A distributed photovoltaic reverse heavy overload prediction method and system
By constructing a reverse heavy overload prediction method that integrates long-term and short-term multidimensional feature vectors and the PatchTST model, the problem of insufficient accuracy in predicting reverse heavy overload in distributed photovoltaic systems is solved. This method enables accurate identification of reverse heavy overload risks and early warning of exceeding thresholds, thereby improving the safe and stable operation capability of the distribution network.
CN122246697APending Publication Date: 2026-06-19GUANGZHOU SHUIMU QINGHUA TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU SHUIMU QINGHUA TECH CO LTD
- Filing Date
- 2026-05-22
- Publication Date
- 2026-06-19
Smart Images
Figure CN122246697A_ABST
Abstract
This invention discloses a method and system for predicting reverse heavy overload in distributed photovoltaic (PV) systems, addressing the technical problem of poor prediction accuracy in existing PV reverse heavy overload prediction methods. The method includes acquiring current load data corresponding to the distribution transformer in the area to be tested and corresponding meteorological forecast data for the area; constructing a multi-dimensional feature vector for both long-term and short-term loads based on the current load data and meteorological forecast data; inputting the multi-dimensional feature vector for both long-term and short-term loads into a preset long-term and short-term reverse heavy overload prediction model for prediction, and outputting long-term and short-term load prediction results; and generating a target PV reverse heavy overload prediction result based on a comparison between the long-term and short-term load prediction results and preset long-term and short-term overload thresholds.
Need to check novelty before this filing date? Find Prior Art