EM Pipe Property Inversion Using Multi-Channel Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for assessing casing integrity and thickness in wellbore operations rely heavily on subjective human interpretation of electromagnetic (EM) tool measurements, leading to inconsistent and time-consuming results due to varying inversion hyperparameters.
Innovation Solution
Employing machine learning algorithms to determine appropriate inversion hyperparameters and potentially replace the inversion calculation, providing consistent and reliable assessments of casing integrity and thickness using EM data collected by logging tools.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If human guidance is used to determine inversion hyperparameters, then the assessment can incorporate prior knowledge and experience, but the results become subjective and inconsistent
Solution Approach 1:
The system enables self-service by allowing the machine learning model to automatically determine inversion hyperparameters without human intervention. The model learns from training data and autonomously selects appropriate hyperparameter values, eliminating subjective human guidance while maintaining reliable and consistent assessments.
Solution Approach 2:
The invention changes the parameter determination approach from manual human selection to automated machine learning-based selection. The system transforms hyperparameter determination into a learnable task where the model adjusts parameters based on patterns in training data, achieving consistency through algorithmic parameter optimization rather than human judgment.
2Productivity
If human guidance is used to determine inversionhyperparameters, then prior knowledge can be applied, but the process becomes time consuming
Solution Approach 1:
The system applies preliminary action by pre-training the machine learning model on extensive training data that encapsulates domain knowledge and experience. This preliminary training phase stores learned patterns and relationships, allowing the model to rapidly determine appropriate hyperparameters during actual assessments without requiring time-consuming human analysis each time.
Solution Approach 2:
The invention substitutes the mechanical human decision-making process with an automated machine learning system. Instead of humans manually analyzing cases and selecting hyperparameters (a time-consuming mechanical process), the system uses computational algorithms to automatically determine parameters, dramatically increasing productivity while reducing time loss.
3Measurement precision
If traditional inversion methods are used, then detailed analysis can be performed, but computational requirements increase
Solution Approach 1:
The machine learning model serves as an intermediary between raw EM measurements and final casing thickness assessments. Instead of directly performing computationally intensive traditional inversion, the model processes measurements through learned relationships, maintaining measurement precision while significantly reducing computational energy requirements by acting as an efficient intermediary layer.
Solution Approach 2:
The system creates a computational copy of the inversion process through the machine learning model. The model learns to replicate the functionality of traditional inversion methods by training on labeled data, producing copies of accurate thickness assessments without requiring the full computational burden of traditional inversion algorithms for each new measurement set.
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
The use of machine learning models reduces computational requirements and ensures reproducible, expedient, and consistent analysis of casing integrity and thickness, overcoming the limitations of human-dependent methods.
Implementation Method 1
One type of corrosion monitoring tool uses electromagnetic (EM) fields to estimate pipe thickness or other corrosion indicators
Data Source
AI summary
A method for estimating a pipe property may include disposing an electromagnetic (EM) logging tool in a wellbore. The EM logging tool may comprise a transmitter disposed on the EM logging tool and a receiver disposed on the EM logging tool. The method may further include transmitting an electromagnetic field from the transmitter into a tubular string to energize the tubular string with the electromagnetic field thereby producing an eddy current that emanates from the tubular string. The method may further include measuring the eddy current in the tubular string with the receiver on at least one channel to obtain one or more measurements from the EM logging tool. The method may further include forming an EM log from the one or more measurements, forming a relationship between a set of one or more previously analyzed measurements and at least one pipe property, and estimating the at least one pipe property from the formed relationship and the one or more measurements from the EM logging tool.