Well Log Correlation Using Multi-Channel ML Depth Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for wellbore log correlation are limited in the number of signals and well logs that can be aligned simultaneously, inefficient for multi-channel data processing, and unable to handle null or missing data, leading to reduced computing efficiency and limited correlation capabilities.
Innovation Solution
A machine-learning model is trained to perform assisted well correlation by aligning sets of well signals, including multiple signals and channels, using input arrays with defined control points and transforming null data to facilitate efficient alignment across multiple wellbores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional methods are used for wellbore log correlation, then the process is simple to implement, but the number of signals and well logs that can be aligned simultaneously is limited
Solution Approach 1:
The patent segments the well log correlation problem into multiple independent channels, where each channel processes a specific well log or signal. This segmentation allows the system to handle multiple signals and well logs simultaneously by processing them in parallel through separate channels, thereby increasing the number of alignable signals without proportionally increasing overall system complexity.
Solution Approach 2:
The patent implements a universal correlation framework that can process multiple types of well logs and signals through a single multi-channel system. The system is designed to handle diverse data types (gamma ray, resistivity, acoustic logs, etc.) using the same underlying correlation algorithms, enabling versatile alignment of numerous signals simultaneously while maintaining manageable complexity through standardized processing procedures.
2Productivity
If conventional methods are used for wellbore log correlation, then the computational requirements are low, but the processing efficiency for multi-channel data is poor
Solution Approach 1:
The patent applies preliminary actions by pre-processing well log data into standardized multi-channel formats before correlation analysis. Control points are pre-identified and data is organized into structured arrays with defined channels, which prepares the data for efficient parallel processing. This preliminary organization enables the system to handle large volumes of multi-channel data more efficiently during the actual correlation computation.
3Reliability
If conventional methods are used for wellbore log correlation, then the system is easy to operate, but the ability to handle null or missing data is poor
Solution Approach 1:
The patent introduces control points as intermediary elements that facilitate the correlation process, especially in the presence of null or missing data. These control points serve as reference markers that anchor the alignment process, allowing the system to reliably correlate well logs even when data gaps exist. The control points act as mediators that maintain correlation reliability without significantly complicating system operation.
Data Source
AI summary
A method for performing wellbore correlation across multiple wellbores includes predicting a depth alignment across the wellbores based on a geological feature of the wellbores. Predicting a depth alignment includes selecting a reference wellbore, defining a control point in a reference signal of a reference well log for the reference wellbore, and generating an input tile from the reference signal, the control points, and a number of non-reference well logs corresponding to non-reference wellbores. The well logs include changes in a geological feature over a depth of a wellbore. The input tile is input into a machine-learning model to output a corresponding control point for each non-reference well log. The corresponding control point corresponds to the control point of the reference log. Based on the corresponding control points output from the machine-learning model, the non-reference well logs are aligned with the reference well log to correlate the multiple wellbores.