Graphical Electric Load Categorization via V-I Trajectory Grid
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying different electric load types, such as non-intrusive load monitoring (NILM), face challenges in distinguishing between loads with similar electrical signatures and lack guidelines for optimized feature selection, leading to inefficiencies in energy management and classification accuracy.
Innovation Solution
A system that maps voltage-current trajectories to a binary cell grid, extracts features from this grid, and uses a hierarchical load feature database to categorize and identify electric load types, optimizing feature extraction and reducing computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional feature extraction methods are used for load identification, then the system can identify load types, but the performance highly depends on the electrical signatures and there is redundancy of information in feature sets
Solution Approach 1:
The patent extracts and emphasizes only the most critical feature - the voltage-current trajectory shape - while discarding redundant features from traditional NILM systems. By focusing solely on the graphical V-I trajectory and its key characteristics (peak points, inflection points, areas), the system achieves accurate load identification without the complexity of extracting and processing multiple redundant electrical signatures.
Solution Approach 2:
The patent segments the V-I trajectory into distinct characteristic points (peak points, inflection points) and regions (areas enclosed by trajectory segments). This segmentation allows the system to analyze the trajectory structure systematically, extracting meaningful features from specific segments rather than processing the entire complex waveform, thereby reducing computational complexity while maintaining identification accuracy.
2Measurement precision
If multiple feature sets are used to improve load identification, then more information is captured, but computational resources are increased and information redundancy occurs
Solution Approach 1:
The patent extracts only the essential information from the V-I trajectory - the graphical shape characteristics including peak points, inflection points, and enclosed areas - while eliminating redundant feature extraction processes. This selective extraction approach captures sufficient information for accurate load classification without the computational burden of processing multiple redundant feature sets.
Solution Approach 2:
Instead of extracting multiple electrical signatures and then analyzing them for redundancy, the patent inverts the approach by directly analyzing the V-I trajectory shape as the primary feature. This inversion eliminates the need for multiple feature extraction stages and reduces computational energy consumption while maintaining or improving classification accuracy.
3Adaptability or versatility
If standardized power supply interfaces are used across devices, then device compatibility is improved, but electrical signatures become similar and load distinction becomes difficult
Solution Approach 1:
The patent applies local quality analysis by examining different regions of the V-I trajectory separately - identifying peak points, inflection points, and enclosed areas in specific segments. Even though devices use standardized power supplies, their operational characteristics create local variations in the V-I trajectory shape, which the patent captures through region-specific analysis to distinguish between load types.
Solution Approach 2:
The patent transforms the analysis from the traditional one-dimensional electrical signature (single parameter) to a two-dimensional V-I trajectory plot, where the shape, area, and geometric characteristics provide additional dimensions for differentiation. This dimensional change enables distinction between loads with similar electrical signatures by analyzing the graphical relationship between voltage and current over time.
Data Source
AI summary
A system for different electric loads includes sensors structured to sense voltage and current signals for each of the different electric loads; a hierarchical load feature database having a plurality of layers, with one of the layers including a plurality of different load categories; and a processor. The processor acquires voltage and current waveforms from the sensors for a corresponding one of the different electric loads; maps a voltage-current trajectory to a grid including a plurality of cells, each of which is assigned a binary value of zero or one; extracts a plurality of different features from the mapped grid of cells as a graphical signature of the corresponding one of the different electric loads; derives a category of the corresponding one of the different electric loads from the database; and identifies one of a plurality of different electric load types for the corresponding one of the different electric loads.