Automated Well Log Depth Matching via Cross-Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current depth matching techniques for well log data sets from multiple logging passes in the oil and gas industry are often inaccurate and require significant manual intervention, lacking robust and automated solutions for correlating gamma-ray logs and other well logs.
Innovation Solution
A process and system that normalize well logs, calculate correlation coefficients, perform feature picking, and generate a shift table for depth matching using normalized cross-correlation optimization, allowing for automated depth matching and analysis of geological structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If traditional optimization algorithms based on cross-correlation and covariance measures are used for depth matching, then the process can be automated to some extent, but the matching accuracy is insufficient and requires additional manual adjustments
Solution Approach 1:
The patent divides the well log data into multiple segments or windows along the depth profile. Each segment is processed independently through cross-correlation to identify local shifts, which are then integrated to produce the overall depth matching transformation. This segmentation allows the algorithm to capture local variations in log response while maintaining global consistency, thereby improving accuracy without sacrificing automation.
Solution Approach 2:
The patent employs a dynamic programming approach where the depth matching transformation is built incrementally by making optimal local decisions at each depth level. The algorithm maintains a cumulative cost function that allows backtracking and refinement, enabling it to adapt to varying log characteristics throughout the well profile. This dynamic approach improves matching accuracy while preserving full automation.
2Extent of automation
If machine learning-based solutions with fully connected neural networks are used for depth matching, then automation can be achieved, but significant resources and human intervention are required for training the model before actual depth matching can be utilized
Solution Approach 1:
The patent implements a self-service mechanism where the depth matching algorithm automatically learns from the data itself without requiring external training. The cross-correlation-based method inherently adapts to the specific characteristics of each well log pair by computing optimal shifts based on actual signal correlations. This eliminates the need for separate training infrastructure, manual model preparation, and extensive human intervention, while maintaining full automation in the depth matching process.
3Measurement precision
If manual adjustments are made to improve depth matching accuracy, then matching precision can be enhanced, but the productivity and efficiency of the process decrease due to time-consuming manual intervention
Solution Approach 1:
The patent replaces the manual mechanical adjustment process with an automated computational system based on cross-correlation analysis. The algorithm automatically identifies optimal depth shifts by comparing log responses mathematically, substituting human visual inspection and manual realignment with objective, repeatable computational procedures. This substitution maintains high accuracy while dramatically improving processing efficiency and eliminating the trade-off between precision and productivity.
4Productivity
If existing depth matching algorithms are used, then automated processing can proceed, but user intervention is still required which contradicts the demand for fully automated data processing and interpretation
Solution Approach 1:
The patent incorporates a feedback mechanism where the cross-correlation results are continuously evaluated and used to refine the depth matching transformation. The algorithm computes correlation coefficients, identifies peak positions, and adjusts shifts iteratively based on the quality of matching achieved. This internal feedback loop allows the system to self-correct and optimize without external user input, achieving the demanded level of full automation while maintaining high processing productivity.
Data Source
AI summary
Processes and systems for correlating well log data sets from well logging passes within a well bore. In some embodiments, a process for well log depth matching can include normalizing a first well log from a first logging pass obtained within a well bore and a second well log from a second logging pass obtained within the well bore, performing a pre-shift, performing feature picking to identify one or more features along the second well log, performing normalized cross-correlation based optimization between the first well log and the second well log to match the one or more features along the second well log to the same one or more features of the first well log and generating a shift table for depth shifting the one or more features of the second well log and the first well log.