Abstract
Power generation from solar and wind energy systems is highly variable due to its dependence on meteorological conditions. An efficient use of these fluctuating energy sources requires reliable forecast information for management and operation strategies. We give an overview of different applications and state-of-the-art models for solar photovoltaic power forecasting. These include physical, regressive, machine learning and time series models and generate either point or probabilistic forecsts. Additionally, techniques of collecting information on solar irradiance, cloud movement, and weather predictions are discussed, including on-site measurements by sensor networks, all-sky imaging, satellite imaging and numerical weather prediction-based models. Finally, we present an overview of error metrics that are commonly used in solar power forecasting to evaluate the performance of the forecast models.
Original language | English |
---|---|
Title of host publication | Comprehensive Renewable Energy, Second Edition |
Subtitle of host publication | Volume 1-9 |
Editors | Trevor M. Letcher |
Publisher | Elsevier |
Chapter | 1.11 |
Pages | 213-233 |
Number of pages | 21 |
Volume | 1 |
Edition | 2nd |
ISBN (Electronic) | 9780128197349 |
ISBN (Print) | 9780128197271 |
DOIs | |
Publication status | Published - Jan 2022 |
Bibliographical note
Publisher Copyright:© 2022 Elsevier Ltd. All rights reserved.
Keywords
- All-sky imaging
- Cloud motion vectors
- Grid integration of PV power
- Machine learning models
- Meteorological variables
- Numerical weather prediction
- Performance metrics
- Photovoltaics
- Point forecasts
- Probabilistic forecasts
- PV power forecasts
- PV systems
- Regression models, time series models
- Satellite imaging