Real-Time Drilling Event Detection via Weighted Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Operators face limitations in monitoring multiple well-site operations simultaneously due to the need for vigilant focus on real-time data monitoring, which can lead to missed events or increased attention requirements, affecting the number of operations they can manage at any given time.
Innovation Solution
Implementing a system with downhole tools and information handling systems that use real-time data processing, machine-learning algorithms, and statistical analysis to automatically detect and alert operators of events by calculating scores based on predetermined rules, allowing for proactive adjustments and increased monitoring capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If operators manually monitor multiple well-site operations simultaneously, then they can detect events in real-time, but the number of operations they can monitor is limited due to focus requirements
Solution Approach 1:
The patent introduces an automated event detection system that acts as an intermediary between the monitored operations and the operator. This system includes data collection modules that gather operational data, event detection modules that analyze the data using machine learning algorithms, and notification modules that alert operators to detected events. By placing this intermediary system in place, the operator no longer needs to directly monitor all operations, thereby resolving the contradiction between reliable event detection and the number of operations that can be monitored.
Solution Approach 2:
The event detection system is designed to autonomously perform data collection, analysis, and event detection without requiring continuous operator attention. The system self-manages the monitoring process by automatically collecting data from multiple operations, analyzing it through predefined algorithms and machine learning models, and generating notifications when events are detected. This self-service capability allows a single operator to effectively monitor multiple operations simultaneously, resolving the limitation imposed by human focus requirements.
2Measurement precision
If operators focus vigilantly on real-time data monitoring, then event detection accuracy improves, but the amount of attention required increases
Solution Approach 1:
The patent replaces the mechanical system of human vigilance and manual data analysis with an automated computational system. The event detection modules use machine learning algorithms and statistical analysis to process operational data, substituting the need for human cognitive resources. This substitution maintains high detection accuracy through sophisticated algorithms while dramatically reducing the operator's attention requirement, as the system operates autonomously and only requires operator intervention when events are detected.
Solution Approach 2:
The system incorporates feedback mechanisms where detected events trigger notifications to operators, who can then provide feedback by confirming or correcting the detection. This feedback loop allows the system to learn from operator responses and improve its detection accuracy over time. The feedback principle enables the system to maintain high measurement precision while requiring minimal ongoing operator attention, as the bulk of the analytical work is performed automatically.
3Reliability
If multiple operations are monitored manually, then comprehensive oversight is achieved, but the risk of missed events increases
Solution Approach 1:
The monitoring system is segmented into multiple specialized modules, each responsible for specific aspects of event detection. Data collection modules are divided into separate components that gather different types of operational data, while event detection modules are segmented to analyze different parameters and patterns. This segmentation allows the system to comprehensively monitor multiple operations simultaneously with dedicated analysis for each, reducing the risk of missed events while maintaining reliable oversight across all monitored operations.
Data Source
AI summary
A method for detecting the occurrence of an event at a well comprising: recording a first set of measurements collected from the well; processing the recorded measurements to provide a first set of data; calculating a first score based on the first set of processed data; recording a second set of measurements collected from the well; processing the second set of recorded measurements to provide a second set of data; calculating a second score based on the second set of processed data; wherein the first set of processed data and the second set of processed data is the number median of absolute difference from normal; combining the first score and the second score using a weighted average; and comparing the combined scores to a predetermined rule based on a specified time frame; and allowing an operator to adjust parameters of the well operation based on the score comparison.