Subsurface Horizon Mapping with Deep Learning for Auto-Geosteering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing drilling systems face inefficiencies and inaccuracies in interpreting subsurface geological features, relying heavily on expert interpretation of downhole data, which is labor-intensive and prone to errors, and struggle to effectively utilize seismic and resistivity data for proactive drilling actions.
Innovation Solution
A horizon mapping system using a resistivity image mapping neural network generates horizon maps and augmented resistivity images in real-time, leveraging deep-learning methods and synthetic training data to accurately predict reservoir boundaries and resistivity change interfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If expert interpretation methods are used to analyze downhole data, then measurement precision is improved, but productivity deteriorates due to labor-intensive processes
Solution Approach 1:
The patent replaces manual expert interpretation (mechanical human analysis) with an automated machine learning system that processes downhole data. The system uses trained models to automatically identify subsurface features, eliminating the need for labor-intensive expert analysis while maintaining or improving interpretation accuracy through consistent algorithmic application.
Solution Approach 2:
The system enables self-service by allowing the automated interpretation system to independently analyze downhole data without requiring continuous expert intervention. The machine learning models are trained once and then autonomously process data streams, generating interpretations that can be reviewed or adjusted by experts only when necessary, thereby significantly reducing manual labor while maintaining high productivity.
2Productivity
If automated interpretation systems are implemented, then productivity is improved, but measurement precision deteriorates due to lack of expert judgment
Solution Approach 1:
The system applies preliminary action by training the machine learning models extensively before deployment using curated datasets with expert-labeled examples. This pre-training ensures the models learn accurate interpretation patterns from expert work, embedding expert judgment into the automated system. Once trained, the models can autonomously perform interpretations with high accuracy while maintaining improved productivity through automation.
3Measurement precision
If complex interpretation methods are used, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent extracts the complex interpretation logic from the overall system by encapsulating it within trained machine learning models. The models contain the sophisticated algorithms and patterns needed for accurate subsurface feature identification, while the host system remains relatively simple in structure. This separation allows complex precision-based interpretation to be achieved without making the entire system overly complex, as the complexity is contained within the trained model components rather than the system architecture.
Data Source
AI summary
The disclosure focuses on a drilling system that uses a horizon mapping system to actively determine resistivity change interfaces that form a horizon in subsurface geological features. In various implementations, the horizon mapping system uses a resistivity image mapping neural network to efficiently and accurately generate horizon maps of subsurface geological features, such as reservoirs, from resistivity images. Additionally, the horizon mapping system may generate augmented resistivity images labeled with a horizon map in real time as data and measurements are received.