engleski

Short-term photovoltaic power forecasting using cloud tracking methods

Concerns about greenhouse gas emissions lead to government incentives and lower prices of photovoltaic (PV) solar panels which in turn causes larger integration of renewable energy sources into the electric power system. Considerable integration of PV power generators into the power grid threatens power system secure operation due to unpredictable power fluctuations following cloud cover variations. Ramp rates of PV power plants can be in the order of seconds to minutes therefore appropriate spinning or other types of non- rotating reserves must be provided to cover production variations (in both direction positive and negative) and to achieve power system balance and economic operation. Having at disposal accurate PV production short-term forecasts, a reduced volume of operating reserves would be necessary with the added possibility to engage slower-ramping units and smaller amount of energy storage facilities. Solar forecasting, in short spatial and time scales, such as 0-1000 m and 0-30 min. is a challenging task for numerical weather predictions using satellite images due to technical restrictions (long intervals and low- resolution satellite images). Hence, in recent time, images from ground-based sky cameras are more and more used. Therefore, methods utilizing ground-based sky images successfully deal with considerable cloud spatial and temporal variability. In this paper an overview of existing methods for short-term solar forecast will be presented following with research on how to improve them in terms of forecast system equipment cost and accuracy. In the proposed research low-cost security cameras will be used to capture pictures of the sky providing input data for very short-term solar forecast at several different temporal scales.

Short-term forecast ; Cloud tracking ; Renewable energy ; PV integration ; Power system control

nije evidentirano