Markov-based stochastic multi-period market settlement with wind uncertainties

To effectively manage uncertainties at high levels of wind penetration, stochastic unit commitment (UC) has been investigated based on scenarios, and multiple sets of prices can be directly obtained as byproducts. For day-ahead markets, however, it is crucial to obtain one set of prices/payments before the realizations of uncertainties. To achieve this goal, several pricing schemes were presented in the literature considering desired economic properties. They included single-period scenario-based models. The extension of such models to multi-period problems, however, suffers from computational complexity. Recently, we developed a Markov-based UC framework where uncertainties are represented by states instead of scenarios, leading to efficient computation. This paper develops two corresponding day-ahead pricing schemes based on sensitivity analysis to reflect the fundamental physical meaning of stochastic marginal prices. To obtain one set of day-ahead payments, ideas from the scenario-based schemes developed by Wong and Fuller are borrowed. Desired economic properties, including economic efficiency and revenue adequacy, are examined assuming truthful bidding. This work lays the foundation for Markov-based multi-period market settlement.