Higher Order Channels for Hydraulic Fracturing Event Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Time-series data from the oil and gas industry, particularly related to drilling and hydraulic fracturing, is often processed manually, limiting the ability to identify hidden features and reducing analysis accuracy and speed due to the lack of consideration for higher-order or derived data channels.
Innovation Solution
A computer-implemented method and system that generates additional data channels from time-series data, including preprocessing and normalizing hydraulic fracturing data to identify events such as proppant ramps, acid pulses, and target slurry rates, using machine learning processes and cloud-based platforms for real-time analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual processing methods are used for time-series data, then operational simplicity is maintained, but analysis accuracy and speed deteriorate due to inability to identify hidden features
Solution Approach 1:
The patent replaces manual mechanical processing with automated computational systems including pre-processing modules, higher-order channel generation algorithms, and machine learning models. This substitution enables automated identification of hidden features through mathematical transformations and pattern recognition, significantly improving analysis accuracy while maintaining operational simplicity through automation.
Solution Approach 2:
The patent introduces higher-order channels as intermediary data structures that transform raw time-series data into enhanced representations. These intermediate channels serve as mediators between raw sensor data and final analysis results, enabling the system to capture hidden features and relationships that are not apparent in the original data without requiring complex direct processing.
2Difficulty of detecting and measuring
If only native time-series channels are used without derived channels, then data processing simplicity is maintained, but feature identification capability deteriorates
Solution Approach 1:
The patent segments the data processing task into distinct stages: pre-processing of raw channels, generation of higher-order derived channels through mathematical transformations, and final feature identification using machine learning models. This segmentation allows each component to specialize in specific transformations, improving feature detection capability while organizing complexity into manageable modular components.
Solution Approach 2:
The patent transforms one-dimensional time-series data into multi-dimensional representations by generating higher-order channels that incorporate temporal derivatives, integrals, and cross-channel relationships. This dimensionality expansion enables the system to detect features across multiple scales and relationships simultaneously, significantly improving feature identification capability.
3Productivity
If automated processing systems are implemented, then analysis speed and accuracy improve, but system complexity increases
Solution Approach 1:
The patent implements a universal automated processing framework that handles multiple data types (pressure, volume, concentration, temperature) and multiple analysis tasks (feature detection, event identification, parameter optimization) through a single integrated system. This multi-functional approach improves analysis speed across all operations while consolidating complexity into a unified architecture rather than requiring separate systems for each function.
Solution Approach 2:
The automated processing system performs self-service through adaptive machine learning models that automatically adjust parameters and thresholds based on the specific characteristics of each well and treatment scenario. This self-adjustment capability enables the system to maintain high analysis speed and accuracy across diverse applications without requiring manual reconfiguration, effectively managing system complexity through autonomy.
Data Source
AI summary
A method for identifying characteristics of well data comprises receiving hydraulic fracturing data comprising data channels from a stage or stages of a hydraulic fracturing sequence and preprocessing the data channel, which may include normalizing and recalibrating the hydraulic fracturing data. The method further involves generating one or more additional higher order channels based on the normalized and recalibrated hydraulic fracturing data, the one or more additional channels derived at least in part from parameters of the normalized and recalibrated hydraulic fracturing data. The system may combine the higher order channels with the accessed data channels, and further process, combine and otherwise identify hydraulic fracturing events at the well being hydraulically fractured based on the received hydraulic fracturing data and the one or more additional channels. From the identified events, the system may alter hydraulically fracturing attributes of a stage being completed and/or subsequent stages of the well or subsequent wells.