Automated Capillary Pressure Curve Classification for Reservoir Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual classification of capillary pressure and relative permeability curves becomes cumbersome and subjective when dealing with a large number of experimental measurements, leading to inefficient and unreliable petrophysical property assignments in reservoir models.
Innovation Solution
An automated method involving principal component analysis, curve representation, classification, and assignment of representative curves to each mesh, using techniques like K-Means or neural networks to categorize curves based on morphology and associate them with lithologies, enabling efficient and quantitative petrophysical behavior characterization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual classification of capillary pressure and relative permeability curves is performed, then the process allows for visual inspection and expert judgment, but the process becomes cumbersome and subjective when dealing with a large number of experimental measurements
Solution Approach 1:
The patent replaces the manual visual inspection mechanism with an automated image processing and pattern recognition system. Curves are transformed into graphical representations and processed through computational algorithms that objectively identify curve characteristics and classify them into families, eliminating the cumbersome and subjective nature of manual classification while maintaining or improving accuracy
Solution Approach 2:
The patent transforms curve data into graphical representations with specific visual parameters (amplitude, shape, inflection points) that can be systematically measured and compared. By changing the representation from raw data tables to standardized graphs with defined characteristics, the system enables automated classification based on quantifiable parameters rather than subjective visual assessment
2Adaptability or versatility
If manual classification of curves is performed, then expert judgment can be applied, but the results become subjective and vary between different analysts
Solution Approach 1:
The patent replaces the human expert judgment mechanism with an automated computational system that applies consistent classification rules to all curves. The system uses objective criteria based on curve morphology and graphical characteristics, ensuring that the same curve will always be classified in the same way regardless of which analyst reviews it, thereby eliminating subjectivity and improving reliability
Solution Approach 2:
The patent implements a feedback mechanism where the automated classification system can be validated and refined by comparing its results with expert classifications. This allows the system to learn from and incorporate expert judgment while maintaining objectivity and consistency, bridging the gap between flexible expert analysis and reliable automated processing
3Quantity of substance
If numerous experimental curves are available for classification, then more data is available for accurate petrophysical property assignment, but the manual classification process becomes increasingly difficult and time-consuming
Solution Approach 1:
The patent replaces the time-consuming manual classification process with an automated computational system that can rapidly process large numbers of curves. The system transforms curves into graphical representations and uses algorithmic methods to classify them based on morphological characteristics, reducing the time required to classify numerous experimental curves from days or weeks to minutes or hours
Solution Approach 2:
The patent segments the classification process into distinct automated steps: curve transformation into graphical representations, extraction of morphological parameters, comparison against reference curves, and assignment to classification families. This segmentation allows each step to be optimized and executed automatically, enabling efficient processing of large datasets without the time penalty that would result from manually processing each curve individually
Data Source
Figure 1~2
Figure 3
Figure 4a~4b
AI summary
The method involves analyzing a component of a data table constructed from curves describing evolution of saturation of fluid to extract factors explaining the data. The curves are represented in a space constructed from the factors, and the curves are classified into classes. A curve representative of each class is defined, and one of the representative curves is associated with each cell. A flow simulator is used to define a development scheme for a reservoir from a representation of the reservoir, and the reservoir is developed according to the development scheme.