Machine-Learning Rock Typing for Real-Time Drilling Geosteering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods struggle to accurately identify rock types in unconventional reservoirs with low porosity and permeability, leading to difficulties in geosteering and optimizing drilling parameters, as subtle changes in rock properties are not easily discernible with existing logging techniques.
Innovation Solution
A data processing system uses unsupervised and supervised machine learning models to analyze well log and core sample data, generating an unconfined compressive strength log and forming rock type clusters, enabling real-time rock type identification during drilling, which informs adjustments to drilling parameters and updates geological models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional logging techniques are used to identify rock types, then the measurement process is simple and quick, but the measurement precision is insufficient to discern subtle changes in rock properties of unconventional reservoirs
Solution Approach 1:
The rock type identification process is segmented into multiple stages: initial rock typing using conventional logs, followed by detailed facies analysis using advanced logging techniques only in target zones. This segmentation allows high-precision measurements to be applied selectively rather than throughout the entire wellbore, improving accuracy while controlling complexity and cost.
Solution Approach 2:
The system integrates multiple logging dimensions (conventional logs, advanced imaging logs, spectroscopic logs) to create a multi-dimensional characterization of rock properties. By combining data from different measurement dimensions, the system achieves superior rock type identification accuracy that cannot be obtained from single-dimension conventional logging alone.
2Measurement precision
If advanced logging techniques are deployed to improve rock type identification, then measurement precision improves, but the cost and time consumption increase
Solution Approach 1:
The logging program is segmented into primary conventional logging for the entire wellbore and secondary advanced logging for specific intervals of interest. This allows the majority of the well to be logged quickly with conventional methods, while advanced techniques are applied only where needed to achieve accurate rock type identification, thereby reducing overall time loss.
Solution Approach 2:
Conventional logging is performed first to establish baseline rock properties and identify potential target zones. Based on these preliminary results, advanced logging is then strategically deployed only in intervals where rock type identification is most critical, avoiding unnecessary time consumption in zones where conventional logging provides sufficient information.
3Adaptability or versatility
If conventional logging is used, then the device complexity is low, but the ability to discriminate different rock types in unconventional reservoirs is insufficient
Solution Approach 1:
The logging system is designed with multi-functionality, where a single integrated platform can perform both conventional logging operations and advanced rock type discrimination tasks. This universal system can adapt its measurement capabilities based on the specific reservoir type being evaluated, providing high discrimination capability for unconventional reservoirs while maintaining simplicity for conventional applications.
Solution Approach 2:
The logging program is dynamic and adaptable, allowing the system to adjust its measurement strategy based on real-time data quality and reservoir characteristics. For unconventional reservoirs where rock type discrimination is critical, the system automatically activates advanced measurement modes, whereas for conventional reservoirs, it operates in a simpler mode, thereby optimizing adaptability without permanently increasing device complexity.
Data Source
AI summary
Systems and methods include obtaining well log data and core sample data of a subsurface formation; generating, based on the well log data and the core sample data, an unconfined compressive strength log for the subsurface formation; using an unsupervised machine learning model to form rock type clusters based on the unconfined compressive strength log and the well log data; forming a training dataset including the well log data, the training dataset labeled based on the rock type clusters; training a supervised machine learning model using the training dataset. While drilling a well in the subsurface formation, logging-while-drilling data is obtained from drilling equipment used to drill the well; and rock types in the subsurface formation are determined using the supervised machine learning model and the logging-while-drilling data.