Machine Learning Relative Permeability Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining relative permeability curves in oilfield development are time-consuming, costly, and lack accuracy, particularly in high water-cut stages where fluid distribution in reservoirs becomes complex, limiting the ability to model seepage characteristics and optimize oil recovery.
Innovation Solution
A machine learning-based method and system that acquire and process logging curve data to train models for predicting relative permeability curves, reducing costs and improving accuracy by using AI algorithms to establish comprehensive predictions without requiring extensive mathematical modeling or nonlinear control simulations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laboratory core test method is used to determine relative permeability curve, then measurement accuracy is improved, but time consumption and cost increase significantly
Solution Approach 1:
The patent replaces the mechanical laboratory core test system with a machine learning-based prediction system. The system uses logging data and reservoir parameters as inputs to train models that predict relative permeability curves, eliminating the need for physical core testing while maintaining prediction accuracy within 5% of laboratory results.
Solution Approach 2:
The patent creates virtual copies of laboratory test results through machine learning models. By training on limited core test data and replicating the seepage behavior patterns, the system generates predicted relative permeability curves that mirror laboratory measurements without requiring actual physical testing.
2Measurement precision
If laboratory core test method is used to determine relative permeability curve, then measurement accuracy is improved, but cost increases significantly
Solution Approach 1:
The patent replaces expensive laboratory core testing infrastructure with computational models running on standard computing hardware. The machine learning system processes logging data and reservoir parameters to generate relative permeability predictions, eliminating costs associated with core acquisition, laboratory equipment, and specialized testing facilities.
Solution Approach 2:
The patent uses readily available logging data and reservoir parameters as inputs instead of expensive, difficult-to-obtain core samples. The machine learning models process these inexpensive data sources to generate predictions, making the methodology cost-effective and accessible for routine oilfield applications.
3Reliability
If professional staffs and equipment are required for core testing, then measurement reliability is improved, but operational complexity increases
Solution Approach 1:
The patent replaces the complex human-operated laboratory testing system with an automated machine learning pipeline. The system automatically processes logging data, selects appropriate models, and generates predictions without requiring specialized staff or equipment, while maintaining reliability through validated algorithms and quality control measures.
Solution Approach 2:
The machine learning system performs self-service by automatically selecting and applying appropriate prediction models based on the input data characteristics. The system handles data preprocessing, model selection, and result generation autonomously, eliminating the need for professional operators to manually conduct tests and interpret results.
4Manufacturing precision
If sufficient core samples are obtained through laboratory testing, then modeling accuracy in reservoir exploitation is improved, but the process becomes less practical due to resource constraints
Solution Approach 1:
The patent replaces the inefficient process of obtaining numerous core samples with a高效的 machine learning prediction system. The models are trained on limited core test data and then applied to predict relative permeability for multiple wells and scenarios, dramatically increasing productivity while maintaining modeling accuracy through the use of transfer learning and model generalization techniques.
Data Source
AI summary
The present disclosure provides a method and system for predicting a relative permeability curve based on machine learning. The present disclosure takes logging curve data as an input, and water saturation endpoint values as an output to establish a first relative permeability curve starting point model, and takes the logging curve data and a predicted water saturation starting value output from the first relative permeability curve starting point model as an input, and relative permeabilities under different water saturations as an output to establish a first relative permeability model, thereby obtaining a comprehensive prediction method for the relative permeability curve based on deep learning, and implying control mechanisms and parameters to a model.