Unsupervised 3D Seismic Facies Classification Without Labels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The extraction and identification of appropriate seismic facies labels for training data in seismic facies classification is expensive in terms of computing resources, and supervised ML models are costly in terms of computing time, while methods like computed self-organizing maps introduce spatial uncertainties in translating two-dimensional maps into 3D seismic volumes.
Innovation Solution
An unsupervised machine learning (USML) approach is used to transform 3D seismic data into a linear representation, allowing for efficient clustering without labeled data, using partition-based models like K-means clustering and Gaussian mixture models, selecting the optimal model and cluster number through the elbow method and silhouette scoring to generate accurate seismic facies classifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If supervised machine learning models are used for seismic facies classification, then classification accuracy is improved, but computing resource consumption and training time increase significantly
Solution Approach 1:
The patent applies unsupervised machine learning models that automatically perform seismic facies classification without requiring manually labeled training data. The system self-organizes seismic data into facies groups based on inherent patterns in seismic attributes, eliminating the time-consuming process of expert labeling while maintaining classification accuracy
Solution Approach 2:
The patent replaces the mechanical process of manual data labeling and supervised model training with automated unsupervised learning algorithms. These algorithms directly process seismic data to identify facies boundaries and characteristics, substituting human-intensive workflows with computational methods that reduce training time while preserving classification quality
2Measurement precision
If supervised machine learning models are used for seismic facies classification, then classification accuracy is improved, but computing resource consumption increases
Solution Approach 1:
The unsupervised learning system performs classification autonomously without requiring extensive pre-labeled training datasets, reducing the computational burden of model training. The algorithm automatically discovers facies patterns from raw seismic data, minimizing resource-intensive preprocessing and training operations
Solution Approach 2:
The patent extracts essential facies classification capabilities from complex supervised learning frameworks by using unsupervised algorithms that focus only on the core pattern recognition task. This extraction removes unnecessary computational overhead associated with labeled data management and supervised training, reducing overall resource consumption
3Productivity
If self-organizing maps are used for seismic facies classification, then classification is achieved, but spatial uncertainties are introduced due to translation from 2D maps to 3D volumes
Solution Approach 1:
The patent processes seismic data in three dimensions from the outset, maintaining the spatial relationships inherent in volumetric seismic data. By avoiding reduction to two-dimensional representations, the method preserves spatial accuracy while achieving automated classification, eliminating the dimensional transformation that causes spatial uncertainty
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Seismic facies modeling of an area of study at an oil and gas exploration site includes obtaining a seismic dataset. A set of unsupervised machine learning (USML) models processes a test dataset of the seismic dataset. Respective USML models of the set are configured with different cluster numbers. A USML model and corresponding elbow point cluster number is selected from the set of USML models. The selected USML model, configured with the elbow point cluster number, processes the seismic dataset to obtain clusters of the data points. Cluster profiles based on seismic cell attributes of the data points of each cluster are generated. Seismic facies labels are assigned to the clusters based on corresponding cluster profiles. The clusters are sampled to a three-dimensional (3D) grid representation of the area of study to obtain a seismic facies model displayed in a visualization tool of a seismic modeling platform.