The needs of curtailing power output of photovoltaic power generation (PV) will be increased by the increase in penetration so that the power supply-demand balancing capability is preserved. This study evaluates the accuracy of a day-ahead forecast of curtailed PV power output compared to the accuracy without curtailment. The aggregated PV power output for a year is considered together with a single point PV power output. The results show that the positive error can be decreased due to the upper limit of PV power output by curtailment. The improvement would be smaller in the case of aggregated PV power output due to the smoothing effect of forecast error among different observation points.
