Well Log Correlation Using Global and Local Marker Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional well-top correlation in wellbore interpretation is time-consuming and error-prone due to reliance on manual visual examination of logs, especially for formations with subtle visual signatures.
Innovation Solution
A method utilizing a combination of global and local machine learning models to predict markers in well logs, merging predictions to align and correlate well logs accurately, incorporating a soft attention-based Convolutional Neural Network (CNN) with U-Net architecture for enhanced accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual visual examination of logs is used for well-top correlation, then human expertise and judgment can be applied, but the process is time-consuming and error-prone
Solution Approach 1:
The patent replaces the manual mechanical process of visual log examination with an automated computer-based system that uses machine learning models (specifically U-Net and attention mechanisms) to automatically identify and correlate well tops, eliminating human manual intervention while maintaining or improving accuracy
Solution Approach 2:
The system enables the well log correlation process to perform itself automatically through trained machine learning models that can independently identify formation tops and perform correlations without requiring human operators to manually examine logs, making the system self-sufficient in the correlation task
2Adaptability or versatility
If manual visual examination is used for well-top correlation, then flexibility in handling diverse log types is maintained, but the process becomes cumbersome for tops with subtle visual signatures
Solution Approach 1:
The patent transforms the correlation task from visual pattern recognition to a mathematical parameter-based approach, where machine learning models analyze numerical parameters and patterns in log data to precisely identify formation tops, enabling accurate detection even when visual signatures are subtle
Solution Approach 2:
The machine learning system is designed to handle multiple types of well logs (gamma ray, resistivity, density, sonic, etc.) and various formation types through a unified automated framework, making the system universally applicable across different log types and geological conditions
3Reliability
If traditional manual correlation methods are used, then human geologists can apply geological knowledge, but the overall productivity of well log correlation is low
Solution Approach 1:
The patent replaces manual geological interpretation with automated machine learning-based interpretation systems that can process multiple well logs simultaneously and rapidly, dramatically increasing throughput while maintaining geological accuracy through trained models
Solution Approach 2:
The system performs preliminary automated correlation and marker identification before any human review, pre-processing the logs and identifying potential formation tops so that geologists only need to verify results rather than perform complete manual correlation, thereby increasing overall productivity
Data Source
AI summary
A method for correlating well logs includes receiving a well log as input to a first machine learning model that is configured to predict first markers in the well log based at least in part on a global factor of the well log, receiving the well log as input to a second machine learning model that is configured to predict second markers in the well log based at least in part on local factors of the well log, generating a set of predicted well markers by merging at least some of the first markers and at least some of the second markers, and aligning the well log with respect to one or more other well logs based at least in part on the set of predicted well markers.