Self-Learning Wellbore Trajectory Calibration for Directional Drilling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wellbore trajectory models for directional drilling are inaccurate due to unmodeled external and unpredictable effects, leading to challenges in reaching geological targets.
Innovation Solution
A self-learning controller with slow and fast learning loops (SLL and FLL) calibrates the wellbore trajectory model using survey and continuous directional measurements, combining slow control inputs for long-term effects and fast control inputs for short-term effects to improve trajectory prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a wellbore trajectory model is used to predict wellbore trajectory, then the drilling tool can be steered to reach the target, but the prediction accuracy is insufficient due to unmodeled external and unpredictable effects
Solution Approach 1:
The patent implements a feedback mechanism where actual wellbore trajectory measurements are continuously compared with predicted trajectory from the model. The differences (errors) are fed back to adjust and recalibrate the model parameters, improving prediction accuracy over time. This closed-loop feedback system addresses the reliability issue by continuously validating and correcting the model against real-world data.
Solution Approach 2:
The patent dynamically adjusts model parameters based on observed trajectory deviations. By changing parameters such as steering ratio, toolface angle, and formation interaction coefficients in response to actual measurements, the model adapts to unmodeled external effects and unpredictable formation conditions, thereby improving prediction accuracy without requiring a complete model redesign.
2Measurement precision
If survey directional measurements are taken frequently to improve calibration accuracy, then the model calibration improves, but the drilling process time increases
Solution Approach 1:
The patent applies partial calibration using only the most critical and reliable measurement data points rather than requiring complete frequent surveying. By selecting key calibration points where measurements have the greatest impact on model accuracy and ignoring less critical data collection opportunities, the system achieves sufficient calibration accuracy without the time penalty of continuous frequent surveying.
Solution Approach 2:
The patent implements continuous model calibration using available directional measurements from the drilling process itself, rather than requiring separate stopping points for surveying. The calibration process runs continuously in the background, utilizing measurements as they become available during normal drilling operations, thus maintaining drilling continuity while improving model accuracy over time.
3Measurement precision
If the wellbore trajectory model accounts for all external and unpredictable effects, then the prediction accuracy improves, but the model complexity increases
Solution Approach 1:
The patent segments the complex external effects into distinct categories: controllable steering inputs, measurable formation interactions, and unpredictable disturbances. Each segment is modeled separately with appropriate complexity, and the results are combined. This segmentation allows the model to account for multiple effect types without requiring a single overly complex unified model, managing complexity through modular decomposition.
Solution Approach 2:
The patent introduces intermediary parameters that represent the net effect of multiple external factors without requiring explicit modeling of each individual factor. These intermediary variables act as mediators between the controlled steering inputs and the actual wellbore trajectory, capturing the cumulative influence of unmodeled effects through empirical relationships rather than detailed physical mechanisms, thus reducing model complexity while maintaining accuracy.
Data Source
AI summary
One or more sensors are positioned downhole in a wellbore of a geological formation. One or more survey directional measurements associated with the wellbore are performed and one or more continuous directional measurements associated with the wellbore are performed. A wellbore trajectory model projects a trajectory of the wellbore to respective measured depths at which at least one of the one or more survey directional measurements and one or more continuous directional measurements were performed or obtained to predict one or more directional measurements of the wellbore at those depths. One or more differences are determined between the predicted one or more directional measurements and at least one of the one or more survey directional measurements and the one or more continuous directional measurements. The wellbore trajectory model is calibrated based on the one or more differences between predicted measurements and performed measurements.