Improving Regional PV Power Curtailment with Better Day-ahead PV Forecasts: An Evaluation of 3 Forecasts

The objective of this study is to evaluate how improvements of day-ahead forecasts of photovoltaic, PV, power accuracy affect PV power curtailment planning, in a scenario of PV overgeneration. To do that, we simulated one year of operation of the power system of Kanto region in Japan, in a scenario of high penetration of PV, and with curtailment. Different ways to operate the power system were simulated using a unit-commitment procedure and 3 types of day-ahead forecasts of PV, each with a different level of accuracy. The results show that, comparing with a naive forecast, current forecasting accuracy can reduce the annual required PV curtailment by 51%. If a perfect forecast could be used, curtailment of PV could be further reduced by 53%, indicating a considerable value in improving current forecasts from the point of view of curtailment scheduling.