Sedimentary Basin Uncertainty Modeling With PCA Surrogates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for quantifying uncertainties in sedimentary basin modeling require numerous simulations, which are computationally intensive and time-consuming, making them impractical for operational studies with tight deadlines.
Innovation Solution
A method combining sequential planning with adaptation and reduced basis decomposition is used to build approximate analytical models for spatial outputs, iteratively improving accuracy while minimizing the number of simulations and computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Monte Carlo type sampling methods are used to generate a large set of models for uncertainty quantification, then the accuracy of uncertainty estimates is improved, but the simulation time and computational resources increase significantly
Solution Approach 1:
The patent creates simplified copies (surrogate models) of the complex basin simulation model. These surrogate models are trained on a subset of simulation data and can rapidly predict basin properties without requiring full-scale numerical simulations. This allows uncertainty quantification to be performed on many more model realizations than would be feasible with the original complex model, thereby improving accuracy while reducing computational time.
Solution Approach 2:
The patent employs inexpensive, computationally lightweight surrogate models that can be rapidly evaluated thousands of times. These surrogate models serve as disposable approximations that replace the need for repeated expensive full-scale simulations, enabling extensive Monte Carlo sampling and sensitivity analysis within operational timeframes.
2Reliability
If a large number of simulations are performed to obtain representative samples for risk analysis, then the reliability of risk assessment is improved, but the productivity of operational studies decreases
Solution Approach 1:
Surrogate models serve as rapid copies that replicate the behavior of complex basin simulations. These copies enable researchers to perform extensive risk assessments with thousands of simulation realizations, achieving high statistical reliability without the computational burden that would reduce productivity in operational settings.
Solution Approach 2:
The patent transforms the computational parameters of the simulation system by replacing the original complex numerical model with simplified surrogate models having different computational characteristics. This parameter change allows the system to evaluate many more scenarios within the same time frame, simultaneously improving reliability through larger sample sizes and maintaining productivity.
3Measurement precision
If sequential planning methods with adaptation are applied to improve the accuracy of approximate analytical models, then the precision of spatial distribution estimates is improved, but the complexity of the modeling process increases
Solution Approach 1:
The patent implements sequential planning with adaptation, where the surrogate model is iteratively improved based on feedback from validation against observed data. The model identifies regions of parameter space or spatial locations where predictions are less accurate and focuses computational resources on improving those specific areas, thereby increasing precision while managing complexity through targeted refinement.
Solution Approach 2:
The modeling process is made dynamic and adaptive rather than static. The surrogate model evolves through iterative training and validation cycles, automatically adjusting its complexity and focus based on the data at hand. This dynamic approach allows the model to achieve high precision where needed while maintaining computational efficiency elsewhere.
Data Source
Figure 1~2
Figure 3
AI summary
The invention relates to a method for determining the uncertainties in a property of a sedimentary basin, wherein a plurality of instances of a spatial distribution of the property is determined for a plurality of combinations of uncertain parameters of a stratigraphic simulation or of a basin simulation, a principal component analysis is applied to this plurality of instances, and an approximate analytical model of the spatial distribution of the property is determined by constructing an approximate analytical model for a selection of components the sum of the eigenvalues of which is higher than a predefined threshold. Subsequently, the approximate analytical model is improved iteratively by determining, in each iteration, at least one additional combination of the uncertain parameters by means of sequential planning with adaptation applied to the approximate analytical models of the selected components considered in decreasing order. Thereafter, on the basis of the approximate analytical model, the uncertainties in the property are determined.