Stochastic Drilling Path Optimization for Uncertain Formations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Directional drilling faces challenges due to information uncertainty from unknown formation properties, sensor noise, and steering inaccuracies, which complicates trajectory planning and optimization of operating parameters, leading to inefficiencies and increased non-productive time.
Innovation Solution
The implementation of stochastic path optimization methods using probability distributions to determine optimal operating parameters and path design parameters, accounting for both operational and environmental uncertainties, through a processing system that retrieves historical and real-time data, models formation properties, and employs Monte Carlo methods for real-time optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional deterministic trajectory planning is used, then the drilling path can be calculated, but the reliability is poor due to information uncertainty from unknown formation properties, sensor noise, and steering inaccuracies
Solution Approach 1:
The patent transforms the deterministic trajectory planning parameters into probabilistic parameters with associated uncertainty distributions. Instead of using fixed formation properties and drilling parameters, the system employs probability density functions to represent uncertainties in formation properties, sensor measurements, and operating parameters, enabling stochastic optimization that accounts for information uncertainty
Solution Approach 2:
The system implements continuous feedback by updating the probability density functions and stochastic models with real-time drilling data and sensor measurements. The uncertainty characterizations are refined as new information becomes available, allowing the trajectory planning to adapt to actual formation conditions and reduce uncertainty over time
2Productivity
If real-time optimization of operating parameters is implemented, then productivity improves, but the device complexity increases due to stochastic modeling and Monte Carlo simulations
Solution Approach 1:
The patent replaces complex mechanical trial-and-error optimization with computational stochastic optimization. Instead of physically testing different operating parameter combinations, the system uses Monte Carlo simulations and probability-based models to computationally determine optimal parameters, reducing physical experimentation while improving optimization accuracy
Solution Approach 2:
The system performs preliminary stochastic modeling and uncertainty characterization before actual drilling operations begin. Probability density functions for formation properties and operating parameters are established in advance, and the stochastic optimization framework is prepared, enabling rapid real-time optimization during drilling without extensive computational overhead
3Measurement precision
If uncertainty characterization is performed using probability density functions, then the measurement precision improves, but the loss of time increases due to computational requirements
Solution Approach 1:
The patent applies partial stochastic optimization by focusing computational efforts on the most critical uncertainties that have the greatest impact on trajectory accuracy and drilling efficiency. Instead of fully optimizing all parameters simultaneously, the system identifies key uncertain parameters and applies stochastic methods selectively to those, reducing computational time while maintaining measurement precision for the most important variables
Data Source
Figure 1~2
Figure 3A~3D
Figure 4
AI summary
A disclosed drilling method includes: obtaining a formation model representing formation properties to be encountered by a drilling assembly being steered towards a target; identifying at least one path-dependent drilling dynamics model for predicting the drilling assembly's response to one or more operating parameters; characterizing uncertainties associated with said formation model and said at least one drilling dynamics model, said characterizing yielding a probability density function for each uncertainty; representing an acceptable range for each of said one or more operating parameters as a probability density function; employing the probability density functions to determine random samples of said uncertainties and of said one or more operating parameters; applying a cost function to the random samples to determine an expected cost as a function said one or more operating parameters; and displaying the randomly sampled operating parameters having a minimum expected cost as optimized operating parameters.