Real-Time Formation Prediction Using Deep Learning Neural Networks

Solution Overview

Problem

Current geosteering methods lack the ability to accurately predict formation properties ahead of the drill bit in real-time, limiting the precision of wellbore placement and hydrocarbon extraction efficiency.

Innovation Solution

A system utilizing a deep learning neural network based on restricted Boltzmann machines to classify lithology and saturation patterns in real-time, processing logging-while-drilling data to provide predictions of formation properties up to 60-90 feet ahead of the drill bit, enabling geosteering adjustments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If real-time classification of formation properties is performed using machine learning models, then geosteering precision is improved, but computational complexity and data processing requirements increase

Engineering Contradiction:

Improvegeosteering precisionVSAvoidcomputational complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary classification of formation properties using machine learning models before the drill bit reaches those formations. By processing LWD data and predicting formation properties ahead of the bit (up to 60-90 feet), the system enables proactive geosteering decisions rather than reactive adjustments, improving precision while managing computational load through advance processing

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The machine learning model acts as an intermediary between the raw LWD data and the geosteering control decisions. It processes and interprets complex geological data, transforming it into actionable predictions about formation properties ahead of the bit, thereby simplifying the decision-making process while maintaining high precision

Inventive Principle:

Principle #24Intermediary (Mediator)

2Manufacturing precision

If formation properties are predicted ahead of the drill bit, then wellbore placement accuracy is improved, but the system requires more sophisticated data processing capabilities

Engineering Contradiction:

Improvewellbore placement accuracyVSAvoiddata processing capabilities

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary classification of formation properties using machine learning models before the drill bit reaches those formations. By processing LWD data and predicting formation properties ahead of the bit (up to 60-90 feet), the system enables proactive geosteering decisions rather than reactive adjustments, improving precision while managing computational load through advance processing

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional mechanical/geophysical measurement methods with machine learning-based computational models. Instead of relying solely on direct physical measurements at the bit, the system uses AI algorithms to interpret LWD data and predict formation properties, achieving higher placement accuracy through intelligent data processing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If real-time geosteering adjustments are made based on predicted formation properties, then hydrocarbon extraction efficiency is improved, but the system response time requirements increase

Engineering Contradiction:

Improvehydrocarbon extraction efficiencyVSAvoidsystem response time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary classification of formation properties using machine learning models before the drill bit reaches those formations. By processing LWD data and predicting formation properties ahead of the bit (up to 60-90 feet), the system enables proactive geosteering decisions rather than reactive adjustments, improving precision while managing computational load through advance processing

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system continuously monitors LWD data and uses machine learning models to predict formation properties, providing real-time feedback for geosteering adjustments. This feedback loop enables dynamic wellbore trajectory modifications based on predicted geological conditions, optimizing hydrocarbon extraction efficiency while maintaining rapid response capability

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20240254874A1Methods and systems for predicting conditions ahead of a drill bit

Publication Date: 2024.08.01 SAUDI ARABIAN OIL CO

AI summary

A method for predicting conditions ahead of a drill bit while drilling a well involves performing, using a machine learning model, a classification of formation properties ahead of the drill bit, based on data that includes logging-while-drilling (LWD) data obtained while drilling the well.