Attenuation of Power Fluctuations in Large PV Power Plants: The Use of Prediction to Optimize the Storage System

The quality and the reliability of the power generated by large grid‐connected photovoltaic (PV) plants are negatively affected by the source characteristic variability. In this paper, we propose a technique to reduce the fluctuations in the power generated by using an energy storage system. The smoothing strategy consists in a simple moving average of past PV production in a certain time window, improved through a PV power output prediction source. We evaluated our proposed technique on the basis of a one year 5-s recorded power output from a 1.1 MW PV plant, calculating the energy storage requirement for smoothing maximum power fluctuations to below 2.5% per minute. Thanks to this day-ahead prediction, it is possible to schedule the time window duration one day in advance in order to reduce the use of the energy storage system by more than 25% and hence, to extend its service life.