Assessing the value of improved forecasts on variable renewable energy (VRE) power generation

Abstract

The variable renewable energy (VRE) has been rapidly penetrated around the world including Korea to reduce greenhouse gas emission in the power generation sector in efforts to cope with climate change and achieve the target of carbon neutrality. However, VRE power generation is inevitably affected by real-time weather conditions, resulting in deviations between the day-ahead forecast and the real-time actual power generation, which incurs additional inefficient operating costs for system operators to resolve these forecast errors. This study examines the marginal effect of VRE forecast errors on the CON and COFF payments which are the balancing cost of the Korea’s power market through econometric analysis in terms of opportunity costs leveraging empirical data from Korea Power Exchange (KPX) from 2018 to 2020, thereby analyzing how much the marginal value of improved VRE forecasts is and how this value can vary depending on power market situations. Our results show that the marginal value of 4.154 KRW/kWh which is derived from our basic model is slightly higher than the current incentive level in Korea, which emphasizes the implementation of more subdivided incentive system that provides more incentives for more accurate forecasts. Moreover, given the continuous trend of VRE penetration, system operators need to prepare more for unexpected increase than unexpected decrease in VRE power generation with considering the introduction of an asymmetric incentive system that can induce over-forecasting rather than under-forecasting of VRE power generation. Finally, we find that the values of improved VRE forecasts during the daytime when solar PV power unexpectedly ramps are significantly greater than the overall value of improved forecasts, which means that higher time-varying incentives are needed to accurate forecast solar PV power generation.