Scenario based uncertainty modeling of electricity market prices

Abstract

Energy trading in liberalized electricity markets is a decision-making problem that is modeled considering price uncertainty. Stochastic programming is a natural platform for modeling such decision-making problems, where uncertainties are characterized through scenarios. Scenarios are possible outcomes of random process with corresponding occurrence probabilities. A large number of scenarios are required for accurate modeling of any uncertainty. However, due to computational complexity and time limitations, generated scenarios are required to be reduced. This paper presents a efficacious algorithm for generation and reduction of electricity market price scenarios. Time series based Auto Regressive Moving Average (ARMA) model is used for scenario generation while Probability Distance Based Backward reduction method is utilized for scenario reduction. Proposed algorithm is illustrated through practical case study based on PJM day-ahead electricity market. Statistical analysis validates the proposed algorithm and comparison between ARMA and heuristic model for scenario generation reflect strength of proposed algorithm for modeling electricity market price uncertainty.