Exception-Based Notification for Transformer Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing dissolved gas analysis (DGA) systems for transformer monitoring generate vast amounts of data, leading to information overload, which can result in delayed analysis and missed fault detections, as utilities struggle to effectively disseminate and manage the data, potentially causing catastrophic events due to inadequate timely notification of fault conditions.
Innovation Solution
A method and system for selective notification that conveys a first notice of a fault condition to responsible parties upon detection and prevents subsequent notifications unless the fault condition changes, ensuring only relevant information is communicated, thereby avoiding data inundation and ensuring timely action.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If on-line dissolved gas analysis units continuously monitor and collect data from multiple transformers, then the completeness and timeliness of fault detection is improved, but the quantity of data generated increases leading to information overload and delayed analysis
Solution Approach 1:
The patent extracts only the essential and critical information from the vast amount of DGA data by implementing exception-based notification. Instead of analyzing and notifying on all data points, the system identifies and extracts only exceptional conditions that deviate from normal operation, thereby reducing information overload while maintaining reliable fault detection.
Solution Approach 2:
The patent segments the continuous data stream into discrete exception events. By dividing the monitoring approach into normal operation (no notification) and exceptional conditions (notification triggered), the system manages the data quantity effectively while ensuring that critical faults are detected and communicated promptly.
2Speed
If data sampling frequency is increased to detect faults more rapidly, then the responsiveness to fault conditions is improved, but the volume of data requiring analysis increases exponentially
Solution Approach 1:
The system extracts only exceptional data points that indicate potential faults, ignoring normal operational data. This approach allows for high sampling frequencies to maintain detection speed while reducing the analyzable data volume to only those instances representing potential problems.
Solution Approach 2:
The patent applies partial action by monitoring all data at high frequency but taking action (notification) only on partial instances where exceptions are detected. This maintains the speed advantage of frequent sampling without requiring analysis of every data point.
3Loss of information
If all detected conditions are notified to responsible parties, then the completeness of information dissemination is improved, but the usefulness of notifications decreases due to repetitive and unnecessary alerts
Solution Approach 1:
The patent extracts only meaningful exception information from the complete set of detected conditions. By filtering out normal variations and repeating conditions, the system maintains information completeness regarding actual faults while eliminating notifications that would not contribute to effective decision-making.
Solution Approach 2:
Instead of notifying on all conditions and expecting users to identify important ones, the patent inverts the approach by notifying only on exceptional conditions that require attention. This reverses the filtering burden from the user to the system, improving notification effectiveness.
4Measurement precision
If manual analysis of large volumes of DGA data is performed, then the measurement precision of fault conditions is improved, but the loss of time increases due to repetitive nature of the work
Solution Approach 1:
The patent implements self-service by having the monitoring system automatically identify and flag exceptional conditions without requiring manual review of every data point. The system serves itself by autonomously filtering and presenting only those conditions requiring human analysis, thereby maintaining precision while reducing time loss.
Solution Approach 2:
The system performs preliminary action by automatically filtering and preparing exception data before presenting it for analysis. This preliminary processing eliminates the need for manual sifting through complete datasets, reducing analysis time while preserving the ability to detect precise fault conditions.
Data Source
Figure 1
Figure 2~4
Figure 5
AI summary
A system (30) includes an analysis engine (34) and a notification engine (38) for selective notification of a condition (72) of an apparatus (26) monitored by a monitoring device (28). A method executed by the system (30) receives (116) data elements (58) from the monitoring device (28), processes (118) the data elements (58) to detect the condition (72) of the apparatus (26), and determines that the condition (72) defines an exception (74) to a normal condition (70) of the apparatus (26). A first notice (60) of the condition (72) is conveyed (146) to a responsible party (59) at a first instance of determination of the exception (74), and conveyance of a second notice (60) is prevented (140) at a second, subsequent, instance of determination of the exception (74). In addition, communication of the normal condition (70) of the apparatus (26) is prevented.