Integrated Unsupervised Models for Gas-Lift Anomaly Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Gas-lift systems in hydrocarbon wells face operational inefficiencies and potential damage due to undetected anomalies, necessitating rapid detection and adjustment of parameters to maintain stability and optimize hydrocarbon production.

Innovation Solution

Employing unsupervised machine learning models, specifically isolation forest and one-class support vector machines, to analyze gas-lift data and generate anomaly metrics, which are aggregated to form an aggregate anomaly prediction, enabling real-time adjustment of operation parameters to address anomalies and restore system stability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If gas-lift systems operate without real-time anomaly detection, then system simplicity is maintained, but operational reliability deteriorates due to undetected faults causing inefficiency and equipment damage

Engineering Contradiction:

Improveoperational reliabilityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces traditional mechanical monitoring systems with machine learning-based anomaly detection models. The system uses ML algorithms to analyze operational parameters and detect anomalies automatically, eliminating the need for complex mechanical monitoring infrastructure while significantly improving operational reliability through continuous intelligent monitoring.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The gas-lift system performs self-diagnosis through integrated anomaly detection models that continuously monitor operational parameters. The system automatically identifies anomalies and generates recommendations without requiring external intervention, enabling self-service maintenance that improves reliability while keeping the monitoring architecture relatively simple.

Inventive Principle:

Principle #25Self-service

2Measurement precision

If multiple machine learning models are integrated for comprehensive anomaly detection, then detection accuracy is improved, but computational complexity and processing time increase

Engineering Contradiction:

Improveanomaly detection accuracyVSAvoidmodel integration complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the anomaly detection task into multiple specialized machine learning models, each responsible for detecting specific types of anomalies. This segmentation allows the system to achieve high detection accuracy by assigning different models to different detection functions, while managing complexity through modular architecture where each model can be independently trained and optimized.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system employs multiple machine learning models beyond what might be minimally required, using a diverse ensemble of models to ensure comprehensive anomaly detection. This excessive action approach guarantees high accuracy by capturing various anomaly patterns through different model perspectives, with the benefit that redundant models can be pruned or combined as needed.

Inventive Principle:

Principle #16Partial or excessive action

3Productivity

If real-time anomaly detection and parameter adjustment are implemented, then hydrocarbon production efficiency is improved, but operational time and computational resources are consumed

Engineering Contradiction:

Improvehydrocarbon production efficiencyVSAvoidoperational time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The anomaly detection models continuously monitor operational parameters in real-time, performing preliminary detection of potential issues before they manifest as serious problems. This allows the system to take corrective action proactively, maintaining optimal hydrocarbon production efficiency by preventing anomalies from developing into failures that would require time-consuming interventions.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements a feedback loop where anomaly detection results directly influence operational parameter adjustments. Real-time feedback from the anomaly detection models enables dynamic optimization of gas-lift parameters, improving hydrocarbon production efficiency by continuously adapting operations to current system conditions without requiring significant time delays.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20250297536A1Methods and systems for gas-lift anomaly detection using integrated unsupervised learning models

Publication Date: 2025.09.25 SAUDI ARABIAN OIL CO

AI summary

A method for determining an anomaly in a gas-lift system. The method includes obtaining gas-lift data from a gas-lift system and associated well, where the gas-lift system injects a gas into a fluid mixture of the well. The method further includes obtaining a set of operation parameters including an injected gas rate and an injected gas pressure. The method further includes determining, with a first machine learned model and a second machine learned model, a first and second anomaly metric each indicative of an anomaly in the gas-lift system or a flow of a production fluid from the well, respectively, based on the gas-lift data. The method further includes forming an aggregate anomaly prediction from the first anomaly metric and the second anomaly metric and adjusting, with a controller, the set of operation parameters based on, at least, the aggregate anomaly prediction.