Deep Neural Network for Nested Tubular Integrity Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for inspecting downhole tubulars for metal loss due to corrosion in oil and gas wells are inefficient, often leading to inaccurate results and requiring time-consuming inversion processes, which can be sensitive to noise and computation-intensive.
Innovation Solution
A deep neural network (DNN) with convolutional layers is pre-trained to directly process response images from electromagnetic pipe inspection tools, avoiding inversion and providing accurate tubular integrity property images of nested tubulars quickly, while enhancing accuracy with further training using real data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional inversion processes are used to process raw measurements from electromagnetic pipe inspection tools, then detailed tubular integrity information can be obtained, but the process is time-consuming and computation-intensive
Solution Approach 1:
The patent applies preliminary action by pre-training the neural network model offline with synthetic data before actual inspection. The pre-trained model contains learned mappings from raw measurements to tubular properties, enabling rapid inference during field operations without requiring time-consuming inversion processes at the time of inspection
Solution Approach 2:
The patent replaces the traditional mechanical inversion process with a data-driven neural network-based system. Instead of using iterative mathematical inversion algorithms that require significant computational resources and time, the system uses a trained neural network that directly maps measurements to tubular properties, dramatically reducing processing time while maintaining accuracy
2Measurement precision
If traditional inversion processes are used to analyze raw measurements, then tubular properties can be determined, but the results are sensitive to noise and may produce artifacts
Solution Approach 1:
The patent replaces the traditional inversion algorithm with a neural network-based system that is inherently more robust to noise. The neural network learns from training data including noisy examples, enabling it to distinguish between actual tubular defects and noise artifacts, thereby improving reliability of the inspection results
Solution Approach 2:
The patent uses synthetic data that copies realistic tubular inspection scenarios for training the neural network. By creating virtual training datasets that mimic real measurement conditions including noise characteristics, the network learns to generalize well to actual field data while being insensitive to noise and artifacts
3Measurement precision
If point-by-point inversion is applied to preserve accuracy, then measurement accuracy is maintained, but processing speed is significantly reduced
Solution Approach 1:
The patent replaces the point-by-point inversion approach with a neural network that processes measurements in a holistic manner. The neural network architecture accepts measurement data and directly outputs tubular property maps, eliminating the need for sequential point-by-point inversion and achieving both accuracy and high processing speed
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
This approach improves accuracy, reduces sensitivity to noise, and saves time by bypassing inversion, enabling faster and more reliable detection of tubular integrity properties, including thickness and conductivity.
Implementation Method 1
transmitting electromagnetic fields at one or more frequencies with the one or more transmitters; and measuring at least one of a real-part, an imaginary-part, an absolute, an amplitude, and a phase of a received signal at the one or more frequencies with the one or more receivers
Data Source
AI summary
Methods and systems for inspecting the integrity of multiple nested tubulars are included herein. A method for inspecting the integrity of multiple nested tubulars can comprise conveying an electromagnetic pipe inspection tool inside the innermost tubular of the multiple nested tubulars; taking measurements of the multiple nested tubulars with the electromagnetic pipe inspection tool; arranging the measurements into a response image representative of the tool response to the tubular integrity properties of the multiple nested tubulars; and feeding the response image to a pre-trained deep neural network (DNN) to produce a processed image, wherein the DNN comprises at least one convolutional layer, and wherein the processed image comprises a representation of the tubular integrity property of each individual tubular of the multiple nested tubulars.