Reservoir Simulation Misfit Classification Using CNN Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Reservoir simulation models often require time-consuming and error-prone manual adjustments to reduce misfit with field data, which is costly and reliant on engineering expertise.
Innovation Solution
Automated classification of misfit in reservoir simulation models using trained classifiers based on image processing of simulation and field data, allowing for efficient and accurate adjustments to improve model accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual adjustments are used to reduce misfit in reservoir simulation models, then model accuracy can be improved, but the process becomes time-consuming and resource-intensive
Solution Approach 1:
The patent replaces manual mechanical review and adjustment processes with an automated image processing system. Trained classifiers automatically analyze misfit patterns in simulation results by processing visual representations of data, eliminating the need for manual engineer review and enabling rapid identification and correction of model errors.
Solution Approach 2:
The system enables self-service through automated classification and adjustment recommendation. The trained classifiers independently identify misfit types and suggest modifications without requiring continuous human intervention, allowing the simulation model to self-correct through automated feedback loops.
2Reliability
If manual review of simulation results is performed, then misfit can be identified, but the process requires expensive engineering expertise and is error-prone
Solution Approach 1:
The patent substitutes manual expert review with automated image processing classifiers. These trained models systematically analyze simulation results and identify misfit patterns objectively, eliminating human error and the need for specialized engineering expertise while maintaining or improving identification accuracy.
Solution Approach 2:
The system creates a visual copy or representation of the simulation data that can be processed and analyzed. By converting simulation results into image formats that classifiers can process, the system enables automated analysis that replicates and enhances human expert capabilities without requiring actual experts to perform the review.
3Productivity
If automated image processing is used to classify misfit, then processing speed increases, but system complexity increases
Solution Approach 1:
The system performs preliminary action by pre-training classifiers on labeled datasets before deployment. This upfront preparation creates ready-to-use models that can rapidly process simulation results without requiring complex real-time decision logic, achieving high processing speed through pre-computed knowledge.
Solution Approach 2:
The patent transforms simulation data into image parameters suitable for classification. By converting numerical simulation results into visual representations with specific graphical parameters, the system enables the use of image processing algorithms that operate efficiently on standardized input formats, simplifying the processing pipeline.
Data Source
AI summary
The present disclosure relates to computer-implemented methods, software, and systems for automatically assessing simulation results obtained from simulation to predict production of a reservoir. Simulation results and observed field data can be obtained based on executing a simulation model for predicting production of a reservoir in a field for a period of time. The observed field data is obtained from the field and for the reservoir in production during the period of time. A type of misfit of the simulation model can be determined when predicting the production. The one or more trained classifiers are trained to classify the type of misfit is based on an image processing of the observed field data and the simulation results. A modification is determined for the simulation model to adjust future simulation results to reduce the misfit of the simulation model. The simulation model is adjusted based on the determined modification.