Machine Learning for Downhole Formation Textural Parameter Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for evaluating formation properties in oil and gas operations rely heavily on expensive and time-consuming core samples, which are ineffective for deep zones and often inaccurate due to near-borehole zone measurements being influenced by flushed fluids, limiting the ability to determine true formation properties like water saturation and textural parameters.
Innovation Solution
A system and method utilizing dielectric measurements processed by a trained machine learning system to determine downhole formation properties, including water saturation and textural parameters, by generating synthetic rock geometries and simulating dielectric responses, allowing for the classification of formations based on textural properties without the need for core samples.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If core samples are used to evaluate formation properties, then deep zone and rock texture properties can be determined, but the process becomes expensive and time consuming
Solution Approach 1:
The patent creates synthetic core samples through numerical modeling that replicates the physical and chemical properties of actual rock formations. These virtual core samples can be generated and evaluated immediately without the time-consuming processes of physical core extraction, transportation, and laboratory analysis, while maintaining measurement precision for formation properties determination
Solution Approach 2:
The patent replaces the mechanical system of physical core sampling and laboratory analysis with a computational system using numerical models and simulations. This substitution eliminates the need for physical core handling and enables rapid evaluation of formation properties through computer-based analysis of synthetic rock samples
2Measurement precision
If core samples are extracted and evaluated, then formation properties can be determined, but the wellbore remains unproductive during evaluation, increasing costs
Solution Approach 1:
The patent creates virtual replicas of core samples through numerical modeling, allowing formation properties to be determined on these synthetic samples instead of requiring the actual wellbore to be shut in for physical core analysis. This enables continuous wellbore productivity while obtaining formation data
Solution Approach 2:
The patent performs formation evaluation calculations in advance using numerical models before the wellbore needs to be taken offline. By pre-computing formation properties from synthetic core samples, the system eliminates the need to shut in the well during evaluation, maintaining productivity throughout the process
3Ease of operation
If shallow downhole measurements are used, then near-borehole zone properties can be evaluated, but deep zone properties and true formation characteristics cannot be determined accurately
Solution Approach 1:
The patent introduces numerical models and synthetic core samples as intermediaries that bridge the gap between shallow downhole measurements and deep zone formation properties. The numerical models process the shallow measurement data through simulated rock physics to infer deep zone characteristics and true formation properties that cannot be directly measured
Solution Approach 2:
The patent replaces the limitation of shallow physical measurements with a computational approach using numerical models. These models transform limited shallow measurement data into comprehensive deep zone formation property evaluations through virtual rock physics simulations, eliminating the need for direct physical measurement of deep zones
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
Enables the determination of formation properties directly from downhole measurements, reducing the need for costly core sampling and providing accurate textural and water saturation data, thereby improving the efficiency and accuracy of wellbore characterization and hydrocarbon recoverability assessments.
Implementation Method 1
obtaining a dielectric measurement of a downhole formation
Data Source
AI summary
A method includes obtaining a dielectric measurement of a portion of a downhole formation. The method also includes processing the dielectric measurement via a trained machine learning system. The method further includes determining at least one of a classification or a textural parameter simultaneously with formation water saturation of the downhole formation, via the machine learning system, based at least in part on the dielectric measurement. The machine learning system is based on a pre-determined dataset from previous measurements or simulated results of synthetic cases. The method determining the correlation between water saturation and formation texture through a frequency cascading training process based on sensitivity of complex dielectric spectrum with respect to desired parameters including water saturation and texture parameter. The method also includes assigning at least one of the classification or the textural parameter to the downhole formation.