Automated Well Data Channel Mapping Using Super and Sub-Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual mapping of signal data from wells undergoing completion operations, such as hydraulic fracturing, is inefficient and often results in inaccurate or incomplete data channel assignments, especially in cases of insufficient human resources or real-time data processing, leading to reduced accuracy in downstream processing.
Innovation Solution
An automated method using a processor to access and classify well data sequences into correct categories through generating summary statistics and processing them with trained super and sub-models, such as random forest classifiers, to accurately map unknown well data channels, thereby improving data quality and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual mapping is used to assign data channels, then accuracy in data channel assignment can be maintained through expert review, but productivity is reduced due to insufficient human resources and time constraints
Solution Approach 1:
The system performs automatic self-mapping of data channels using machine learning models that analyze data characteristics and assign channels without human intervention. The automated system serves itself by generating summary statistics, processing them through trained models, and producing channel assignments independently, eliminating the need for manual expert review while maintaining high accuracy through algorithmic classification.
2Productivity
If automated mapping is implemented to improve productivity, then processing speed increases, but measurement precision may deteriorate due to potential incorrect mappings
Solution Approach 1:
The patent replaces the mechanical system of manual expert review with an automated machine learning system that processes data channels through trained models. The system substitutes human cognitive processes with computational algorithms that generate summary statistics and apply classification models to assign data channels automatically, achieving both high speed and accuracy through algorithmic decision-making rather than human judgment.
Solution Approach 2:
The system incorporates feedback mechanisms where the machine learning models are trained on historical data and continuously improve their classification accuracy. The automated mapping process provides feedback loops that allow the system to learn from previous mappings, adjust its classification criteria, and refine channel assignments over time, ensuring high precision while maintaining automated productivity.
3Loss of time
If real-time data processing is required, then response time is reduced, but the complexity of accurate channel mapping increases making it difficult or impossible without automation
Solution Approach 1:
The system performs preliminary actions by pre-training machine learning models on historical data before real-time processing is needed. Summary statistics generation and model training are completed in advance, so that during real-time operations, the system only needs to apply pre-trained classification rules to incoming data channels. This preliminary preparation reduces the computational burden during real-time processing, enabling fast response times while managing system complexity through advance setup.
Data Source
AI summary
A method of mapping unknown well data channels involves generating, with a processor, summary statistics for unknown sequences of well data values where the well data values pertain to respective particular types of data and where each type belongs to a category of well data. The system processes the respective summary statistics with a trained super model to classify the respective sequences of well data values into categories of well data. The system then processes the summary statistics for a given sequence using a trained sub-model of the category to which the sequence belongs in order to further classify the sequence as to its particular type within the category thereby mapping originally unknown sequences of well data values to the correct category and type of well data.