Deep Learning Spectroscopic Prediction of Subsurface Properties
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for rock characterization and classification in geo-exploration are limited in accuracy and efficiency, particularly in predicting geological formation properties such as rock type, geomechanical properties, sonic velocity, and permeability, which are crucial for effective energy resource exploration and mineral deposit development.
Innovation Solution
A computer-implemented method using deep learning models, specifically convolutional neural networks, trained with spectroscopic infrared (IR) data from core samples to predict geological formation properties, enabling more accurate and high-resolution predictions compared to traditional methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used for rock characterization and classification, then the process is simpler, but the accuracy in predicting geological formation properties deteriorates
Solution Approach 1:
The patent replaces traditional mechanical and manual rock characterization methods with a computational system using spectroscopic data and machine learning algorithms. The system substitutes physical analysis methods with optical spectroscopy (FTIR, Raman) combined with automated deep learning models, achieving higher accuracy in predicting geological properties without manual intervention.
Solution Approach 2:
The patent transforms rock characterization by changing the measurement parameters from traditional physical methods to spectroscopic parameters (infrared absorption, Raman scattering). By measuring molecular vibrations and chemical bonds through spectroscopy, the system extracts detailed compositional information that enables more accurate prediction of geological formation properties.
2Measurement precision
If spectroscopic data and deep learning models are integrated, then the prediction accuracy improves, but the computational complexity increases
Solution Approach 1:
The patent segments the complex prediction task into multiple specialized deep learning models, each trained for specific geological properties (rock type classification, porosity prediction, permeability estimation, strength characterization). This segmentation allows each model to focus on specific patterns in the spectroscopic data, improving overall accuracy while making the system more manageable through modular architecture.
Solution Approach 2:
The system performs preliminary processing and feature extraction from spectroscopic data before feeding it to the deep learning models. Preprocessing steps including noise filtering, spectral normalization, and key feature identification are executed in advance, reducing the computational burden on the prediction models and improving their efficiency.
3Measurement precision
If high-resolution spectroscopic analysis is performed, then the characterization precision improves, but the time required for analysis increases
Solution Approach 1:
The patent replaces time-consuming manual rock analysis with automated spectroscopic measurement and machine learning prediction. The system captures spectral data rapidly using FTIR or Raman spectroscopy instruments, then immediately processes the data through pre-trained deep learning models to generate predictions, reducing analysis time from days to minutes while maintaining high precision.
Solution Approach 2:
The system creates a spectral fingerprint copy of the rock sample that contains all necessary compositional information. Instead of physically analyzing the entire rock sample through multiple tests, the spectroscopic copy captures molecular-level characteristics that can be rapidly analyzed computationally, preserving diagnostic information while enabling fast processing.
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 method provides enhanced accuracy in predicting rock type, geomechanical properties, sonic velocity, and permeability, offering improved guidance for energy resource exploration and mineral deposit development by integrating IR data with well logs and geophysical data, thereby improving the precision of subsurface property estimation.
Implementation Method 1
spectroscopic infra-red (IR) data, wherein at least portions of the plurality of geo-exploration data are based on measurements of core samples
Data Source
AI summary
A computer-implemented method includes: accessing a plurality of geo-exploration data from a first drilling site, wherein the plurality of geo-exploration data include spectroscopic infra-red (IR) data and well logs, wherein at least portions of the plurality of geo-exploration data are based on measurements of core samples taken from the first drilling site; based on, at least in part, the plurality of geo-exploration data, training a set of deep learning models, each deep learning model comprising multiple layers and configured to predict one or more geological formation properties; applying the set of deep learning models to newly received geo-exploration data that also includes spectroscopic IR data; and predicting the one or more geological formation properties based on, at least in part, the newly received geo-exploration data.