-and-development (R&D) efforts on applying unstructured grids to better represent the near-well flow in reservoir simulation, but their applications are mainly in single-well study or sector modeling with a few wells. Some of the reasons cited for this include (1) the lack of an effective, easy-to-use full-field complex well-gridding tool; (2) the lack of supporting unstructured workflow for full-cycle reservoir simulation; (3) the cost of unstructured-grid simulation; and (4) the availability of post-analysis and visualization tools for unstructured-grid simulation. The paper describes a novel method to automatically generate unstructured grids that conform to complex well paths in field-scale simulation. The method uses a multilevel approach to place cells optimally within the solution domain on the basis of the “regions of interests.” The wellbore geometry is honored by means of the construction of a near-well grid that is complemented with multilevel quad-tree (Fig. 1) refinements to achieve the desired resolution in grid transition zones. The method includes an algorithm to remove small cells and pinching cells on the basis of local grid quality measures and cell prioritization to honor well paths. The gridding process forms a component of a production-level reservoir-simulation workflow. The use of unstructured grid results in computational savings by placing cells where the resolution is needed. An in-house massively parallel simulator is used to run the unstructured-grid models. Simulation examples for full-field applications with hundreds of complex wells by use of both structured grids and unstructured grids will be used to compare results, accuracy, and performance of the gridding method for reservoir simulation.

Reservoir simulation predominately uses structured grids, but they have difficulties for field-scale simulation with maximum reservoir contact (MRC) wells. In this work, the grid resolution for numerical convergence using structured grids is studied. Results showed that sufficiently fine grid is needed to obtain converged solution. In common practice, the geocellular models are simply upscaled and used in history matching. This study highlights the need for grid resolution to resolve flow dynamics in reservoir simulation and improve well inflow performance calculation. Otherwise, the near-well flow may not have converged and the fidelity of the models for performance prediction is in question.Improving the fidelity of simulation results by adequately modeling the complex multiphasic flow dynamics near complex wells in full-field simulation is of paramount importance nowadays. This work introduces a full-field unstructured gridding method which can optimally place unstructured grid cells where the resolution is needed. The method produces a consistent discretization that is efficient to compute by using a parallel unstructured reservoir simulator. For a giant Middle-East carbonate reservoir that was developed primarily using complex MRC wells, unstructured grid models were used to improve the accuracy of near-well flow and to better represent well inflow performances. The unstructured grid models are compared against the original structured grid models that show computational cost saving and require few grid cells.Simulation results demonstrate an unstructured workflow that is practical and can be used to validate near-well modeling accuracy for the existing structured grid simulation results. The method is particular attractive for situation with denselyspaced complex wells where a structured local grid refinement (LGR) method will be ineffective. An unstructured grid is well suited to honor the near-well flow geometry and to focus grid resolution where it is needed. Perpendicular bisection (PEBI) grids are orthogonal by construction. This reduces computational complexity because two-point flux approximation (TPFA) can be applied. In field-scale simulation, this results in a significant improvement to accuracy and computational cost savings. TX 75083-3836, U.S.A., fax +1-972-952-9435

No abstract

Machine learning based intelligent automation is developed by extending a prior workflow of unstructured grid reservoir simulation (Ding et al. 2015; Fung et al. 2014, 2013). Reservoir heterogeneity, geological internal boundary features and well geometry complexity are being taken into account to automatically detect well zone and focusing reservoir area by calculating the region-of-interests in the model and define cell spacing for grid coarsening and refinement in the reservoir. Automated workflow is demonstrated by using an unstructured grid reservoir simulation example. Previously, an unstructured grid reservoir simulation workflow is introduced (Ding et al. 2015; Fung et al. 2014, 2013). One major component of the workflow is the near-well unstructured grid modeling framework which consists of a 2.5D unstructured PEBI grid engine and its input criteria, such as the locations of reservoir where grid coarsening and refinement are being applied in, along with respective cell spacing being allocated. The creation of such regions of interest and selection of cell spacing involve user's manual interaction, which is user-experience dependent and not intended to serve as a long term solution in the simulation workflow. This paper enhances this process by automatically detecting the target area in the reservoir by computing the convex hull of the well dataset or modified convex hull if concave exists in the dataset. The convex hull is used as the basis for reservoir polygon with defined cell spacing computed by cell density control scheme. The automation considers the heterogeneity and complexity of the reservoir, such as geological internal boundaries and complicated well geometry. The targeted locations in the reservoir cover necessary area for grid refinement with high density grids to capture the accurate flow dynamics near the well, whereas unimportant area in the reservoir are detected for being coarsened to avoid extreme large model size and long simulation runtime. The result of this work enhances the unstructured grid modeling process by automatically computing the local reservoir areas for grid coarsening and refinement with respective grid density on the multi-level hierarchical grids, it avoids user's manual interaction, which is neither efficient nor user friendly. The automated workflow improves unstructured gridding efficiency and enhances user's simulation experience. Utilizing emerging technology breakthrough such as machine learning is important towards a successful era of the Fourth Industrial Revolution (4IR), this work of workflow automation is an example of using machine learning for enhancing problem solving in reservoir simulation.