Coupled Pipe Network Reservoir Modeling for Multi-Branch Wells
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current reservoir simulators face challenges in modeling fluid flow in multilateral wells due to the distinct flow characteristics of porous media and pipe flow, leading to difficulties in achieving stable pressure distribution and convergence in computational models, particularly due to ill-conditioned matrices and time step size limitations.
Innovation Solution
A new data processing system and methodology that treats multilateral well flow as two-dimensional steady flow, using a novel numbering scheme to create nearly tridiagonal coefficient matrices, and employs the Newton-Raphson method with a linear iterative solver to determine well pressures and flow rates, considering the fluid flow characteristics of porous media and pipe flow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a decoupled sequential approach is used to model reservoir and well flow, then computational stability is improved, but solution accuracy deteriorates due to time step size limitations and inability to capture strong interactions
Solution Approach 1:
The model segments the wellbore into discrete intervals along its length, allowing the coupled system to be divided into manageable computational units that can be solved simultaneously while maintaining stability and accuracy
Solution Approach 2:
The reservoir and wellbore flow equations are merged into a single coupled system of equations that is solved simultaneously, capturing the strong interactions between the two media while maintaining computational stability through proper matrix formulation
2Measurement precision
If a fully coupled solution is used to model reservoir and well flow, then solution accuracy is improved, but computational cost increases and convergence problems occur due to ill-conditioned matrices
Solution Approach 1:
The model applies different numerical treatment to different parts of the system - using a specialized numbering scheme and matrix formulation for the wellbore portion while maintaining standard reservoir modeling for the formation, optimizing computational efficiency for each domain
Solution Approach 2:
The model transforms the governing equations into a form with improved numerical properties by changing the arrangement of unknowns and the structure of the coefficient matrix, converting an ill-conditioned system into a well-conditioned one that converges efficiently
3Measurement precision
If pipe flow equations are solved with Newton Raphson iterations, then solution accuracy is improved, but convergence fails unless a good initial estimate is provided which is difficult to obtain for complex networks
Solution Approach 1:
The model performs preliminary ordering and arrangement of the system equations before applying Newton Raphson iteration, creating a structured initial formulation that ensures reliable convergence without requiring sophisticated initial estimates
Solution Approach 2:
The model incorporates feedback mechanisms through the iterative solution process where the structured matrix formulation and numbering scheme provide continuous correction guidance, ensuring convergence even when initial estimates are not perfect
Data Source
Figure 1~2
Figure 3~5
Figure 6~7
AI summary
A convergent solution is provided for a coupled system where oil flow from a subsurface reservoir formation enters a number of pipes of a multi-branch well in the formation. An iterative linear system solver computer implemented methodology is developed, capable of handling a large number of unknowns which are present when modeling a multi -branch well A systematic approach which defines proper boundary conditions at the reservoir level and at the wellhead is prov4 l ided and utilized.