Accurate operating reserve forecasting helps the system operator to make decisions contributing to the security of the power system. It also helps market participants to adopt proper strategic bidding for the day-ahead ancillary services market to enhance their financial profit. This article proposes a stochastic hybrid method to forecast the operating reserve requirement in day-ahead electricity markets. At the first stage, based on using a modified Gray model, the day-ahead operating reserve is forecasted. In order to improve the accuracy of the operating reserve forecasting, at the next stage, a Markov chain model is used to predict the forecasting error of the Gray model. These two models are linked to each other using two different approaches—classical and fuzzy. The proposed approach is verified by the historical data of the operating reserve for spring and autumn seasons in the Khorasan Electricity Network located in Khorasan Province, Iran.

Keywords