Rock Crack Identification via Variational Autoencoder 3D Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for monitoring rock crack information during hydraulic fracturing, such as acoustic emission and CT scanning, are either inaccurate or costly, making it difficult to accurately describe the spatial position, quantity, trend, length, and width of cracks in rock specimens.
Innovation Solution
A method and system using a variational autoencoder that combines micro-resistivity scanning and CT scanning to train a variational autoencoder with two-dimensional and three-dimensional images, allowing for the reconstruction of three-dimensional CT scanning images from two-dimensional micro-resistivity distribution images to determine crack information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If acoustic emission technology is used to monitor rock cracks, then monitoring can be performed, but the technology cannot accurately locate cracks and describe rock crack information such as spatial position, quantity, trend, length, and width
Solution Approach 1:
The patent combines micro-resistivity scanning imaging technology with variational autoencoder deep learning technology to merge the advantages of both approaches. The micro-resistivity scanning provides spatial resolution while the variational autoencoder reconstructs three-dimensional crack information, achieving both accurate location and comprehensive crack description capabilities.
Solution Approach 2:
The variational autoencoder acts as an intermediary that transforms two-dimensional micro-resistivity distribution images into three-dimensional CT scanning images. This intermediary process enables accurate crack localization and characterization without requiring direct three-dimensional scanning of every specimen.
2Measurement precision
If CT scanning technology is used to obtain rock crack information, then accurate three-dimensional crack information can be obtained, but the test cost is high
Solution Approach 1:
The patent uses micro-resistivity scanning to create a two-dimensional copy of the rock specimen's internal structure, then employs variational autoencoder to reconstruct a three-dimensional representation from this two-dimensional data. This copying approach avoids the high cost of direct three-dimensional CT scanning while maintaining accurate crack information extraction.
Solution Approach 2:
The patent changes the measurement parameters from three-dimensional CT scanning to two-dimensional micro-resistivity scanning, then uses variational autoencoder to reconstruct the three-dimensional information. This parameter change reduces test cost while maintaining the ability to obtain accurate three-dimensional crack information.
3Loss of energy
If micro-resistivity scanning imaging is used to obtain two-dimensional images, then low cost is achieved, but three-dimensional crack information cannot be directly obtained
Solution Approach 1:
The patent uses variational autoencoder to perform dimensionality change, transforming two-dimensional micro-resistivity distribution images into three-dimensional CT scanning images. This dimensional transformation recovers the lost three-dimensional crack information while maintaining the cost advantage of two-dimensional scanning.
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
Accurately determines rock crack information with low costs by leveraging the variational autoencoder to reconstruct three-dimensional images from two-dimensional data, improving crack identification precision and reducing operational expenses.
Implementation Method 1
performing a micro-resistivity scanning imaging test on the rock specimen by using a micro-resistivity scanning imaging test system, to obtain a two-dimensional micro-resistivity distribution image of the rock specimen
Implementation Method 2
performing a computed tomography (CT) scanning test on the rock specimen, to obtain a three-dimensional CT scanning image of the rock specimen
Data Source
AI summary
The present disclosure pertains to a rock crack information identification method and system based on a variational autoencoder, and belongs to the technical field of rock mechanical fracturing tests in petroleum engineering. A micro-resistivity scanning imaging test and a computed tomography (CT) scanning test are performed on each rock specimen to obtain a two-dimensional micro-resistivity distribution image and a three-dimensional CT scanning image of the rock specimen for training, and an initial variational autoencoder is trained to obtain a trained variational autoencoder. In an actual application, a two-dimensional micro-resistivity distribution image of a rock to be tested is obtained first, and the three-dimensional CT scanning image is reconstructed by using the trained variational autoencoder, such that rock crack information of the rock to be tested is accurately determined with low costs.