Development of Unit Commitment Model Considering Confidence Intervals of Photovoltaics Forecast and Analysis of a Large Scale Power System

Abstract

We have developed a unit commitment (UC) model which can consider the confidence intervals of photovoltaics (PV) forecast data. We also analyzed supply-demand balance for the Tokyo Electric Power Company area with highly penetrated variable renewable energy sources in 2030. For 2030, 53GW of installed PV capacity in Japan (17.5GW in Tokyo area) was assumed. The case study was carried out to investigate the effect of the UC considering forecast error on day ahead supply-demand balance. As a result, the number of operating thermal power plants increased and the rate of operation of coal-fired power plants decreased to reserve the Load-Frequency Control capacity whose amount increased due to highly penetrated PV. Pumped storage plants stored water in the daytime, when the power of PV is peak, not in midnight to increase the amount of electric load and the operation rate of thermal power plants. Therefore broaden confidence intervals or increasing the amount of forecast errors considered in UC raised the expected operational cost.