EM Formation Classification for Real-Time Geosteering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing geosteering operations face challenges in navigating complex subterranean formations with non-planar boundaries, as 1D inversion models are inadequate and 2D/3D models are too data-intensive for real-time applications.
Innovation Solution
Employing a machine learning-based approach to classify subterranean formations using electromagnetic logging tools, distinguishing between 1D and non-1D formations, and applying appropriate inversion algorithms (1D for 1D formations, higher-order models for non-1D) for accurate drilling guidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D, 2.5D, or 3D forward and inversion models are used to evaluate EM measurements in complex formations, then measurement precision and reliability improve, but device complexity and loss of time increase due to extensive data requirements and time-intensive computations
Solution Approach 1:
The system performs preliminary classification of formation geometry (1D vs. non-1D) before executing the full inversion process. By pre-identifying complex formation structures using machine learning algorithms trained on synthetic data, the system can then selectively apply appropriate inversion models only where needed, avoiding unnecessary computational overhead in simple formations.
Solution Approach 2:
The evaluation process is segmented into distinct stages: (1) machine learning-based classification of formation type, (2) selection of appropriate inversion model based on classification, and (3) execution of inversion only for complex formations. This segmentation allows the system to maintain high accuracy for complex cases while achieving real-time performance for simpler cases.
2Reliability
If 2D, 2.5D, or 3D forward and inversion models are used to evaluate EM measurements in complex formations, then measurement precision and reliability improve, but device complexity increases due to extensive data requirements
Solution Approach 1:
A machine learning classifier serves as an intermediary between raw EM measurements and the inversion process. This intermediary analyzes measurement patterns to identify complex formation structures, enabling the system to activate complex inversion models only when necessary. The classifier acts as a gatekeeper that simplifies the overall system by preventing unnecessary execution of computationally intensive algorithms.
Solution Approach 2:
The system dynamically changes operational parameters based on formation classification. When complex formations are detected, the system transitions from using simple 1D inversion to more sophisticated 2D/3D models. This parameter adaptation allows the system to maintain reliability for complex cases while minimizing computational complexity for standard formations.
3Productivity
If 1D inversion model is used to evaluate EM measurements in simple layered formations, then productivity and ease of operation improve due to real-time feedback capability, but measurement precision deteriorates for complex formation structures with faults and non-planar boundaries
Solution Approach 1:
The system dynamically adapts its processing approach based on real-time classification of formation geometry. The machine learning classifier continuously monitors EM measurements and adjusts the inversion model selection accordingly. This dynamic behavior enables the system to maintain real-time productivity for simple formations while automatically switching to more accurate methods when complex structures are detected.
Solution Approach 2:
The integrated system combines multiple functions: machine learning classification, formation geometry identification, and selective inversion model execution. This multi-functional approach allows a single system to handle both simple and complex formations effectively, providing real-time feedback for simple cases and high-precision evaluation for complex cases without requiring separate systems.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables real-time, efficient classification of complex formations, improving drilling accuracy by reducing computational demands and enhancing geosteering capabilities in challenging geological structures.
Implementation Method 1
an electromagnetic logging tool is deployed in a wellbore penetrating the subterranean formation
Data Source
AI summary
A method for classifying a subterranean formation includes deploying an electromagnetic logging tool in a wellbore penetrating the subterranean formation, causing the electromagnetic logging tool to make electromagnetic logging measurements in the wellbore, and evaluating the electromagnetic logging measurements with a trained machine learning primary classifier to classify the subterranean formation as either a 1D formation or a non-1D formation.