Transformer DGA Alarm Detection With Adaptive Edge Thresholds

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing transformer fault detection systems using static alarm thresholds for dissolved gases generate false positives due to age-related gas concentration changes and load variations, leading to unnecessary maintenance and shutdowns, and lack computational efficiency and cybersecurity.

Innovation Solution

Implementing adaptive DGA alarms at an edge device using dynamic thresholds based on rolling window averages, rate-of-change (RoC), and log-ratio changes, with configurable percentile thresholds, enabling accurate fault detection and reducing false alarms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If static alarm thresholds are used for dissolved gas detection, then the system is simple to implement, but false positives increase due to age-related gas concentration changes and load variations

Engineering Contradiction:

Improvealarm threshold systemVSAvoidfault detection accuracy

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent transforms static alarm thresholds into dynamic thresholds that automatically adapt to transformer conditions. The system calculates rolling window averages of historical gas concentration data and uses these dynamic baselines to determine alarms, allowing the thresholds to change with transformer age and operational conditions while maintaining detection reliability

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the parameter of alarm thresholds from fixed values to variable values based on historical data statistics. By computing rolling window averages and standard deviations, the thresholds dynamically adjust to reflect normal variations in gas concentrations due to aging and load changes, reducing false positives while maintaining sensitivity to actual faults

Inventive Principle:

Principle #35Parameter changes

2Reliability

If dynamic thresholds with rolling window averages are implemented, then false alarm reduction is achieved, but computational overhead increases

Engineering Contradiction:

Improvefault detection accuracyVSAvoidcomputational overhead

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system performs self-service by automatically calculating rolling window averages and dynamic thresholds using its own historical data, without requiring external intervention or complex external systems. The edge device autonomously processes the data and adjusts thresholds based on its accumulated knowledge of normal transformer behavior

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system applies partial action by using a configurable rolling window size that balances computational load with detection accuracy. Rather than processing all historical data indefinitely, the system uses a finite window that provides sufficient statistical basis for dynamic thresholds while limiting computational requirements to what is necessary for reliable operation

Inventive Principle:

Principle #16Partial or excessive action

3Measurement precision

If multiple analysis methods (measurement flag, RoC flag, acceleration flag) are combined, then detection precision is improved, but device complexity increases

Engineering Contradiction:

Improvefault detection precisionVSAvoidanalysis system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the fault detection process into three distinct analytical components: measurement flag (comparing current values to dynamic thresholds), RoC flag (detecting rate of change), and acceleration flag (detecting changes in rate of change). Each segment handles a specific aspect of anomaly detection, and their combination provides comprehensive precision while keeping individual components relatively simple and modular

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20250341504A1Adaptive combined approach for intelligent transformer dissolved gas alarm detection

Publication Date: 2025.11.06 GE INFRASTRUCTURE TECH LLC

AI summary

Devices methods for triggering alarms and cautions for electrical equipment may include receiving, at edge device, data of an electrical device, including dissolved gas data or electrical data; setting, based on a comparison of a measurement value of the data to an upper rolling window-based threshold or to a lower rolling window-based threshold, a measurement flag for the electrical device; determining a rate-of-change (RoC) of the data; setting a RoC flag for the electrical device based on a comparison of the RoC to a delta RoC-based threshold; determining an acceleration of the data; setting an acceleration flag for the electrical device based on a comparison of the acceleration to a percentile log-ratio change of measurements threshold; and setting one of a no flag, a caution flag or an alarm flag for the electrical device based on the measurement flag, the ROC flag, and the acceleration flag.