一种电力负荷聚类方法及系统
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA ELECTRIC POWER RESEARCH INSTITUTE CO LTD
- Filing Date
- 2019-11-29
- Publication Date
- 2026-07-17
AI Technical Summary
Existing load clustering methods suffer from low computational efficiency and insufficient robustness when dealing with load time series data that are collected at asynchronous times, have unequal time lengths, or contain noise, resulting in non-unique clustering results.
By acquiring the user's load time series, inserting phase shifts to generate embedded sequences, and using linear neural regression to generate mapping function parameters, clustering is performed based on the similarity of the mapping function parameters to obtain typical load curves.
It achieves high computational performance and robustness to noise when processing asynchronous and unequal-length load data, can preserve data features to the maximum extent, and supports dynamic expansion and online clustering.
Smart Images
Figure CN110910029B_ABST