Visual Event Classification in Power System Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current monitoring systems for electrical systems rely on cumbersome and inaccurate manual or automated setpoints to detect events, failing to consider the contextual information such as trends and patterns in readings, which limits their effectiveness in identifying similar events in the future.
Innovation Solution
A system that allows users to visually classify events by selecting and tagging portions of readings on a graphical interface, enabling the central processing unit to identify similar characteristics in subsequent readings and apply the same classifications, along with the option to define notification profiles and perform actions based on these classifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual or automated setpoints are used to detect events, then event detection capability is provided, but accuracy and ease of operation deteriorate due to cumbersome configuration and inability to consider contextual information
Solution Approach 1:
The system automatically learns event characteristics and classifications from user interactions with the graphical interface. Users visually identify and classify events by selecting portions of readings on the graph, and the system autonomously extracts characteristics and creates detection rules without requiring manual setpoint configuration. This eliminates the cumbersome configuration process while improving detection accuracy through context-aware pattern recognition.
2Adaptability or versatility
If manual setpoints are configured to detect events, then basic event detection is achieved, but adaptability deteriorates because setpoints cannot account for contextual variations in different operational cycles
Solution Approach 1:
The system performs preliminary learning by analyzing user-classified events and extracting their characteristics in advance. When users visually select and classify events on the graphical interface, the system proactively learns the patterns, trends, and contextual features of these events, storing them as reference models for future automatic detection and classification across varying operational conditions.
Solution Approach 2:
The system incorporates feedback loops where user classifications of visually selected events are fed back into the learning mechanism. This feedback enables the system to continuously refine its understanding of event characteristics and improve its adaptability to different operational contexts, thereby enhancing both precision and versatility in event detection.
3Adaptability or versatility
If automated setpoint adjustment is implemented, then some adaptability is improved, but ease of operation worsens due to lack of contextual understanding of reading sets
Solution Approach 1:
The system replaces traditional mechanical setpoint adjustment mechanisms with a visual graphical interface. Users interact with readings through graphical selections on a displayed graph rather than manipulating numerical setpoints. The system automatically translates these visual selections into characteristic extractions and classification rules, substituting complex automated adjustment mechanics with intuitive visual interaction that simplifies operation while maintaining adaptability.
Data Source
AI summary
A system comprises a central processing unit configured to receive, from at least one component monitor, a first set of readings for a component of a power system; provide the first set of readings to be displayed as a graph on a user interface; receive, from the user interface, a user selection of a portion of the first set of readings, and at least one user classification of the portion of the first set of readings; determine at least one characteristic of the portion of the first set of readings; detect the at least one characteristic in a second set of readings; and apply the user classification to the second set of readings.