Parallel Multiscale Reservoir Simulation via Domain Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Reservoir simulations are computationally expensive and time-consuming, particularly when dealing with large datasets and complex models, due to the need for extensive processing resources and memory.
Innovation Solution
The approach involves dividing the reservoir grid into domains and generating coarse grids for each domain, allowing for parallel processing across multiple computer systems, processors, or cores, with calculations such as pressure, flux, and fluid transport performed using coarse grids, and utilizing shared memory for efficient thread scheduling and communication between processing units.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If reservoir simulation uses detailed fine grids to maintain accuracy, then measurement precision is improved, but computational time and resource requirements increase
Solution Approach 1:
The reservoir grid is divided into multiple domains, each processed independently on different processors or computer systems. This segmentation allows parallel computation while maintaining the accuracy benefits of detailed grid representation in each domain, thereby reducing overall computational time without sacrificing simulation precision.
Solution Approach 2:
The patent introduces a new dimension of processing by utilizing parallel computing across multiple processors, cores, or computer systems. This dimensional expansion allows the simulation to maintain fine grid detail for accuracy while distributing computational workload to reduce time requirements.
2Measurement precision
If reservoir simulation uses detailed fine grids to maintain accuracy, then measurement precision is improved, but processing power requirements increase
Solution Approach 1:
By dividing the computational domain into segments that can be processed independently, the patent distributes processing power requirements across multiple processors or systems. Each processor handles a portion of the simulation with adequate detail, maintaining overall accuracy while reducing the processing power burden on any single system.
Solution Approach 2:
The parallel processing framework allows the simulation system to utilize multiple processors, cores, or even different computer systems interchangeably. This multi-functionality enables the system to scale processing power dynamically based on available resources while maintaining simulation accuracy through consistent application of the simulation methodology across all processing units.
3Measurement precision
If reservoir simulation uses short interval timesteps to improve accuracy, then measurement precision is improved, but productivity decreases
Solution Approach 1:
The patent segments the computational workload into independent domain processing tasks that can be executed in parallel. This allows short interval timesteps to be used for accurate simulation while multiple timesteps are processed simultaneously across different processors, thereby maintaining precision while improving overall productivity through parallel execution of timestep calculations.
4Productivity
If reservoir simulation uses parallel processing across multiple systems, then productivity is improved, but device complexity increases
Solution Approach 1:
The patent divides the reservoir simulation into independent domains that can be processed in parallel on multiple systems. This segmentation enables productivity improvement through parallel processing while managing device complexity by creating modular, independently processable units with standardized interfaces for combining results.
Solution Approach 2:
The patent combines results from multiple parallel processing operations to produce the final reservoir simulation. This merging approach allows productivity gains from parallel processing while managing system complexity through a unified framework that integrates results from distributed processing units.
Data Source
AI summary
Systems, computer-readable media, and methods for performing a reservoir simulation by obtaining reservoir data; translating the reservoir data into grid properties to create a grid; dividing the grid into domains; generating coarse grids corresponding to each domain; processing the domains, where processing a domain includes: calculating pressure for the domain using a coarse grid corresponding to the domain, calculating flux for the domain using a coarse grid corresponding to the domain, and calculating transport of fluids for the domain using a coarse grid corresponding to the domain; and generating a reservoir simulation corresponding to the grid based on processing each domain. The domains can be processed in parallel on different computer systems, different processors, or different cores.