ML-Based Inversion Model Selection for Accurate Earth Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating earth models from seismic data are limited in accuracy and efficiency, hindering effective resource extraction and characterization of subsurface regions.
Innovation Solution
A method utilizing a trained machine learning model to select the most appropriate model from a plurality of models and invert seismic data to determine parameters, enhancing the accuracy of earth model generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional seismic tomography methods are used to generate earth models, then subsurface imaging can be achieved, but the accuracy and precision of the models are limited due to data interpretation challenges and model complexity
Solution Approach 1:
The patent introduces machine learning models as an intermediary between raw seismic data and earth model generation. The ML models preprocess and interpret seismic data, automatically selecting appropriate inversion models and parameters, thereby reducing interpretation complexity while improving model accuracy through learned patterns from training data.
Solution Approach 2:
The patent changes the parameters of the inversion process by using ML-predicted model selections and parameter optimizations. Instead of relying on traditional fixed-parameter inversion methods, the system dynamically adjusts inversion parameters based on ML model predictions, improving measurement precision while managing complexity through automated parameter selection.
2Measurement precision
If complex inversion models are used to improve subsurface characterization accuracy, then model precision improves, but computational resources and processing time increase
Solution Approach 1:
The patent applies partial action by using ML models to select only the necessary level of inversion complexity for each specific geological scenario. Instead of always applying the most complex inversion models, the system uses ML predictions to determine the appropriate model complexity, achieving sufficient accuracy while reducing unnecessary computational resource consumption.
Solution Approach 2:
The patent performs preliminary action by using machine learning models to pre-select appropriate inversion models and parameters before the actual inversion process. This preliminary ML-based model selection prevents the system from wasting computational resources on inappropriate complex models, thereby reducing energy consumption while maintaining the necessary precision for accurate subsurface characterization.
3Productivity
If manual model selection and interpretation processes are used, then flexibility in analysis can be maintained, but productivity and efficiency of resource extraction operations decrease
Solution Approach 1:
The patent implements self-service by enabling the system to automatically select inversion models and interpret seismic data without requiring manual intervention. The machine learning models autonomously analyze seismic data, select appropriate inversion parameters, and generate earth models, significantly improving productivity and resource extraction efficiency while maintaining adequate flexibility through the ability to review and adjust ML-generated results.
Data Source
AI summary
A method can include receiving data for a geologic region; based at least in part on the data, selecting a model from a plurality of models using a trained machine learning model, and inverting the data using the selected model to determine parameters of the selected model.