Solar photovoltaic power output forecasting using machine learning technique
| Parent link: | Journal of Physics: Conference Series Vol. 1327 : Innovations in Non-Destructive Testing (SibTest 2019).— 2019.— [012051, 5 р.] |
|---|---|
| Autor principal: | |
| Autor corporatiu: | |
| Sumari: | Title screen Photovoltaic (PV) systems are used around the world to generate solar power. Solar power sources are irregular in nature due to the output power of PV systems being intermittent and depending greatly on environmental factors. These factors include, but are not limited to, irradiance, humidity, PV surface temperature, speed of the wind. Due to uncertainties in the photovoltaic generation, it is critical to precisely envisage the solar power generation. Solar power forecasting is necessary for supply and demand planning in an electric grid. This prediction is highly complex and challenging as solar power generation is weather-dependent and uncontrollable. This paper describes the effects of various environmental parameters on the PV system output. Prediction models based on Artificial Neural Networks (ANN) and regression models are evaluated for selective factors. The selection is done by using the correlation-based feature selection (CSF) and ReliefF techniques. The ANN model outperforms all other techniques that were discussed. |
| Publicat: |
2019
|
| Matèries: | |
| Accés en línia: | https://doi.org/10.1088/1742-6596/1327/1/012051 http://earchive.tpu.ru/handle/11683/57043 |
| Format: | Electrònic Capítol de llibre |
| KOHA link: | https://koha.lib.tpu.ru/cgi-bin/koha/opac-detail.pl?biblionumber=661314 |
| Sumari: | Title screen Photovoltaic (PV) systems are used around the world to generate solar power. Solar power sources are irregular in nature due to the output power of PV systems being intermittent and depending greatly on environmental factors. These factors include, but are not limited to, irradiance, humidity, PV surface temperature, speed of the wind. Due to uncertainties in the photovoltaic generation, it is critical to precisely envisage the solar power generation. Solar power forecasting is necessary for supply and demand planning in an electric grid. This prediction is highly complex and challenging as solar power generation is weather-dependent and uncontrollable. This paper describes the effects of various environmental parameters on the PV system output. Prediction models based on Artificial Neural Networks (ANN) and regression models are evaluated for selective factors. The selection is done by using the correlation-based feature selection (CSF) and ReliefF techniques. The ANN model outperforms all other techniques that were discussed. |
|---|---|
| DOI: | 10.1088/1742-6596/1327/1/012051 |