GIS Failure Diagnosis Using Noise and Discharge Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing gas insulated switchgear (GIS) failure are inaccurate and lack effective preventive diagnostic technologies, leading to increased risk of equipment damage and safety hazards due to insufficient maintenance.
Innovation Solution
A method involving a noise classification model and a failure mode classification model using convolutional neural networks to distinguish between noise and failure signals, specifically classifying corona, floating, and void discharges, without assigning a separate class to noise signals, thereby improving signal identification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional diagnostic methods are used for GIS failure detection, then the diagnostic process is simple, but the diagnosis accuracy is low leading to misclassification of noise signals
Solution Approach 1:
The patent segments the diagnostic task into two distinct classification models: a noise classification model and a failure mode classification model. This segmentation allows each model to specialize in its specific function, improving overall diagnosis accuracy while managing complexity through modular architecture
Solution Approach 2:
The patent introduces an intermediary noise classification model that acts as a mediator between the raw signal input and the failure mode classification. This intermediary filters and categorizes noise signals before they reach the failure detection system, preventing misclassification and improving diagnostic precision
2Measurement precision
If noise signals are assigned separate classes in the classification model, then the model structure is simple, but noise signals are misclassified as failure signals reducing diagnostic accuracy
Solution Approach 1:
The patent separates noise classification and failure mode classification into two distinct models rather than using a single unified model. This segmentation prevents noise signals from being misclassified as failure signals while maintaining clear model boundaries and functionality
Solution Approach 2:
The patent extracts and isolates noise signals from the failure detection process by creating a dedicated noise classification model. This extraction prevents noise interference in the failure mode classification, improving diagnostic accuracy without complicating the core failure detection logic
Data Source
AI summary
A method of diagnosing GIS (gas insulated switchgear) failure and apparatus for performing method can include classifying, by a noise classification model, a noise signal and a failure signal. The method can also include classifying, by a failure mode classification model, a failure signal type based on the failure signal.