A photovoltaic power prediction method and medium
By acquiring and grouping image-power aligned samples, a multi-step prediction model was used to solve the problem of information interference in short-term photovoltaic power prediction, achieving more accurate and stable prediction results.
Patent Information
- Application Number
- CN202610846189.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-12
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies struggle to effectively combine sky images, historical photovoltaic power, and image-derived physical diagnostic features in short-term photovoltaic power prediction, leading to information interference and instability in the prediction model. This is especially true when cloud cover changes rapidly, resulting in response lag and significant prediction bias.
By acquiring timestamped sky images and photovoltaic power data, a clear-sky reference power is calculated, image-derived physical diagnostic features are extracted and grouped according to physical mechanisms, image-power aligned samples are constructed, and a multi-step prediction model is used to predict photovoltaic power.
It improves the accuracy and interpretability of short-term photovoltaic power prediction, reduces interference from heterogeneous information, provides a more complete input basis, and ensures that the model can respond in a timely manner when cloud conditions change rapidly.
Smart Images
Figure CN122418633A_ABST