Stochastic Inversion for Well Success Probability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining well success rely on subjective or non-physics based statistical approaches, leading to potentially erroneous estimates of well success probabilities, lacking a quantitative decision-risk method.
Innovation Solution
A computer-implemented method using stochastic inversion to generate posterior distributions of earth models, calculating well production and cost distributions based on earth parameters, and applying probability-weighted values to determine the best well location from multiple possible locations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional non-physics based statistical methods are used to estimate well success probabilities, then the decision-making process is simplified, but the accuracy and reliability of the probability estimates deteriorate due to subjective assessment
Solution Approach 1:
The patent replaces conventional non-physics based statistical methods with stochastic inversion that incorporates physics-based rock formation models. This substitution transforms the decision support system from subjective statistical assessment to objective physics-based probability estimation, thereby improving measurement precision while maintaining computational feasibility through automated inversion algorithms
Solution Approach 2:
The patent changes the fundamental parameters used in probability estimation from subjective statistical metrics to physics-based parameters derived from stochastic inversion of geophysical data. By transforming the input parameters from conventional well logs and seismic data into probabilistic rock properties (permeability, porosity, saturation) through physics-based models, the system achieves more accurate and reliable well success probability estimates
2Measurement precision
If stochastic inversion is used to generate posterior distributions for quantitative risk assessment, then the accuracy of well success probability estimation is improved, but the computational complexity and data processing requirements increase
Solution Approach 1:
The patent segments the complex stochastic inversion process into modular computational steps: (1) generating multiple plausible earth models consistent with observed data, (2) performing reservoir simulation on each model, (3) aggregating results to compute posterior distributions, and (4) integrating with decision support systems. This segmentation enables the complex computational task to be distributed and managed efficiently, reducing the practical burden of computational complexity
Solution Approach 2:
The patent introduces an intermediary layer of probabilistic rock property models that bridge the gap between raw geophysical data and well success predictions. This intermediary stochastic inversion framework transforms complex multi-parameter geophysical datasets into simplified probabilistic distributions of critical reservoir properties, thereby managing computational complexity while preserving measurement precision
3Reliability
If subjective risk assessment is replaced with quantitative stochastic inversion, then the reliability of decision-making is improved, but the ease of operation deteriorates due to complex data processing requirements
Solution Approach 1:
The patent implements self-service through automated stochastic inversion algorithms that independently process geophysical data and generate posterior distributions without requiring manual statistical analysis. The system automatically performs rock property inversion, reservoir simulation, and probability calculation, reducing the operational burden on users while maintaining high reliability through consistent physics-based computations
Data Source
AI summary
A system and a computer implemented method for determining a best well location from a plurality of possible well locations are described herein. The method includes drawing a plurality of earth models from a posterior distribution, wherein the posterior distribution is generated by stochastic inversion of existing data; calculating a well production at a plurality of proposed well locations within an earth model in the plurality of earth models using a relationship between the well production and earth parameters; calculating from the plurality of earth models, cost distributions using the relationship between well cost and the earth parameters; and calculating probability weighted values for the proposed well locations using probabilities from location dependent stochastic inversions as weights.