Forecasting Hourly Photovoltaic Generation On Day Ahead

According to the new Law of Electricity Market of Ukraine, photoelectric stations will be obliged to declare their generation graphic one day ahead. Proceeding from this, the task of hourly prediction of generation of PV arises a day ahead. Since such a graph significantly depends on the change of meteorological parameters, in the work it was investigated which of them most influence generation. On the basis of the analysis it was determined that this is solar radiation, cloudiness, humidity, wind speed and temperature. The determined meteorological parameters included the construction of the neural network for forecast the hourly generation of PV on day ahead. Neural networks which proposed that is capable of predicting the generation of photovoltaic stations with a fairly high accuracy