Transformer Oil Monitoring With Adaptive Dissolved Gas Alert Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing monitoring systems for electrical equipment immersed in insulating liquid lack a simple, reproducible, and traceable method for setting and adjusting alert thresholds, leading to poorly configured alerts that can miss potential issues, causing more serious problems and safety risks.
Innovation Solution
A monitoring system with an analysis and processing system that obtains and processes concentration measurements of dissolved molecular species, determines if predetermined conditions are met, causes alerts, and adjusts thresholds based on historical measurements without human intervention, using trained artificial intelligence for optimal threshold settings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If default alert thresholds provided by the monitor manufacturer are used, then the monitoring system can operate without configuration, but the thresholds are not optimized to any specific transformer and may provide nuisance notifications or miss actual problems
Solution Approach 1:
The monitoring system automatically collects historical concentration data, analyzes patterns specific to each transformer, and self-configures optimized alert thresholds without requiring manual technician input. The system serves itself by learning from its own operational data and automatically adapting thresholds to each specific transformer's characteristics.
Solution Approach 2:
The system performs preliminary data collection and analysis during an initial monitoring period to establish baseline concentrations and variability patterns before final threshold optimization is implemented. This preliminary action enables the system to be better prepared and customized before full operational use.
2Reliability
If alert thresholds are manually configured by technicians, then customization to specific transformers is possible, but the configuration is highly dependent on professional skill and commonly poorly configured
Solution Approach 1:
The patent replaces the manual mechanical process of technician configuration with an automated computational system that uses algorithms to analyze historical data and determine optimal thresholds. This substitution eliminates variability in professional skill and provides consistent, data-driven threshold configuration.
Solution Approach 2:
The system continuously monitors concentration data, compares actual readings against established thresholds, and uses this feedback to refine and optimize thresholds over time. The feedback loop enables the system to learn from actual operational patterns and adjust thresholds to minimize both false alarms and missed detections.
3Reliability
If alert thresholds are set to provide earliest possible notification, then detection sensitivity is improved, but nuisance notifications increase due to measurement noise
Solution Approach 1:
The system dynamically adjusts alert thresholds based on measured concentration parameters and their historical variability. By changing threshold parameters according to actual data patterns rather than using fixed values, the system maintains high detection sensitivity while adapting to the specific noise characteristics of each transformer.
Solution Approach 2:
The patent applies different threshold settings and analysis methods to different dissolved gas species and different transformers based on their individual characteristics. Each transformer receives customized thresholds optimized for its specific operational patterns and noise levels, rather than using universal thresholds.
4Measurement precision
If conventional monitors are used, then basic concentration measurement is provided, but no simple, reproducible, traceable way of setting alert thresholds is provided
Solution Approach 1:
The system discards the complexity of manual threshold configuration and recovers a simplified, automated process. By eliminating the need for manual technician input and using algorithm-based threshold determination, the system recovers a much simpler and more reproducible method that can be consistently applied across all transformers.
Data Source
AI summary
A monitoring system for a piece of electrical equipment having components immersed in electrical insulating liquid, associated methods and uses thereof are described. The monitoring system includes alert threshold determination capabilities. The monitoring system may comprise: an analysis system configured for obtaining measurements conveying concentrations of dissolved molecular species in the electrical insulating liquid; and a processing system programmed for: processing the measurements conveying concentrations of dissolved molecular species and one or more alert thresholds to determine if a predetermined condition corresponding to the one or more alert thresholds is met; causing an alert event in response to the predetermined condition being met; and performing a threshold adjustment determination process to derive new values for the one or more alert thresholds at least in part by processing historical measurements of concentrations of dissolved molecular species. Alternate embodiments may make use of an artificial intelligence engine trained using historical concentration measurements to process the measurements conveying concentrations of dissolved molecular species to adjust alarm thresholds.