3D Grid Generation for Heterogeneous Reservoirs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for constructing 3D grids in subsurface reservoir modeling require complex geometric post-processing and inverse mapping, which are time-consuming and inefficient, especially when dealing with heterogeneous material properties and faulted formations.
Innovation Solution
The method generates a grid geometry directly in physical space, eliminating the need for a grid in design space, allowing for flexible handling of complex geometries and reducing the complexity of mapping by only requiring forward mapping, with grid cells being truncated or conforming to fault surfaces based on property contrasts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If complex geometric post-processing and inverse mapping are used to construct 3D grids in faulted formations, then the grid can conform to fault surfaces, but the process becomes time-consuming and inefficient
Solution Approach 1:
Instead of mapping from physical space to design space and back (inverse mapping), the patent inverts the approach by directly constructing the grid in physical space using forward mapping only. The grid cells are defined by mapping design space coordinates to physical space coordinates, eliminating the need for complex inverse mapping and geometric post-processing while maintaining conformity to fault surfaces.
Solution Approach 2:
The patent extracts and removes the problematic inverse mapping step from the traditional workflow. By directly constructing grids in physical space using only forward mapping, the method eliminates the time-consuming geometric post-processing and inverse mapping operations while preserving the essential functionality of creating fault-conforming grids.
2Stability of the object's composition
If a grid is constructed in design space and then mapped back to physical space, then the material properties can be described continuously, but the process requires complex inverse mapping and post-processing
Solution Approach 1:
The patent inverts the traditional mapping sequence by constructing the grid directly in physical space rather than in design space. This is achieved by using forward mapping to transform grid cells from design space coordinates to physical space coordinates, thereby eliminating the need for inverse mapping and complex geometric post-processing while maintaining continuous material property descriptions.
Solution Approach 2:
The patent segments the complex mapping problem into simpler components by constructing the grid cell by cell in physical space. Each grid cell is independently mapped using forward mapping, and the collection of these mapped cells forms the complete grid structure, avoiding the need for complex global inverse mapping operations.
3Loss of information
If traditional mapping methods are used to handle heterogeneous material properties, then the material properties can be represented, but the process requires well-defined inverse mapping and complex geometric post-processing
Solution Approach 1:
The patent inverts the traditional approach by directly constructing grids in physical space using forward mapping, thereby eliminating the need for inverse mapping. This maintains accurate representation of heterogeneous material properties while significantly simplifying the grid generation process and reducing computational complexity.
Data Source
AI summary
Method for generating a 3D grid, and for defining a material property model on the grid, to use, for example, in a reservoir simulator. A mapping is defined (61,71) to a design space in which the material property is described as a piecewise smooth implicit or explicit function in three dimensions. Grid geometry is constructed only in the physical space of the model (62-65,73-76), and no grid is required in the design space. The material property, for example permeability, is sampled in the design space (66,77) to populate the cells in the grid constructed in the physical domain. Prismatic grid cells may be truncated based on faults and horizons (65), or maybe conformed to fault surfaces using a 3D parameterization of the model (76). Only forward mapping, i.e. from the physical domain to the design space, is required.