Time Series Event Flagging for Downhole Operations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual analysis of time series data from wellsite operations is time-consuming and repetitive, requiring experts to visually inspect parameters to identify relevant events, which limits efficiency and increases the risk of operational failures.
Innovation Solution
A method and system that automatically classify unlabeled subsequences of time series data into predetermined clusters using a loss function, allowing for the assignment of events of interest and enabling real-time decision-making by extracting and labeling subsequences representative of specific operations, such as pressure tests, and sending commands to the wellsite operating system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual visual inspection by experts is used to analyze time series data, then measurement precision and event recognition accuracy are improved, but loss of time and productivity deteriorate
Solution Approach 1:
The patent creates a digital copy of the expert's analytical capability through a machine learning model. The model is trained on time series data labeled by experts, capturing the patterns and features that experts use to identify events of interest. This digital copy can then process large volumes of data continuously without fatigue, achieving both high accuracy and efficiency.
Solution Approach 2:
The patent replaces the mechanical process of human visual inspection with an automated computational system. The machine learning model processes time series data through mathematical operations, substituting the expert's cognitive and visual processing with algorithmic analysis that can operate continuously and scale indefinitely.
2Measurement precision
If manual selection of relevant data portions is performed, then measurement precision is improved, but productivity and ease of operation deteriorate
Solution Approach 1:
The system performs self-service by automatically identifying and extracting relevant data portions without human intervention. The machine learning model independently analyzes time series data, selects relevant segments, and prepares them for further processing, eliminating the need for manual data selection and significantly improving productivity.
Solution Approach 2:
The patent applies preliminary action by pre-processing and preparing data automatically before it is needed for analysis. The system continuously monitors time series data and pre-identifies relevant events and patterns, so that when analysis is required, the data is already organized and ready, eliminating manual preparation time.
3Reliability
If continuous monitoring of parameters is performed, then reliability is improved, but loss of time for expert review deteriorates
Solution Approach 1:
The patent implements feedback by continuously monitoring system parameters and automatically providing insights and alerts back to operators. The machine learning model analyzes incoming data in real-time and provides immediate feedback about events of interest, allowing experts to receive summarized information rather than raw data streams, thus maintaining reliability while reducing review time.
Solution Approach 2:
The patent extracts only the most critical information from continuous monitoring data. The machine learning model filters through vast amounts of continuous data and extracts only the relevant events and anomalies that require expert attention, separating signal from noise and allowing experts to focus only on what matters.
Data Source
AI summary
The disclosure relates to a method for flagging at least an event of interest in an unlabeled time series of a parameter relative to a wellsite (including to the well, formation or a wellsite equipment), wherein the time series of the parameter is a signal of the parameter as a function of time. The disclosure also relates to a method for evaluation a downhole operation such as a pressure test using a pressure time series. Such methods comprises collecting a time series, extracting at least an unlabeled subsequence of predetermined duration in the time series, and assigning an event of interest a label, in particular representative of the status of the downhole operation, to at least one of the unlabeled subsequences. A command may be sent to a wellsite operating system based on assigned label.