基于负荷特征的电能表非侵入式识别方法及系统
Patent Information
- Application Number
- CN202610560212.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-27
- Publication Date
- 2026-07-17
- Estimated Expiration
- 2046-04-27
AI Technical Summary
Existing technologies suffer from severe overlap of transient signals and the inability of fixed windows to adaptively segment overlapping segments when multiple devices are switched on and off intensively, leading to feature contamination and chain-like false detections, making it impossible to accurately identify load characteristics in complex power environments.
By using a non-intrusive energy meter identification method based on load characteristics, the method locks the switching events by sliding the detection window and locking the maximum likelihood ratio. It also constructs a pollution index to adaptively prune the window, accurately extracts load characteristics, and inputs them into the classification model.
It improves the purity and recognition accuracy of load features in scenarios with dense multi-device switching, avoids false detections triggered by abnormal background noise, and ensures the stability of adaptive detection and the robustness of the classification model.
Smart Images
Figure CN122221008B_ABST
Abstract
Citation Information
Patent Citations
Non-invasive load monitoring method based on event detection
CN110954744A
Power load event detection algorithm based on adaptive window model
CN121434912A