Optimal Coarse Grid Proxy for Reservoir Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Reservoir simulation models require detailed grids for accuracy, leading to long simulation run times, which is exacerbated by repeated calls to the reservoir simulator during optimization, necessitating a method to reduce computational complexity while preserving simulation model outputs.
Innovation Solution
The development of a 'Coarsening Software' that establishes an optimal coarse grid proxy by averaging material properties and using optimizers to find the best fit of coarse grid outputs to a training set, reducing grid dimensions and simulation time while maintaining accuracy of predefined simulation model outputs like cumulative oil production.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a detailed grid is used in reservoir simulation to capture geological structure complexities, then simulation accuracy is improved, but simulation run time increases
Solution Approach 1:
The patent applies segmentation by dividing the reservoir model into multiple grid blocks of varying sizes. Fine grid blocks are used in areas with complex geological structures where high accuracy is critical, while coarse grid blocks are used in simpler areas to reduce computational burden. This segmented approach allows the model to achieve high accuracy where needed while maintaining acceptable run times overall.
Solution Approach 2:
The patent implements local quality by assigning different grid resolutions to different spatial locations within the reservoir model. Areas with complex fault structures, heterogeneous permeability, or critical production zones are assigned fine grid blocks for detailed representation, while uniform or low-value areas use coarse grid blocks. This localized differentiation optimizes the balance between accuracy and computational cost.
2Measurement precision
If a detailed grid is used to resolve geological complexities, then simulation model output accuracy is improved, but computational complexity increases
Solution Approach 1:
The computational domain is segmented into multiple grid blocks with different resolution levels. This segmentation reduces the overall computational complexity by allowing the solver to process fewer cells in coarse blocks while maintaining detailed representation only where geologically necessary, thus reducing total computational operations required.
Solution Approach 2:
The patent applies partial action by providing high-resolution detailing only to the extent necessary for capturing critical geological features and maintaining accurate simulation outputs. Rather than uniformly high resolution throughout, the model provides fine grid detail selectively in areas where it materially impacts simulation accuracy, accepting approximate representations in less critical areas.
3Measurement precision
If the reservoir simulator is repeatedly called during optimization, then optimization accuracy is improved, but total computational time increases
Solution Approach 1:
The optimal grid configuration is determined through preliminary action by performing multiple simulation runs with different grid block configurations before final optimization. The grid block sizes and arrangements are pre-optimized to balance accuracy and speed, creating an efficient computational framework that minimizes total time across repeated optimization calls.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting grid block configuration parameters (size, distribution, resolution) based on the specific optimization problem requirements and geological characteristics. This allows the model to adapt the computational grid to match the particular simulation scenario, optimizing the trade-off between repeated call accuracy and total computational time for each specific optimization task.
Data Source
AI summary
A method is disclosed for performing optimal gridding in reservoir simulation, the method comprising: establishing an optimal coarse grid proxy that can replace all or parts of a fine grid with a coarse grid while preserving an accuracy of a predefined simulation model output, the step of establishing an optimal coarse grid proxy including finding, by using an optimizer, a best fit of a coarse grid output to the output of a training set.