Multi-Well Log Processing Using RNNs for Real-Time Geosteering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing logging-while-drilling (LWD) and measurement-while-drilling (MWD) techniques face inaccuracies due to interference and heterogeneity in rock formations, limiting effective geosteering and reservoir characterization.
Innovation Solution
The use of a recurrent neural network, such as a long short-term memory (LSTM) network, processes well log data to generate adjusted logging and drilling parameters in real-time, accounting for interference and heterogeneity, thereby improving drilling performance and image coherence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional LWD/MWD measurements are used to determine subsurface formation properties, then drilling operations can proceed with real-time data, but the measurements become inaccurate due to interference and unique wellbore conditions
Solution Approach 1:
The patent introduces an intermediary system consisting of a recurrent neural network and image processing algorithms that act as a mediator between the raw LWD/MWD measurements and the final formation property determination. This intermediary processes the measurements to remove interference effects and reconstruct accurate formation properties, effectively isolating the measurement system from the harmful wellbore conditions.
Solution Approach 2:
The patent replaces the traditional mechanical/physical measurement interpretation methods with a computational approach using recurrent neural networks and image processing. This substitution allows the system to handle complex interference patterns and wellbore conditions that traditional methods cannot resolve, improving measurement reliability in challenging environments.
2Productivity
If real-time processing of well log data is implemented to improve drilling decision-making, then drilling operations can be optimized, but the complexity of processing and analyzing data from multiple wells increases
Solution Approach 1:
The patent segments the complex task of multi-well data processing into distinct functional modules: data acquisition from multiple wells, recurrent neural network processing, image generation, and drilling parameter determination. This segmentation allows each module to be optimized independently and simplifies the overall system architecture, making real-time processing more manageable despite the inherent complexity.
Solution Approach 2:
The patent implements a dynamic processing system where the recurrent neural network continuously adapts to new data as it becomes available during drilling operations. The system dynamically adjusts its processing based on the incoming data stream from multiple wells, enabling real-time decision-making while managing complexity through adaptive rather than static processing approaches.
Data Source
AI summary
A method may include obtaining well log data regarding a geological region of interest. The well log data may correspond to logging-while-drilling (LWD) measurements or measurement-while-drilling (MWD) measurements acquired from various wellbores. The method may further include generating image data regarding the geological region of interest using the well log data and a recurrent neural network. The method may further include determining a drilling parameter for a wellbore among the wellbores in real-time using the image data. The drilling parameter may be determined while the well log data is being acquired in the wellbore. The method may further include transmitting, based on the drilling parameter, a command to a drilling system coupled to the wellbore.