2D Domain Decomposition for Parallel Reservoir Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 2D domain decomposition methods in parallel reservoir simulation often result in disproportionate numbers of active cells across domains, leading to poor performance and inefficient parallel processing due to inadequate consideration of active cell distribution.
Innovation Solution
A method for balancing active cells during 2D domain decomposition by iteratively adjusting decomposition boundaries and using stair stepping to ensure even distribution, thereby optimizing the assignment of active cells across domains.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional 2D domain decomposition divides the reservoir model into equal parts between domains, then the decomposition is simple and straightforward, but the number of active simulation cells becomes disproportionate across domains causing poor reservoir simulation performance
Solution Approach 1:
The patent applies local quality by transitioning from uniform equal-partition decomposition to localized decomposition that adapts to the specific distribution of active cells in each domain. The decomposition method now considers local characteristics (active cell concentration) and creates domain boundaries that reflect these local variations, ensuring each domain receives an appropriate number of active cells rather than simply dividing the total model into equal geometric parts.
2Loss of time
If the number of processors is increased to reduce computation time, then parallel processing capability improves, but the complexity of domain decomposition and active cell balancing increases
Solution Approach 1:
The patent applies preliminary action by performing active cell counting and domain decomposition planning before the actual parallel simulation computation begins. The system pre-calculates the optimal distribution of active cells across domains and establishes the decomposition scheme in advance, so that when multiple processors are activated, the workload is already balanced and ready for efficient parallel execution without requiring complex real-time adjustments.
Solution Approach 2:
The patent implements feedback by using the counted active cells information to adjust and optimize the domain decomposition strategy. The system counts active cells in different regions, uses this feedback information to identify imbalances, and then modifies the decomposition boundaries accordingly to achieve better load distribution across processors, creating a closed-loop optimization process.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
Systems and methods for 2D domain decomposition during parallel reservoir simulation to balance the active cells in a reservoir model.