Day-Ahead Photovoltaic Power Forecasting Based on Multi-Model Ensemble and NWP Spatial Downscaling

View at publisher Summarise with VIRAT
Authors: Jingjie Cao, Tao You, Weizhou Chen, Haiyang Yuan, Ruiying Gong
Cited by: 0
References: 11

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study presents a comprehensive framework for day-ahead photovoltaic power forecasting integrating physical mechanism modeling and machine learning approaches. Four key aspects are investigated: (1) PV generation characteristics analysis revealing solar irradiance as the dominant factor with seasonal efficiency varying from 68.7% (winter) to 85.2% (summer); (2) historical power-based forecasting using ensemble methods with Neural Network achieving optimal performance (RMSE$=\mathbf{0. 0 0 2}$, QR$=100 {\%}$); (3) NWP-enhanced forecasting demonstrating significant accuracy improvement during overcast conditions; (4) spatial downscaling techniques addressing scale mismatch with machine learning downscaling achieving$\mathbf{0. 8 {\%}}$RMSE improvement. Experimental results validate the framework effectiveness, providing technical support for intelligent PV plant operation and power grid stability.

Fields of Study & Subjects

Solar Radiation and Photovoltaics Energy Load and Power Forecasting Photovoltaic System Optimization Techniques

Publication details

Journal / source
International Conference on Electrical Automation and Artificial Intelligence (ICEAAI)
Pages
274–277
Publication date
16 January 2026
Language
EN