.png){kind=icon}
.png){kind=icon}
Return
An event-aware and physics-guided multimodal framework for short-term photovoltaic power forecasting
.png){kind=icon}
PRE
AI
W
M
Y
X
Y
DOI：10.1016/j.asoc.2026.116171.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• PI-ECM-Mamba enables event-aware short-term PV forecasting under cloud transitions. • Physics-guided proxy cues encode cloud-transition context from sky-image observations. • Event-conditioned fusion adaptively regulates modality relevance during cloud transitions. • Conditional diffusion enriches rare cloud-transition patterns for event-aware learning. • A compact Mamba backbone supports robust 5–60 min forecasting with low latency.
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W
Organization
Cited Papers