Optimal Allocation of Curtailment Levels of PV Power Output in Different Regions in Consideration of Reduction of Aggregated Fluctuations

Due to the high penetration of photovoltaic power generation system (PV) anticipated in the future, the curtailment of PV power output becomes crucial, not only to maintain supply-demand balance but also to preserve an adequate capacity for the frequency control. When the curtailment level (CL) of the aggregated PV power output is determined in a day-ahead unit commitment (UC) scheduling, different CL should be applied to different regions with distinctive weather modes in the power system area so that the fluctuations of aggregated PV power output are minimized. The objective of this study is to optimally allocate the CL to each region based on the short-term forecasting of the weather modes so that the hourly maximum fluctuation of the aggregated PV power output (MFagg) is minimized as long as the aggregated average power output (Avgagg) becomes the same as the scheduled value in UC. Based on the past observations of PV power output, the proposed method employs relations between the regions’ MF and CL (MF-CL patterns), and relations between the regions’ Avg and CL (Avg-CL patterns) for several typical weather modes. Thus, a specific MF-CL pattern and Avg-CL pattern are determined for each region based on the short-term forecasting of the weather mode, and the CL optimization is proceeded by using these patterns. The proposed methods are tested by using the time-series of PV power output at 61 observation points in the central region of Japan for one year. As a result, it is demonstrated that merely acknowledging the weather mode of each region enabled the proposed methods to reduce MFagg significantly and these results are practically similar to the method where perfect short-term forecasting of PV power output was utilized in the entire year.