Modeling Real-Time Balancing Power Market Prices Using Combined SARIMA and Markov Processes

This paper describes modeling of real-time balancing power market prices by using combined seasonal auto regressive integrated moving average (SARIMA) and discrete Markov processes. The combination of such processes allows generation of price series with periods where no demand for balancing power exists. The purpose of the model is simulation of prices to construct scenario trees representing possible realization of the stochastic prices. Such scenario trees can be used in planning models based on stochastic optimization to generate bid sequences to the balancing market. The spread of the prices in the tree and the shape of the scenarios are of central importance. Model parameter estimation methods reflecting the demands on scenario trees have therefore been used. The proposed model is also applied to data from the Nordic power market. The conclusion of this paper is that the developed model is appropriate for modeling real-time balancing power prices.