Machine Learning Wellbore Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for evaluating wellbore defects in multi-barrier wells, such as plug and abandonment wells, face challenges due to weak nuclear measurement signals, leading to the need for expensive and complex evaluation methods.
Innovation Solution
The use of machine learning systems to process nuclear measurement data, including feature engineering, normalization, and classification techniques, to identify wellbore defects like cement defects and tubing eccentricity, leveraging multiple weak signals to improve defect detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If nuclear measurement modalities are used in multi-barrier wells, then defect detection capability is improved, but signal strength becomes insufficient
Solution Approach 1:
The patent segments the defect detection process into multiple classification stages (first classifier for feature of interest, second classifier for defect presence, third classifier for defect properties). This segmentation allows the system to process weak nuclear signals systematically, extracting useful information at each stage rather than requiring strong single-pass signals.
Solution Approach 2:
The patent transitions from direct signal interpretation to a multi-dimensional feature space by extracting features of interest through the first classifier, then processing these features through additional classifiers. This dimensional transformation enables the system to detect defects even when original nuclear signals are weak.
2Measurement precision
If traditional evaluation methods are deployed for multi-barrier wells, then defect detection accuracy is maintained, but system complexity and cost increase
Solution Approach 1:
The patent replaces complex physical evaluation methods with a machine learning-based classification system. Instead of using complicated hardware or invasive measurement techniques, the system uses trained classifiers to process nuclear measurement data, achieving accurate defect detection with simpler equipment.
Solution Approach 2:
The machine learning system is trained on log data to automatically identify features of interest and detect defects without requiring complex manual evaluation procedures. The system serves itself by learning from historical data and applying that knowledge to new measurements, reducing the need for expert intervention and complex procedural methods.
Data Source
AI summary
A method for identifying defects in a multi-barrier wellbore includes receiving log data, the log data corresponding to one or more wellbore operations, the log data including data from at least one measurement modality corresponding to a present measurement modality. The method also includes training, using the log data, a machine learning model. The method further includes acquiring wellbore data, via the present measurement modality, during a logging operation. The method also includes processing at least a portion of the wellbore data using the trained machine learning model. The method includes identifying one or more features of interest in the wellbore data, via the trained machine learning model.