Seismic Subsurface Prediction via Deep Learning Dimensionality Reduction

Solution Overview

Problem

Existing backpropagation-enabled processes for seismic data analysis are inefficient and computationally intensive, especially when requiring detailed and refined classification of subsurface features, leading to increased computational time and potential human error in interpretation.

Innovation Solution

A method for training a backpropagation-enabled regression process that inputs multi-dimensional seismic data and computes predicted attribute values with a reduced dimensionality, utilizing deep learning techniques such as convolutional neural networks to improve accuracy and efficiency while reducing computational resources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If detailed and refined classification of subsurface features is performed using existing backpropagation-enabled processes, then prediction accuracy is improved, but computational time and resource consumption increase significantly

Engineering Contradiction:

Improveprediction accuracyVSAvoidcomputational time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the high-dimensional seismic data into multiple lower-dimensional subspaces using dimensionality reduction techniques. By dividing the complex prediction task into smaller dimensional components, the system achieves detailed classification accuracy while reducing the computational burden associated with processing full-dimensional data.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transforms the problem from high-dimensional space to lower-dimensional space through dimensionality reduction. This dimensional transformation allows the system to maintain prediction accuracy by preserving essential feature relationships while operating in a computationally more efficient lower-dimensional environment.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Measurement precision

If detailed and refined classification of subsurface features is performed using existing backpropagation-enabled processes, then prediction accuracy is improved, but computational resource consumption increases

Engineering Contradiction:

Improveprediction accuracyVSAvoidcomputational resource consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the high-dimensional seismic data into multiple lower-dimensional subspaces using dimensionality reduction techniques. By dividing the complex prediction task into smaller dimensional components, the system achieves detailed classification accuracy while reducing the computational burden associated with processing full-dimensional data.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transforms the problem from high-dimensional space to lower-dimensional space through dimensionality reduction. This dimensional transformation allows the system to maintain prediction accuracy by preserving essential feature relationships while operating in a computationally more efficient lower-dimensional environment.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

3Adaptability or versatility

If conventional seismic interpretation processes are used, then human interpretation flexibility is maintained, but human error and time-intensive analysis occur

Engineering Contradiction:

Improveinterpretation flexibilityVSAvoidhuman error rate

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent implements self-service through automated machine learning models that perform seismic data interpretation independently. The system learns from training data and automatically performs classification and prediction tasks, eliminating reliance on human interpreters while maintaining adaptability through continuous learning and model updating capabilities.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent substitutes human mechanical interpretation processes with automated computational systems. By replacing human analysts with machine learning models, the system eliminates human error while maintaining interpretative capabilities through algorithmic pattern recognition and decision-making processes.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11698471B2Method for predicting subsurface features from seismic using deep learning dimensionality reduction for regression

Publication Date: 2023.07.11 SHELL USA INC

AI summary

A method for training a backpropagation-enabled regression process is used for predicting values of an attribute of subsurface data. A multi-dimensional seismic data set with an input dimension of at least two is inputted into a backpropagation-enabled process. A predicted value of the attribute has a prediction dimension of at least 1 and is at least 1 dimension less than the input dimension.