Drillstring Off-Bottom Detection Using Adaptive Surface Data Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing drilling technologies face challenges in efficiently navigating and maintaining the trajectory of boreholes, particularly in complex subsurface environments, leading to issues such as stuck pipe and inefficient resource extraction due to inadequate real-time data integration and adaptive drilling strategies.
Innovation Solution
Implementing a system that integrates machine learning models with drilling operations to predict and adapt to subsurface conditions, utilizing sensors and real-time data to optimize drillstring conditions and trajectory, incorporating geosteering methods for precise wellbore navigation and resource extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If real-time data integration and adaptive drilling strategies are implemented, then drilling efficiency and accuracy are improved, but system complexity increases
Solution Approach 1:
The system segments drilling operations into multiple measurable parameters (weight on bit, torque, hook load, RPM, flow rate, pressure, temperature) and processes each parameter separately through machine learning models. This segmentation allows complex drilling operations to be managed through modular, independent analysis of individual parameters while maintaining overall system efficiency.
Solution Approach 2:
The system implements continuous feedback loops where real-time sensor data from the drillstring is fed into machine learning models that predict optimal drilling parameters. These predictions are then fed back to adjust drilling operations dynamically, creating a closed-loop control system that improves efficiency through adaptive decision-making based on actual subsurface conditions.
2Measurement precision
If machine learning models are used to predict subsurface conditions, then measurement precision is improved, but computational requirements increase
Solution Approach 1:
The system performs preliminary training of machine learning models using historical drilling data before actual drilling operations. This preliminary action creates pre-trained models that can make rapid predictions during real-time operations without requiring intensive computational resources during the actual drilling process, thus balancing accuracy with energy consumption.
Solution Approach 2:
The system uses virtual representations (digital twins) of the drillstring and subsurface conditions to simulate and predict drilling behavior. These virtual copies allow complex subsurface conditions to be analyzed computationally without requiring physical experimentation, reducing energy consumption while maintaining prediction accuracy.
3Manufacturing precision
If geosteering methods are implemented for precise wellbore navigation, then manufacturing precision is improved, but operational complexity increases
Solution Approach 1:
The geosteering system automatically adjusts wellbore trajectory using real-time data and machine learning predictions without requiring constant manual intervention. The system self-corrects trajectory deviations by analyzing sensor data and autonomously determining optimal steering actions, reducing operational complexity while maintaining high precision in wellbore navigation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method can include identifying a threshold value for a drillstring off-bottom condition determination; filtering time series surface data of a post-connection drilling state of the drillstring using the threshold value to generate filtered time series surface data; and statistically determining a drillstring off-bottom condition value using the filtered time series surface data.