Two-Point Flux Approximation for Stable Reservoir Flow Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Reservoir simulation using two-point flux approximation is destabilized by undefined flow directions, particularly in scenarios with thermal variations, leading to increased computational intensity and inefficiency.
Innovation Solution
A modified two-point flux approximation scheme that includes a test for flow stability, applying weighted averaging when flow differences exceed a threshold, and using a modified discretization method to stabilize the simulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If two-point flux approximation is used for reservoir simulation, then computational simplicity is maintained, but simulation stability deteriorates when flow directions are undefined
Solution Approach 1:
The patent applies dynamics by making the discretization scheme adaptive rather than fixed. The method dynamically switches between two-point flux approximation and weighted averaging based on flow direction stability. When flow directions are well-defined, the simpler two-point scheme is used; when undefined (particularly in thermal scenarios), the method transitions to the more stable weighted averaging approach, allowing the system to adapt its computational complexity to maintain stability.
Solution Approach 2:
The patent changes the discretization parameter based on flow conditions. By monitoring flow direction stability and switching between different discretization schemes (two-point flux approximation vs. weighted averaging), the method adjusts its numerical approach to match the physical conditions, thereby maintaining both simplicity when possible and stability when required.
2Reliability
If two-point flux approximation is used with undefined flow directions, then computational overhead increases due to extra compute cycles, but the method continues to attempt flow direction determination
Solution Approach 1:
The patent extracts the problematic scenario (undefined flow directions) from the general simulation process by identifying it through a stability test. When undefined flow directions are detected, the method removes the two-point flux approximation approach and replaces it with weighted averaging, thereby eliminating the source of computational inefficiency while maintaining reliable flow direction determination through the alternative method.
Solution Approach 2:
The patent implements feedback through a stability test that monitors flow direction definition. Based on the test results, the method provides feedback to select the appropriate discretization scheme. This feedback mechanism prevents unnecessary computational cycles by avoiding the application of two-point flux approximation in scenarios where it would fail, thereby improving computational efficiency while maintaining reliability.
3Stability of the object's composition
If weighted averaging is applied when flow differences exceed a threshold, then simulation stability is improved, but computational complexity increases
Solution Approach 1:
The patent applies dynamics by making the discretization scheme adaptive rather than fixed. The method dynamically switches between two-point flux approximation and weighted averaging based on flow direction stability. When flow directions are well-defined, the simpler two-point scheme is used; when undefined (particularly in thermal scenarios), the method transitions to the more stable weighted averaging approach, allowing the system to adapt its computational complexity to maintain stability.
Solution Approach 2:
The patent applies local quality by applying different discretization methods to different spatial locations or scenarios based on local flow conditions. Rather than using a single method throughout the entire simulation, the approach tailors the numerical method to local requirements - using simple two-point approximation where flow is well-defined and switching to weighted averaging only where flow directions are undefined, thereby minimizing overall computational complexity while maintaining stability where needed.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for performing a modified two point flux approximation scheme is disclosed. The method includes: obtaining a first pressure value for a first neighbor cell and a second pressure value for a second neighbor cell, where the first neighbor cell has a first value of a reservoir property and the second neighbor cell as a second value of the reservoir property; determining a first weight using the first pressure value and a second weight using the second pressure value; calculating a third value of the reservoir property as a weighted average of the first value and the second value; and applying the third value to the first neighbor cell.