Salt Water Disposal Well Control for Real-Time Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for monitoring and managing salt water disposal (SWD) wells lack real-time data processing and predictive analytics, leading to potential casing leaks and reduced oil and gas production due to untimely detection of performance declines.
Innovation Solution
A data processing system utilizing machine learning models to analyze real-time operational data from SWD wells, including injection pressure, rate, and casing-casing annulus pressure, to detect anomalies and automate control measures, such as adjusting choke settings, thereby minimizing manual intervention and improving well integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time data processing and machine learning models are implemented, then detection accuracy and response time improve, but system complexity and computational requirements increase
Solution Approach 1:
The system performs preliminary actions by pre-training machine learning models with historical well performance data before deployment. The models are pre-configured with knowledge of normal and abnormal well patterns, enabling them to quickly analyze real-time data without requiring complex real-time computation for pattern recognition. This preliminary preparation reduces the complexity of real-time processing while maintaining high detection accuracy.
Solution Approach 2:
The patent introduces machine learning models as intermediary components between raw sensor data and anomaly detection. These models act as intelligent mediators that automatically process and interpret complex operational data, transforming raw measurements into meaningful insights. This intermediary layer simplifies the overall system architecture by consolidating complex analysis functions within the trained models rather than requiring complex real-time processing infrastructure.
2Productivity
If automated control measures are implemented, then manual intervention is minimized, but automation extent and system complexity increase
Solution Approach 1:
The system implements self-service automation where the machine learning models autonomously detect anomalies and trigger appropriate control measures without requiring manual engineer intervention. The automated system serves itself by continuously monitoring well performance, identifying issues, and executing corrective actions such as adjusting injection rates or alerting operators only when necessary. This self-service approach improves operational efficiency while keeping automation at an appropriate level that maintains human oversight for critical decisions.
3Reliability
If real-time monitoring of multiple parameters is implemented, then well integrity is improved, but data processing requirements and system complexity increase
Solution Approach 1:
The system extracts and focuses on the most critical parameters for well integrity monitoring, such as injection pressure, temperature, and flow rate, rather than processing all available data equally. The machine learning models are trained to identify and prioritize the most informative parameters for detecting casing leaks and other integrity issues. This selective extraction approach maintains high well integrity monitoring while reducing the overall data processing load by concentrating computational resources on the most relevant measurements.
Data Source
AI summary
Systems and methods for monitoring and managing day to day performance of a group of salt water disposal wells in a network include performing operations including performing a well measurement to generate salt-water disposal (SWD) well data; retrieving a machine learning model that is trained using labeled SWD data, the labeled SWD data representing one or more SWD well faults causing respective data signatures in the SWD data, each respective data signature being associated with a corresponding label identifying the SWD well fault for a particular SWD well; inputting the SWD well data generated in real-time based on SWD well operation, into the machine learning model; generating, by the machine learning model based on the inputting, a classification output representing a predicted well injection pressure or injection rate of the particular SWD well including SWD wells faults/anomalies; and generating, based on the classification output, control data for changing operation of the particular SWD well.