A study on utilization of probabilistic forecasting of photovoltaic power generation output based on regulation reserve in unit commitment scheduling

In order to maintain the electricity supply and demand balancing of an electric power system with high penetration photovoltaic (PV) power generation, the improvement of unit commitment (UC) scheduling based on a highly accurate and reliable forecasting of PV power generation is essentially important. Considering the wide variety of PV power generation depending on the change in weather conditions, the probabilistic forecast of PV power generation should be applied to UC scheduling. When the larger confidence interval is used in UC scheduling, the power supply reliability can be improved although the operation cost would be increased. In order to improve the power supply reliability while avoiding the increase in operation cost as much as possible, the proper confidence interval should be used. For this purpose, this paper proposes a novel UC scheduling method based on the adaptive confidence interval, which is selected so that the power supply flexibility (or reserve capacity) exceeds the predetermined level. As a first step of developing such a UC scheduling method, this paper demonstrates the effect of adaptive confidence interval for several different situations in terms of price of electricity purchased for compensating for the shortage of electricity supply and PV penetration level. The results suggest that the proposed method is useful when the PV penetration is huge but acceptable level, and the electricity price for shortage compensation is not expensive so much. © 2019 The Institute of Electrical Engineers of Japan.