Reservoir Model Optimization Under Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In oilfield operations, existing computer models of wells face challenges in effectively managing uncertainty, which affects decision-making and reservoir development, as they often require simulating complex interactions between multiple wells and equipment parameters, leading to increased computational complexity and reduced efficiency.
Innovation Solution
A method and system that receive realizations of a reservoir model with uncertainty, select a portion of these realizations in a reduced dimensional space to preserve uncertainty, optimize an objective function based on the selected realizations, and generate a field operations plan using the optimized parameter values, employing techniques like smart sampling and multidimensional scaling to reduce dimensionality and account for equipment conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If computer models simulate complex interactions between multiple wells and equipment parameters to manage uncertainty, then the reliability of decision-making is improved, but the device complexity and computational overhead increase
Solution Approach 1:
The patent applies dimensionality reduction techniques to transform high-dimensional uncertainty representations into lower-dimensional spaces while preserving essential uncertainty characteristics. This allows the system to maintain reliable decision-making under uncertainty without the computational burden of processing full high-dimensional realizations.
Solution Approach 2:
The system extracts and selects only the most representative uncertainty realizations from the full set, filtering out redundant information while preserving the essential uncertainty structure needed for reliable decisions. This extraction process reduces computational complexity while maintaining decision-making reliability.
2Measurement precision
If the system processes all model realizations to accurately represent uncertainty, then the measurement precision of uncertainty is improved, but the loss of time and computational efficiency worsen
Solution Approach 1:
Instead of processing all model realizations, the system selectively processes only the necessary subset that provides sufficient uncertainty representation. This partial action approach maintains adequate measurement precision while significantly reducing computational time and efficiency losses.
Solution Approach 2:
By transforming uncertainty representations into reduced dimensional spaces, the system achieves accurate uncertainty characterization with fewer computational operations, thereby reducing time loss while maintaining precision.
3Device complexity
If the system uses reduced dimensional space to select realizations, then the device complexity is reduced, but the loss of information about uncertainty may increase
Solution Approach 1:
The system incorporates feedback mechanisms to verify that the reduced dimensional representation adequately captures uncertainty characteristics. This feedback loop ensures that information loss is minimized while maintaining reduced computational complexity.
Solution Approach 2:
The system transforms uncertainty representations by changing parameters and dimensions, creating a reduced dimensional space that preserves essential uncertainty information while reducing computational complexity. The transformation is designed to maintain information fidelity.
Data Source
AI summary
A method can include receiving realizations of a model of a reservoir that includes at least one well where the realizations represent uncertainty in a multidimensional space; selecting a portion of the realizations in a reduced dimensional space to preserve an amount of the uncertainty; optimizing an objective function based at least in part on the selected portion of the realizations; outputting parameter values for the optimized objective function; and generating at least a portion of a field operations plan based at least in part on at least a portion of the parameter values.