Machine Learning Geosteering for Accurate Pay-Zone Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing geosteering methods struggle to accurately control the directional wellbore within a pay zone, leading to potential gas or water breakthrough and reduced hydrocarbon production efficiency.
Innovation Solution
Utilizing downhole geological logging measurements with a trained machine learning model to invert data and generate structural features of the subsurface region, enabling precise control of the drill bit operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional geosteering methods are used to control directional wellbore, then the control process is simpler, but the accuracy of maintaining wellbore within pay zone deteriorates
Solution Approach 1:
The patent replaces traditional mechanical/geological interpretation methods with machine learning algorithms that automatically process downhole logging measurements. The ML model substitutes complex manual analysis with automated pattern recognition, achieving higher positioning accuracy without proportionally increasing system complexity.
Solution Approach 2:
The patent introduces machine learning models as an intermediary between raw downhole measurements and geosteering decisions. This intermediary layer processes and interprets complex geological data, translating it into actionable wellbore positioning information that improves control accuracy.
2Loss of information
If machine learning model is used to invert downhole data, then the structural feature identification improves, but the computational complexity increases
Solution Approach 1:
The patent performs preliminary training of machine learning models using synthetic and field data before actual geosteering operations. This pre-processing step creates ready-to-use models that can quickly invert downhole measurements during drilling, reducing real-time computational complexity while maintaining high structural feature identification accuracy.
Solution Approach 2:
The patent uses synthetic data that copies real subsurface conditions to train machine learning models. This approach creates virtual training scenarios that teach the models to recognize actual geological structures, improving feature identification without requiring exhaustive real-data processing during operations.
3Productivity
If real-time data-driven control is implemented, then the hydrocarbon production efficiency improves, but the operational complexity increases
Solution Approach 1:
The patent implements a closed-loop feedback system where downhole measurements are continuously processed by machine learning models, and control decisions are automatically adjusted based on real-time wellbore positioning relative to pay zones. This continuous feedback maintains high production efficiency while automating operations to reduce manual complexity.
Solution Approach 2:
The patent enables the geosteering system to self-adjust wellbore trajectory using automated machine learning-based interpretation of downhole data. The system autonomously determines optimal drilling directions to maintain wellbore within pay zones, reducing the need for continuous manual intervention and simplifying operations while maximizing hydrocarbon production.
Data Source
AI summary
A system and method may include receiving data acquired by a downhole tool of a tool string disposed at least in part in a borehole in a subsurface region. The system and method may also include inverting the data using a trained machine learning model to generate a structural feature of the subsurface region. The system and method may further include controlling operation of the tool string based at least in part on the structural feature of the subsurface region.