Geosteering Neural Network Data Reconciliation

Solution Overview

Problem

Current geosteering methods face challenges in reconciling and accurately predicting subsurface lithology and physical properties ahead of a drill bit, as existing data from logging while drilling, electromagnetic, and seismic surveys often yield inconsistent and noisy results, hindering precise wellbore trajectory planning.

Innovation Solution

The implementation of a deep learning method involving multiple neural networks for data reconciliation, where physical parameters from various geophysical data sources are estimated, preprocessed, and converted into reconciled parameters, allowing for improved geosteering decisions based on consistent and accurate subsurface interpretations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If multiple geophysical data sources (LWD, EM, seismic) are used to estimate physical parameters, then the quantity and variety of information increases, but the consistency and reliability of the parameters deteriorate due to noisy and conflicting results

Engineering Contradiction:
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent introduces an intermediary reconciliation process that mediates between multiple geophysical data sources. The system uses trained neural networks to transform and harmonize parameters from different sources (LWD, EM, seismic) into a unified set of reconciled physical parameters, resolving conflicts and improving consistency while preserving the benefits of using multiple data sources.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements a feedback mechanism where reconciled physical parameters are continuously compared against training data and previous interpretations. The neural networks are trained using feedback from known subsurface conditions to improve the reconciliation process, creating a closed-loop system that enhances reliability through iterative refinement.

Inventive Principle:
Principle #23Feedback

2Productivity

If traditional data processing methods are used to interpret subsurface properties, then the process is simpler and faster, but the measurement precision and accuracy of lithology and physical parameter predictions deteriorate

Engineering Contradiction:
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent replaces traditional mechanical data processing methods with an intelligent system based on neural networks. The system automates the reconciliation process using machine learning models that have been trained on extensive datasets, achieving both high speed (through automated processing) and high precision (through intelligent pattern recognition and reconciliation of multiple data sources).

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system performs preliminary action by pre-training neural networks on extensive training datasets before actual geosteering operations. This pre-processing creates ready-to-use reconciliation models that can quickly and accurately process new data during drilling operations, achieving both speed and precision through advance preparation.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If manual reconciliation of physical parameters is performed, then the interpretation quality may be higher, but the time consumption and operational complexity increase

Engineering Contradiction:
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system implements self-service by enabling automated reconciliation of physical parameters through trained neural networks. The system performs the reconciliation process autonomously without requiring manual intervention, achieving both high quality interpretation (through intelligent algorithms) and rapid processing (through automation), thereby eliminating the time loss and complexity associated with manual methods.

Inventive Principle:
Principle #25Self-service

4Reliability

If complex neural network models are trained to reconcile physical parameters, then the reliability of predictions improves, but the device complexity and computational requirements increase

Engineering Contradiction:
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the complex reconciliation task into multiple specialized neural networks, each trained to handle specific aspects of parameter reconciliation. This modular approach improves reliability through specialized processing while managing complexity by organizing the system into discrete, manageable components that can be independently trained and optimized.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS20240255668A1Geosteering using improved data conditioning
Publication Date: 2024.08.01
  • US20240255668A1 patent drawing
  • US20240255668A1 patent drawing
  • US20240255668A1 patent drawing

AI summary

Systems and methods for geosteering using improved data conditioning are disclosed. The methods include estimating physical parameters from a training dataset including remote sensing data; preprocessing the estimated physical parameters; training a first neural network; training a second neural network; training a third neural network; converting estimated physical parameters into the rock characteristics with the first neural network; and converting rock characteristics into reconciled physical parameters with the second neural network. The methods further include obtaining new remote sensing data; estimating new estimated physical parameters from the new remote sensing data; converting new estimated physical parameters into new reconciled physical parameters with the third neural network; and performing geosteering of a well based on a subsurface geology interpreted from the new reconciled physical parameters.