The potential of forecasting in reducing the LCOE in PV plants under ramp-rate restrictions

Abstract

An increasing number of grid codes are requiring the limitation of the PV output power fluctuation over a given time scale. Batteries represent the most obvious solution to smooth power fluctuations, with the corresponding negative impact on the PV energy cost. However, short-term forecasting is currently being proposed as a tool to reduce battery capacity requirements or even completely remove it. Although these solutions decrease or avoid the battery cost, it also entails some energy curtailment losses which obviously raise the final cost of PV energy. This energy losses, currently unknown, are independent of the forecasting accuracy and represent the minimal additional cost in the hypothetical case of a perfect prediction. Thus, this paper compares Levelized Cost of Energy (LCOE) of three ramp-rate control strategies in order to determine which would give the lowest cost: battery-based, ideal short-term forecasting, or a combination of both. Results show that curtailment losses would be small enough to make battery-less strategy an appropriate choice, so it is worthwhile improving short-term forecasting in view of the potential LCOE savings. Database is taken from high resolution measurements recorded for over a year at 8 PV plants ranging from 1 to 46 MWp.