Reservoir Mesh Creation Using Extended Anisotropic Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current reservoir modeling techniques face challenges in accurately simulating fractures within hydrocarbon reservoirs, particularly due to the complexity of fracture networks and the computational inefficiencies of existing mesh generation methods, which hinder precise fluid flow predictions and reservoir management decisions.
Innovation Solution
The implementation of an extended anisotropic, geometry-adaptive refinement (AGAR) algorithm for polyhedra, which allows for the creation of hybrid computational meshes that adaptively refine cells around fractures, enabling more accurate and efficient modeling of complex fracture networks while maintaining computational speed and orientation control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If structured-based meshing or extrusion based meshing is used, then mesh generation is computationally efficient and robust, but the ability to accurately model complex fracture networks is limited due to inherent 2.5 dimensional constraints and difficulty handling arbitrary 3D fracture surfaces
Solution Approach 1:
The mesh is segmented into different regions: structured hexahedral elements in the bulk reservoir and unstructured polyhedral elements around fracture surfaces. This allows each region to be optimized independently - structured mesh for computational efficiency and unstructured mesh for geometric flexibility in capturing complex fracture topologies.
Solution Approach 2:
The patent employs a composite mesh structure combining two different mesh types (structured hexahedral and unstructured polyhedral) within a single reservoir model. This composite approach leverages the advantages of both mesh types to simultaneously achieve computational efficiency and accurate representation of complex fracture networks.
2Adaptability or versatility
If fully unstructured meshing techniques such as tetrahedralization or polyhedral meshing are used, then the ability to model arbitrary 3D fracture surfaces is improved, but robustness decreases and the process becomes more complex especially with imperfect geometry input
Solution Approach 1:
Different mesh qualities and types are applied locally to different regions of the reservoir model. High geometric adaptability with unstructured polyhedral elements is applied only where needed (around fracture surfaces), while simpler structured hexahedral elements are used in the bulk reservoir where geometric complexity is lower, maintaining overall robustness.
Solution Approach 2:
The patent introduces an intermediate processing stage that includes geometry repair and mesh quality assessment steps between the input geometry and final mesh generation. This intermediary process cleans up imperfect geometry input and ensures that both structured and unstructured meshing operations can proceed robustly.
3Measurement precision
If local grid refinement is applied to structured techniques, then local regions can be resolved with finer grids, but the process becomes time-consuming and computationally expensive especially for general reservoir geometries
Solution Approach 1:
The mesh structure is made dynamic through adaptive refinement capabilities. The system can automatically adjust mesh density in response to geometric features, applying finer unstructured polyhedral meshes only in regions with complex fracture surfaces while maintaining coarser structured hexahedral meshes elsewhere, optimizing both resolution and generation speed.
4Manufacturing precision
If higher mesh resolution is used to improve fluid flow simulation accuracy, then fracture properties and connectivity are better captured, but computational requirements increase
Solution Approach 1:
High mesh resolution is applied locally only in regions containing fracture surfaces where accurate fluid flow simulation is most critical. The bulk reservoir regions maintain lower mesh resolution, significantly reducing the total number of cells and computational resources required while preserving simulation accuracy in the most important regions.
Data Source
AI summary
One example herein involves a method that includes identifying, based on a received reservoir specification: a set fractures including 2.5 dimensional (2.5D)-permitting fractures; and other fractures. The method further includes generating an intermediate reservoir model including an extrusion mesh which models the 2.5D-permitting fractures in a three-dimensional (3D) space. In response to determining that cells in the mesh should be refined in a direction within the 3D space, the method anisotropically refines cells in the mesh corresponding to the other fractures. The method also includes resolving a fracture network within the intermediate reservoir model using the refined cells and then generating a reservoir earth model using the fracture network.