Porosity Estimation in Organic Shale Formations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is no reliable method to compute porosity associated with organic matter (PAOM) from commonly available well log data or core CT scan data, particularly in shale formations, as existing methods are time-consuming, costly, and require advanced technology logs that are often unavailable or unreliable.
Innovation Solution
A method that uses scanning electron microscope (SEM) or focused ion beam (FIB)-SEM images to determine PAOM, total porosity, and organic matter content, followed by curve-fitting to generate mathematical functions for computing PAOM values from well logging or core CT scan data, reducing the need for individual SEM image analysis and enabling PAOM estimation along the length of a wellbore or formation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ion milled SEM image analysis is used to evaluate PAOM, then measurement precision is improved, but productivity deteriorates due to time-consuming and costly individual sample analysis
Solution Approach 1:
The patent creates a digital replica of the SEM image analysis process by training a machine learning model on a dataset of ion-milled SEM images with manually measured PAOM values. The trained model then generates predictions for new samples without requiring actual SEM imaging, thus copying the analytical capability while eliminating the time-consuming imaging and manual measurement steps.
Solution Approach 2:
The patent performs preliminary action by creating a comprehensive training dataset in advance with known PAOM values from ion-milled SEM analysis. This pre-processing work enables the machine learning model to learn the relationship between well log data and PAOM, so that future predictions can be made rapidly without repeating the complex SEM analysis process.
2Reliability
If advanced technology well logs (NMR or dielectric properties) are used to compute PAOM, then reliability is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces the physical measurement systems (NMR or dielectric property logs) with a computational approach using machine learning. Instead of requiring complex hardware to directly measure PAOM-related properties, the system uses a trained model that processes data from simpler, more widely available well logs to predict PAOM, thereby substituting mechanical/physical measurement systems with an information-processing system.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between readily available well log data and PAOM estimation. This intermediary translates data from conventional well logs into reliable PAOM predictions without requiring direct measurements from complex advanced logging tools, thus bridging the gap between simple input data and reliable output predictions.
3Ease of operation
If traditional well log analysis methods are used for shale formations, then ease of operation is maintained, but measurement precision deteriorates because traditional methods do not work well for organic shales
Solution Approach 1:
The patent changes the parameters used in well log analysis by training a machine learning model to recognize patterns in conventional well log data that are specific to organic shale formations. Instead of applying traditional parameters and formulas designed for sandstones and carbonates, the model learns formation-specific relationships from training data, thereby adapting the analysis parameters to work effectively for shale while maintaining ease of operation.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method for determining porosity associated with organic matter (PAOM) in a well or formation is provided. A system for performing the method is also provided.