Unsupervised Well Log Reconstruction via Autoencoder Outlier Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The acquisition of labeled high-quality well log training data is expensive and time-consuming, and the applicability of such data is limited to specific fields, making it challenging for machine learning models to reconstruct well logs in new or different fields due to varying underground geochemical properties.
Innovation Solution
An autoencoder-based workflow for well log reconstruction and outlier detection that uses neural networks to identify and remove outliers and incomplete sections from well log data, allowing for unsupervised learning and reconstruction without relying on labeled data, thereby enabling efficient and accurate log reconstruction across different fields.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If labeled high-quality well log training data is acquired manually by domain experts, then the training data quality is improved, but the cost and time consumption increase significantly
Solution Approach 1:
The system performs self-service by automatically identifying and labeling high-quality well log sections without human intervention. The machine learning model autonomously processes well log data, detects quality sections, and prepares training data independently, eliminating the need for manual expert labeling while maintaining data quality.
2Measurement precision
If labeled training data is acquired manually by domain experts, then the training data quality is improved, but the cost increases significantly
Solution Approach 1:
The system performs self-service by automatically identifying and labeling high-quality well log sections without human intervention. The machine learning model autonomously processes well log data, detects quality sections, and prepares training data independently, eliminating the need for manual expert labeling while maintaining data quality.
3Measurement precision
If labeled high-quality well log data from a specific field is used for training, then the model performance for that field is improved, but the applicability to other fields with different geochemical properties deteriorates
Solution Approach 1:
The system achieves universality by developing a machine learning model that can process and analyze well log data across multiple fields with different geochemical properties. The model is designed to be field-agnostic, learning general patterns from unlabeled data that apply universally across different geological environments, enabling it to reconstruct well logs in new fields without field-specific training data.
Solution Approach 2:
The system adapts to different fields by dynamically adjusting model parameters and learning patterns specific to each field's geochemical properties. The unsupervised learning approach allows the model to automatically adapt to varying data distributions and characteristics across different fields, maintaining performance while achieving broad applicability.
4Measurement precision
If machine learning models are trained with labeled data, then the reconstruction accuracy is improved, but the processing time increases from minutes to days
Solution Approach 1:
The system performs self-service by automatically identifying and labeling high-quality well log sections without human intervention. The machine learning model autonomously processes well log data, detects quality sections, and prepares training data independently, eliminating the need for manual expert labeling while maintaining data quality.
Data Source
Figure 1A~1D
Figure 2
Figure 3A
AI summary
A method includes receiving well log data comprising a plurality of well logs, identifying one or more sections of one or more well logs of the plurality of well logs that have substantially complete data, training a reconstruction neural network to reconstruct incomplete well logs based on the one or more sections of the one or more well logs that have substantially complete data, and reconstructing one or more incomplete well logs of the plurality of well logs using the reconstruction neural network.