3D Vision Morphology Decoder for Heterogeneous Rock Permeability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting rock permeability, particularly in heterogeneous materials like Cretaceous carbonate, face challenges due to high computation requirements and insufficient image resolution, limiting the effectiveness of convolutional neural networks (CNN) in analyzing 3D rock morphology.
Innovation Solution
A novel approach using Magnetic Resonance Imaging (MRI) and Micro-Computed Tomography (μCT) data interpretation, combined with machine learning and computer vision, to autonomously segment and label rock sections, and apply geometrical analysis and permeability equations to predict 3D permeability, incorporating a morphology decoder process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If direct numerical simulation including Lattice Boltzmann method is used to model fluid flow through porous material, then accuracy of permeability prediction is improved, but computation power requirement increases significantly
Solution Approach 1:
The patent replaces complex numerical simulation methods (Lattice Boltzmann method) with a machine learning-based computational model. The ML model learns permeability predictions from training data without requiring intensive numerical simulations, thereby reducing computation power requirements while maintaining prediction accuracy.
Solution Approach 2:
The patent performs preliminary actions by training the machine learning model on a dataset of rock images with known permeability values before actual prediction. This pre-training phase allows the model to capture complex relationships in advance, enabling fast and accurate predictions without requiring intensive computation during the actual permeability estimation process.
2Extent of automation
If convolutional neural network is used for rock morphology analysis, then automation of permeability prediction is improved, but measurement precision deteriorates due to insufficient image resolution for capturing heterogeneity in 3D
Solution Approach 1:
The patent transitions from 2D image analysis to 3D volumetric analysis by processing a stack of 2D rock slice images to reconstruct and analyze the 3D pore structure. This dimensional enhancement allows the CNN to capture heterogeneity and spatial relationships in three dimensions, improving measurement precision while maintaining automation.
Solution Approach 2:
The patent implements a nested architecture where a CNN model is embedded within a larger processing pipeline that handles 3D image reconstruction, segmentation, and analysis. The CNN operates on individual slices or small volumes, while the outer system integrates results across multiple slices to produce accurate 3D permeability predictions, combining automation with precision.
3Ease of operation
If 2D image analysis is used for permeability determination, then ease of operation is improved, but measurement precision deteriorates due to inability to capture 3D heterogeneity
Solution Approach 1:
The patent enhances the analysis from 2D to 3D by processing stacks of rock slice images. The system reconstructs the 3D pore network structure from multiple 2D slices, allowing accurate capture of heterogeneity and spatial relationships while maintaining computational feasibility through automated processing.
Solution Approach 2:
The patent creates a digital 3D copy of the rock pore structure from 2D image slices. This virtual 3D model preserves the complex heterogeneity of the rock structure, enabling accurate permeability prediction without requiring physical 3D scanning or complex experimental measurements.
Data Source
AI summary
Methods and systems are disclosed to determine a permeability value and a permeability image of a rock. The process can include determining labeling of a 3D segment of a rock. The labeled segments can be determined using, for example, computer vision and machine learning and can be used to determine a grain size of a 3D rock segments. The grain size value can be used to determine a permeability value for the 3D rock segment. The permeability value of the 3D rock segment can be used to determine the heterogeneous permeability of the rock.