Real-time quantitative characterization method, equipment and medium for rock mass evolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to provide real-time, high-precision imaging and transparent characterization of multiphase field coupling rock mass evolution during carbon dioxide geological storage, leading to a 'black box' understanding of fracture formation and evolution under multiphase field coupling environments.
Innovation Solution
A real-time quantitative characterization method using a deep learning model driven by physical principles, combining data fusion of multiphase field detection and monitoring data with feature extraction and Transformer models, to generate evolution distribution images and quantify physical and mechanical parameters through a mathematical model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional imaging methods (X-ray tomography, nuclear magnetic resonance, geophysical detection, acoustic emission monitoring) are used to characterize rock mass evolution, then structural information can be provided, but imaging lag and poor imaging accuracy occur, failing to achieve real-time transparent characterization
Solution Approach 1:
The patent transforms the characterization approach by changing from direct imaging parameters to indirect physical field parameters (temperature, strain, resistivity, chemical ion concentration, microbial cell concentration, P wave velocity, S wave velocity, acoustic emission). These parameter changes enable real-time monitoring without imaging lag while maintaining characterization accuracy through deep learning reconstruction.
Solution Approach 2:
The patent replaces traditional mechanical/imaging-based detection systems with a data-driven deep learning system. Instead of using physical imaging mechanisms that cause lag, the system uses neural networks to reconstruct evolution distribution images from real-time physical field data, eliminating imaging lag while improving accuracy.
2Loss of information
If multiple detection technologies are combined to improve characterization comprehensiveness, then more rock mass information can be obtained, but system complexity increases and real-time processing becomes difficult
Solution Approach 1:
The patent merges multiple detection technologies (temperature sensors, strain sensors, resistivity sensors, chemical sensors, microbial sensors, ultrasonic sensors, acoustic emission sensors) into a unified deep learning framework. The Transformer model integrates these diverse data sources, processing them simultaneously to achieve comprehensive characterization without proportionally increasing system complexity.
Solution Approach 2:
The deep learning model serves multiple functions: it processes diverse sensor data, reconstructs evolution distribution images, characterizes rock mass evolution, and provides real-time predictions. This multi-functionality eliminates the need for separate processing systems for each detection technology, reducing overall system complexity while maintaining comprehensiveness.
3Productivity
If deep learning models are used for real-time data processing, then processing speed improves, but model training complexity and computational requirements increase
Solution Approach 1:
The patent performs model training in advance during the preliminary stage, creating a pre-trained deep learning model. Once trained, the model can perform real-time inference rapidly without requiring continuous retraining. This preliminary action separates the computationally intensive training phase from the real-time processing phase, achieving both high processing speed and manageable training complexity.
Data Source
AI summary
A real-time quantitative characterization method, equipment and a medium for rock mass evolution are disclosed. The method includes: the multiphase field detection and monitoring data is fused; a deep learning model driven by physical principles is established based on the physical principles of fluid density; the deep learning model is trained using real-time fused data as input and corresponding evolution distribution images as output; a trained deep learning model is used to obtain an evolution distribution image based on multi-phase field detection and monitoring data of different time periods and types; a mathematical model is used to quantitatively characterize of physical and mechanical parameters in the whole process of progressive failure of the dynamic evolution of the rock mass based on the macroscopic mechanical parameters and evolution distribution images synchronized with multiphase field detection and monitoring data.