Neural Network Seismic Layer Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current drilling technologies lack the accuracy and efficiency in determining the locations and boundaries of subsurface seismic layers, which are crucial for precise resource extraction, as existing methods struggle to effectively predict the presence and characteristics of subsurface layers like salt layers.

Innovation Solution

A system utilizing machine learning processes, specifically neural networks, to analyze seismic cubes and predict the locations and boundaries of subsurface seismic layers, improving upon current drilling technology by providing accurate and efficient predictions for optimal drill placement.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional drilling technologies are used to determine subsurface layer locations, then the drilling process can be performed with existing equipment, but the accuracy and efficiency in determining subsurface layer boundaries is insufficient

Engineering Contradiction:
Improveaccuracy of determining subsurface layer locationsVSAvoidefficiency of drilling operations
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces traditional mechanical and manual interpretation methods with machine learning algorithms and neural networks. The system uses trained machine learning models to automatically analyze seismic cube data and predict subsurface layer boundaries, substituting human interpretation and conventional processing methods with automated intelligent systems that provide both high accuracy and efficient processing

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

Solution Approach 2:

The patent transforms the approach by changing from direct seismic data interpretation to using machine learning models that process multiple seismic attributes simultaneously. The system converts raw seismic cube data into predicted layer location predictions through trained neural networks, changing the analytical parameters from individual seismic measurements to integrated multi-attribute analysis

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If machine learning processes are implemented to predict subsurface layer locations, then the accuracy of determining subsurface layers is improved, but the complexity of the system increases

Engineering Contradiction:
Improveprediction accuracy of subsurface layer locationsVSAvoidcomplexity of machine learning system
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by training the machine learning models in advance using labeled seismic data before deployment. The neural networks are pre-trained with known subsurface layer locations and seismic characteristics, so that during actual operation, the system can directly apply these pre-learned patterns to new seismic cubes without requiring complex real-time training processes

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary layer between raw seismic data and final predictions by using trained machine learning models as mediators. These models act as intermediate processing systems that translate complex seismic cube data into interpretable layer location predictions, simplifying the overall system architecture while maintaining high accuracy

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3529639B1Determining subsurface layers using machine learning
Publication Date: 2022.03.30 SERVICES PETROLIERS SCHLUMBERGER SA
  • EP3529639B1 patent drawingFigure 1
  • EP3529639B1 patent drawingFigure 2
  • EP3529639B1 patent drawingFigure 3

AI summary

A method is disclosed and includes receiving a seismic cube. The seismic cube includes a three-dimensional image of a portion of a subsurface area. The method further includes providing the seismic cube to a machine learning process. The machine learning process includes one or more neural networks used for predicting a location of a subsurface seismic layer in the received seismic cube. The method also includes receiving, from the machine learning process, the prediction of the location of the subsurface seismic layer in the seismic cube.