Machine Learning for 3D Imbibition Saturation Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for predicting three-dimensional imbibition phase saturation profiles in porous rock media are time-consuming and costly, and existing technologies do not effectively utilize machine learning for fluid flow simulations in heterogeneous reservoirs.

Innovation Solution

A computer-implemented method using a machine learning algorithm trained with simulated and measured phase saturation profiles to approximate three-dimensional imbibition phase saturation profiles, incorporating a porous media fluid flow simulator and data from medical-CT core flooding equipment, to predict fluid flow properties in porous rock media.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If medical-CT core flooding equipment is used to capture three-dimensional phase saturation information, then measurement precision is improved, but productivity deteriorates due to time-consuming and expensive extraction processes

Engineering Contradiction:
Improvethree-dimensional phase saturation informationVSAvoidextraction speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent creates a digital copy (virtual model) of the core plug that replicates the three-dimensional phase saturation information obtained from medical-CT scanning. This virtual model allows repeated analysis and extraction of saturation profiles without the need to physically re-scan or re-extract data from the original core sample, significantly improving productivity while maintaining measurement precision.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent performs preliminary action by conducting the time-consuming medical-CT scanning and creating the three-dimensional phase saturation model upfront, before any specific analysis is needed. This pre-computed virtual model can then be used for multiple different fraction flow rates and conditions without repeating the expensive and time-consuming scanning process, thus resolving the productivity issue.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If multiple fraction flow rates are analyzed to understand fluid flow properties, then reliability is improved, but loss of time increases due to repeated extraction processes

Engineering Contradiction:
Improvefluid flow property predictionVSAvoidextraction time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The virtual model serves as a reusable copy that can be analyzed for multiple fraction flow rates without requiring repeated physical extraction or scanning. Researchers can simulate different flow conditions on the same three-dimensional phase saturation data, maintaining reliable fluid flow property predictions while eliminating time loss from repeated extraction processes.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent enables analysis of multiple fraction flow rates by changing the simulation parameters in the virtual model rather than physically re-extracting data for each condition. This allows comprehensive fluid flow property prediction across different operating conditions without the time penalty of repeated measurements.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If three-dimensional phase saturation profiles are extracted for each fraction flow rate, then manufacturing precision is improved, but productivity deteriorates due to repetitive extraction processes

Engineering Contradiction:
Improvephase saturation profile accuracyVSAvoiddata generation speed
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The three-dimensional phase saturation model acts as a master copy from which accurate phase saturation profiles can be extracted for any fraction flow rate on demand. This eliminates the need to re-extract or re-measure data for each condition, maintaining manufacturing precision while dramatically improving productivity by allowing rapid generation of profiles through virtual model analysis.

Inventive Principle:
Principle #26Copying

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

This approach enables efficient and cost-effective generation of three-dimensional phase saturation profiles, improving the accuracy of fluid flow predictions in heterogeneous reservoirs and aiding in hydrocarbon extraction planning.

Implementation Method 1

A computer implemented method using a machine learning algorithm trained with simulated and measured phase saturation profiles to approximate three-dimensional imbibition phase saturation profiles

Methodology Applied
Scientific EffectMachine learning:

Implementation Method 2

incorporating a porous media fluid flow simulator and data from medical-CT core flooding equipment, to predict fluid flow properties in porous rock media

Methodology Applied
Scientific EffectFluid flow through porous media: Permeation

Implementation Method 3

Medical computer tomography (medical-CT) equipment is therefore used for conducting of so-called 'core-flooding experiments' aiming at capturing three-dimensional phase saturation information of the core plugs extracted from the porous rock medium

Methodology Applied
Scientific EffectComputed tomography: Tomography

Implementation Method 4

System and method for correlating oil distribution during drainage and imbibition using machine learning

Methodology Applied
Scientific EffectCapillary action: Capillary Action

Implementation Method 5

predicting a three-dimensional imbibition phase saturation profile for imbibition of the porous rock medium

Methodology Applied
Scientific EffectPhase saturation: Phase Change

Data Source

PatentEP4113117B1System and method for correlating oil distribution during drainage and imbibition using machine learning
Publication Date: 2024.01.03 ADNOC
  • EP4113117B1 patent drawingFigure 1A
  • EP4113117B1 patent drawingFigure 1B
  • EP4113117B1 patent drawingFigure 1C

AI summary

A method (10) and system (100) for approximating a predicted three-dimensional imbibition phase saturation profile (35p) from a measured three-dimensional drainage phase saturation profile (25m), a derived one-dimensional drainage phase saturation profile (20d), a measured one-dimensional imbibition phase saturation profile (30m) using a trained machine-learning algorithm (130) are disclosed. A method (5) for training of the machine learning algorithm (130) is also disclosed.